AI网关新增xAI格式TTS端点与过滤弃用模型开关

- POST /ai/v1/tts:xAI 官方 TTS 格式转换层(text/voice_id→input/voice),复用 Speech 编排,model 缺省 xai.grok-tts;实测 output_format/speed 透传生效 - 「过滤弃用模型」开关(GET/PUT /api/v1/ai-settings,settings 表持久化):开启后已宣布弃用(未退役)模型从列表与路由排除;同步入库与退役提醒不受影响 - AI 网关文档更名 docs/AI网关.md,补 tts 矩阵段与开关说明;README 引用同步 - CHANGELOG 0.5.0(0.4.0 已发版冻结);DASH_VERSION v0.5.0
AI网关新增TTS/重排/审核端点,xAI工具扩展,swagger修缺
2026-07-14 09:49:28 +08:00 · 2026-07-13 20:17:06 +08:00 · 2026-07-13 15:06:01 +08:00 · 2026-07-13 12:23:18 +08:00 · 2026-07-13 10:13:44 +08:00 · 2026-07-12 17:48:28 +08:00
133 changed files with 21915 additions and 4174 deletions
@@ -23,19 +23,20 @@ jobs:
          echo "IMAGE=${REGISTRY}/${{ gitea.repository }}" >> "$GITHUB_ENV"
          echo "BUILD_TIME=$(date -u +%Y-%m-%dT%H:%M:%SZ)" >> "$GITHUB_ENV"

-      - name: 下载前端最新 dist.zip 解压进嵌入目录
+      - name: 下载前端 dist.zip
        env:
          TOKEN: ${{ secrets.BUILD_TOKEN }}
        run: |
-          DASH_TAG=$(curl -fsSL -H "Authorization: token ${TOKEN}" \
-            "${{ gitea.server_url }}/api/v1/repos/${DASH_REPO}/releases/latest" | jq -r '.tag_name')
-          curl -fsSL -H "Authorization: token ${TOKEN}" \
-            -o dist.zip "${{ gitea.server_url }}/${DASH_REPO}/releases/download/${DASH_TAG}/dist.zip"
+          DASH_TAG=$(head -1 DASH_VERSION)
+          BASE="${{ gitea.server_url }}/${DASH_REPO}/releases/download/${DASH_TAG}"
+          curl -fsSL -H "Authorization: token ${TOKEN}" -o dist.zip "${BASE}/dist.zip"
+          curl -fsSL -H "Authorization: token ${TOKEN}" -o dist.zip.sha256 "${BASE}/dist.zip.sha256"
+          sha256sum -c dist.zip.sha256
          rm -rf internal/webui/dist && mkdir -p internal/webui/dist
          unzip -q dist.zip -d internal/webui/dist
          test -f internal/webui/dist/index.html

-      - name: 构建双架构二进制(artifact stage 导出)
+      - name: 构建双架构二进制
        run: |
          docker buildx build --pull \
            --platform linux/amd64,linux/arm64 \
@@ -46,8 +47,9 @@ jobs:
          mkdir -p assets
          mv output/linux_amd64/oci-portal-server assets/oci-portal-server-linux-amd64
          mv output/linux_arm64/oci-portal-server assets/oci-portal-server-linux-arm64
+          (cd assets && sha256sum * > SHA256SUMS)

-      - name: 构建并推送多平台镜像(与上一步共享构建缓存)
+      - name: 构建并推送多平台镜像
        env:
          TOKEN: ${{ secrets.BUILD_TOKEN }}
        run: |
@@ -65,7 +67,7 @@ jobs:
          awk -v ver="${TAG#v}" '$0 ~ "^## \\[" ver "\\]" {flag=1; next} /^## \[/ && flag {exit} flag {print}' CHANGELOG.md > release-notes.md
          [ -s release-notes.md ] || echo "Release ${TAG}" > release-notes.md

-      - name: 创建 Release 并上传二进制附件(重跑幂等,同名附件自动替换)
+      - name: 创建 Release 并上传二进制附件
        uses: https://gitea.com/actions/gitea-release-action@v1
        with:
          token: ${{ secrets.BUILD_TOKEN }}
@@ -10,7 +10,7 @@ permissions:

 env:
  BUILDX_NO_DEFAULT_ATTESTATIONS: "1"
-  DASH_DOWNLOAD: https://github.com/wangdefaa/oci-portal-dash/releases/latest/download/dist.zip
+  DASH_REPO: wangdefaa/oci-portal-dash
  IMAGE: ghcr.io/wangdefaa/oci-portal

 jobs:
@@ -21,14 +21,18 @@ jobs:

      - uses: docker/setup-buildx-action@v3

-      - name: 下载前端最新 dist.zip 解压进嵌入目录
+      - name: 下载前端 dist.zip
        run: |
-          curl -fsSL -o dist.zip "$DASH_DOWNLOAD"
+          DASH_TAG=$(head -1 DASH_VERSION)
+          BASE="https://github.com/${DASH_REPO}/releases/download/${DASH_TAG}"
+          curl -fsSL -o dist.zip "${BASE}/dist.zip"
+          curl -fsSL -o dist.zip.sha256 "${BASE}/dist.zip.sha256"
+          sha256sum -c dist.zip.sha256
          rm -rf internal/webui/dist && mkdir -p internal/webui/dist
          unzip -q dist.zip -d internal/webui/dist
          test -f internal/webui/dist/index.html

-      - name: 构建双架构二进制(artifact stage 导出)
+      - name: 构建双架构二进制
        run: |
          BUILD_TIME=$(date -u +%Y-%m-%dT%H:%M:%SZ)
          echo "BUILD_TIME=${BUILD_TIME}" >> "$GITHUB_ENV"
@@ -41,6 +45,7 @@ jobs:
          mkdir -p assets
          mv output/linux_amd64/oci-portal-server assets/oci-portal-server-linux-amd64
          mv output/linux_arm64/oci-portal-server assets/oci-portal-server-linux-arm64
+          (cd assets && sha256sum * > SHA256SUMS)
          ls -lh assets/

      - name: 提取 CHANGELOG 版本段作为 Release 描述
@@ -1 +1 @@
-0.6.5
+0.6.6
@@ -1,6 +1,6 @@
 # Backend Development Guidelines(oci-portal 后端规范)

-> Go 后端(`oci-portal/`)编码规范入口。规范提炼自 Google / Uber Go Style Guide、Effective Go 与 Go 官方模块布局指南,按本项目裁剪;条目与项目约定冲突时,以本 spec 为准。迁自 docs/开发指南.md 与根目录 AGENTS.md(2026-07)。
+> Go 后端(`oci-portal/`)编码规范入口。规范提炼自 Google / Uber Go Style Guide、Effective Go 与 Go 官方模块布局指南,按本项目裁剪;条目与项目约定冲突时,以本 spec 为准。

 ---

@@ -51,7 +51,7 @@ go mod tidy                     # 依赖有变更时
 全部 HTTP 接口带 swaggo 注释(@Summary 中文/@Tags 按域/@Param/@Success/@Router;JWT 组接口标 @Security BearerAuth)。**新增或修改接口必须同步注释**,并重新生成 spec:

 ```bash
-go tool swag init -g cmd/server/main.go -o docs --parseInternal --parseDependency
+go tool swag init -g cmd/server/main.go -o docs --parseInternal --parseDependency --overridesFile docs/.swaggo
 ```

 生成的 `docs/` 一并提交;UI 由 `SWAGGER=1` 开启(/swagger/index.html)。同名 handler 方法(list/create/get/update/remove)分布在多个 struct 上,写注释时确认 @Router 路径与该 receiver 的真实注册一致(routes_*.go)。
@@ -183,7 +183,7 @@ Complex task: ask the user if you can create a Trellis task and enter the planni
 - 1.0 Create task `[required · once]` (only after task-creation consent)
 - 1.1 Requirement exploration `[required · repeatable]` (`prd.md`; complex tasks also need `design.md` + `implement.md`)
 - 1.2 Research `[optional · repeatable]`
- 1.3 Configure context `[required · once]` — Claude Code, Cursor, OpenCode, Codex, Kiro, Gemini, Qoder, CodeBuddy, Copilot, Droid, Pi, ZCode, Reasonix (sub-agent-dispatch platforms only; inline platforms skip)
+- 1.3 Configure context `[required · once]` — Claude Code, Cursor, OpenCode, Codex, Kiro, Gemini, Qoder, CodeBuddy, Copilot, Droid, Pi, Oh My Pi, ZCode, Reasonix (sub-agent-dispatch platforms only; inline platforms skip)
 - 1.4 Activate task `[required · once]` (review gate, then `task.py start`; status → in_progress)
 - 1.5 Completion criteria

@@ -272,13 +272,13 @@ Code committed. Run `/trellis:finish-work`; if dirty, return to Phase 3.4 first.

 When a user request matches one of these intents inside an active task, route first, then load the detailed phase step if needed.

-[Claude Code, Cursor, OpenCode, codex-sub-agent, Kiro, Gemini, Qoder, CodeBuddy, Copilot, Droid, Pi, ZCode, Reasonix, Trae]
+[Claude Code, Cursor, OpenCode, codex-sub-agent, Kiro, Gemini, Qoder, CodeBuddy, Copilot, Droid, Pi, Oh My Pi, ZCode, Reasonix, Trae]

 - Planning or unclear requirements -> `trellis-brainstorm`.
 - `in_progress` implementation/check -> dispatch `trellis-implement` / `trellis-check`.
 - Repeated debugging -> `trellis-break-loop`; spec updates -> `trellis-update-spec`.

-[/Claude Code, Cursor, OpenCode, codex-sub-agent, Kiro, Gemini, Qoder, CodeBuddy, Copilot, Droid, Pi, ZCode, Reasonix, Trae]
+[/Claude Code, Cursor, OpenCode, codex-sub-agent, Kiro, Gemini, Qoder, CodeBuddy, Copilot, Droid, Pi, Oh My Pi, ZCode, Reasonix, Trae]

 [codex-inline, Kilo, Antigravity, Devin]

@@ -353,7 +353,7 @@ Return to this step whenever requirements change and revise the relevant artifac

 Research can happen at any time during requirement exploration. It isn't limited to local code — you can use any available tool (MCP servers, skills, web search, etc.) to look up external information, including third-party library docs, industry practices, API references, etc.

-[Claude Code, Cursor, OpenCode, codex-sub-agent, Kiro, Gemini, Qoder, CodeBuddy, Copilot, Droid, Pi, ZCode, Reasonix, Trae]
+[Claude Code, Cursor, OpenCode, codex-sub-agent, Kiro, Gemini, Qoder, CodeBuddy, Copilot, Droid, Pi, Oh My Pi, ZCode, Reasonix, Trae]

 Spawn the research sub-agent:

@@ -361,7 +361,7 @@ Spawn the research sub-agent:
 - **Task description**: Research <specific question>
 - **Key requirement**: Research output MUST be persisted to `{TASK_DIR}/research/`

-[/Claude Code, Cursor, OpenCode, codex-sub-agent, Kiro, Gemini, Qoder, CodeBuddy, Copilot, Droid, Pi, ZCode, Reasonix, Trae]
+[/Claude Code, Cursor, OpenCode, codex-sub-agent, Kiro, Gemini, Qoder, CodeBuddy, Copilot, Droid, Pi, Oh My Pi, ZCode, Reasonix, Trae]

 [codex-inline, Kilo, Antigravity, Devin]

@@ -380,7 +380,7 @@ Brainstorm and research can interleave freely — pause to research a technical

 #### 1.3 Configure context `[required · once]`

-[Claude Code, Cursor, OpenCode, codex-sub-agent, Kiro, Gemini, Qoder, CodeBuddy, Copilot, Droid, Pi, ZCode, Reasonix, Trae]
+[Claude Code, Cursor, OpenCode, codex-sub-agent, Kiro, Gemini, Qoder, CodeBuddy, Copilot, Droid, Pi, Oh My Pi, ZCode, Reasonix, Trae]

 Curate `implement.jsonl` and `check.jsonl` so the Phase 2 sub-agents get the right spec/research context. These files were seeded on `task create` with a single self-describing `_example` line; your job here is to fill in real entries.

@@ -425,7 +425,7 @@ Ready gate: both `implement.jsonl` and `check.jsonl` must contain at least one r

 Skip this step only when both files already have real curated entries.

-[/Claude Code, Cursor, OpenCode, codex-sub-agent, Kiro, Gemini, Qoder, CodeBuddy, Copilot, Droid, Pi, ZCode, Reasonix, Trae]
+[/Claude Code, Cursor, OpenCode, codex-sub-agent, Kiro, Gemini, Qoder, CodeBuddy, Copilot, Droid, Pi, Oh My Pi, ZCode, Reasonix, Trae]

 [codex-inline, Kilo, Antigravity, Devin]

@@ -458,11 +458,11 @@ If `task.py start` errors with a session-identity message (no context key from h
 | `design.md` exists (complex tasks) | ✅ |
 | `implement.md` exists (complex tasks) | ✅ |

-[Claude Code, Cursor, OpenCode, codex-sub-agent, Kiro, Gemini, Qoder, CodeBuddy, Copilot, Droid, Pi, ZCode, Reasonix, Trae]
+[Claude Code, Cursor, OpenCode, codex-sub-agent, Kiro, Gemini, Qoder, CodeBuddy, Copilot, Droid, Pi, Oh My Pi, ZCode, Reasonix, Trae]

 | `implement.jsonl` and `check.jsonl` each contain at least one real curated entry (seed row does not count) | ✅ |

-[/Claude Code, Cursor, OpenCode, codex-sub-agent, Kiro, Gemini, Qoder, CodeBuddy, Copilot, Droid, Pi, ZCode, Reasonix, Trae]
+[/Claude Code, Cursor, OpenCode, codex-sub-agent, Kiro, Gemini, Qoder, CodeBuddy, Copilot, Droid, Pi, Oh My Pi, ZCode, Reasonix, Trae]

 ---

@@ -472,7 +472,7 @@ Goal: turn reviewed planning artifacts into code that passes quality checks.

 #### 2.1 Implement `[required · repeatable]`

-[Claude Code, Cursor, OpenCode, CodeBuddy, Droid, Pi]
+[Claude Code, Cursor, OpenCode, CodeBuddy, Droid, Pi, Oh My Pi]

 Spawn the implement sub-agent:

@@ -484,7 +484,7 @@ The platform hook/plugin auto-handles:
 - Reads `implement.jsonl` and injects referenced spec/research files into the agent prompt
 - Injects `prd.md`, `design.md` if present, and `implement.md` if present

-[/Claude Code, Cursor, OpenCode, CodeBuddy, Droid, Pi]
+[/Claude Code, Cursor, OpenCode, CodeBuddy, Droid, Pi, Oh My Pi]

 [codex-sub-agent, Gemini, Qoder, Copilot, ZCode, Reasonix, Trae]

@@ -526,7 +526,7 @@ The platform prelude auto-handles the context load requirement:

 #### 2.2 Quality check `[required · repeatable]`

-[Claude Code, Cursor, OpenCode, codex-sub-agent, Kiro, Gemini, Qoder, CodeBuddy, Copilot, Droid, Pi, ZCode, Reasonix, Trae]
+[Claude Code, Cursor, OpenCode, codex-sub-agent, Kiro, Gemini, Qoder, CodeBuddy, Copilot, Droid, Pi, Oh My Pi, ZCode, Reasonix, Trae]

 Spawn the check sub-agent:

@@ -540,7 +540,7 @@ The check agent's job:
 - Auto-fix issues it finds
 - Run lint and typecheck to verify

-[/Claude Code, Cursor, OpenCode, codex-sub-agent, Kiro, Gemini, Qoder, CodeBuddy, Copilot, Droid, Pi, ZCode, Reasonix, Trae]
+[/Claude Code, Cursor, OpenCode, codex-sub-agent, Kiro, Gemini, Qoder, CodeBuddy, Copilot, Droid, Pi, Oh My Pi, ZCode, Reasonix, Trae]

 [codex-inline, Kilo, Antigravity, Devin]

@@ -11,7 +11,8 @@
      "env": {
        "DATA_KEY": "dev-data-key",
        "JWT_SECRET": "dev-jwt-secret",
-        "ADMIN_PASSWORD": "admin123"
+        "ADMIN_PASSWORD": "admin123",
+        "GIN_MODE": "debug"
      }
    },
    {
@@ -1,8 +1,120 @@
 # Changelog

-格式遵循 [Keep a Changelog](https://keepachangelog.com/zh-CN/1.1.0/)，版本号遵循语义化版本。
+格式参考 [Keep a Changelog](https://keepachangelog.com/zh-CN/1.1.0/)（版本段不记日期），版本号遵循语义化版本。

-## [0.0.1] - 2026-07-09
+## [0.5.0]
+
+### Added
+
+- xAI 官方格式文本转语音端点 `POST /ai/v1/tts`：接受 xAI 官方 TTS 请求（`text` / `language` 必填，`voice_id`、`output_format`{codec, sample_rate, bit_rate}、`speed` 等），网关转换为 OpenAI 兼容形态后与 `/ai/v1/audio/speech` 走同一上游与渠道调度；`model` 为网关扩展字段（缺省 `xai.grok-tts`）；实测 `output_format` 对象与 `speed` 透传生效，xAI SDK / 客户端可直接指向网关
+- 「过滤弃用模型」开关（`GET` / `PUT /api/v1/ai-settings`，持久化）：开启后 OCI 已宣布弃用（即使未到退役日）的模型从模型列表与路由中同时排除，关闭恢复；渠道同步入库与 30 天退役提醒不受影响
+
+### Changed
+
+- AI 网关文档更名为 `docs/AI网关.md`（原 `docs/ai-gateway.md`），README 引用同步
+
+## [0.4.0]
+
+### Added
+
+- 文本转语音端点 `POST /ai/v1/audio/speech`：OpenAI Audio Speech 兼容，直通 OCI 兼容面（模型 `xai.grok-tts`，voice 取 xAI Grok Voice 列表 `ara`/`eve`/`leo`/`rex`/`sal`）；上游必填的 `language` 缺省时自动注入 `"auto"`，xAI 专属参数平铺透传；响应为一次性完整音频（Content-Type 透传上游，缺省 audio/mpeg），不提供流式
+- 文档重排端点 `POST /ai/v1/rerank`：Jina / Cohere 通行协议，走 OCI typed 面（`cohere.rerank-v4.0-pro` / `-fast`），支持 `top_n` 与 `return_documents`，`results[].index` 指向入参 `documents` 下标
+- 内容审核端点 `POST /ai/v1/moderations`：OpenAI moderations 外壳映射 OCI Guardrails（内容审核 / PII / 提示注入），`input` 为字符串或字符串数组（单次至多 8 条），categories 用 OCI 原生维度 `overall` / `blocklist` / `prompt_injection`（任一得分 ≥0.5 判 `flagged`），PII 命中放扩展字段 `results[].pii`（不参与 flagged；实测中文人名 / 手机号识别较弱，英文正常）
+- Responses 服务端工具扩展：放行 Oracle 文档化的 xAI `code_interpreter` 与远程 `mcp` 工具，并解除 `web_search` / `x_search` 仅非流式的限制——四类服务端工具非流式与流式均实测可用；`code_interpreter` 的命名容器管理与 File Search 不提供
+- 模型能力映射扩展：`TEXT_RERANK` → RERANK、`TEXT_TO_AUDIO` → TTS，渠道探测 / 同步自动发现重排与语音模型入池
+
+### Changed
+
+- AI 网关文档独立成篇 `docs/ai-gateway.md`（端点定位、兼容边界、已知上游限制、字段兼容矩阵——矩阵只列支持项，不支持 / 忽略项以文字简述），README 精简为概览并引用；实测披露：Responses `input_file` 内容块被上游以 ZDR 形态拒绝，能力不可用
+- swagger 全面修缺：135 处响应注解从泛型 map 改为具体类型（补文档用途响应结构与泛型列表外壳），AI 网关端点请求 / 响应 schema 完整可见；`json.RawMessage` 与联合类型统一渲染为任意 JSON 值（AnyJSON），不再误显示为 byte 数组；swag overrides 配置移至 `docs/.swaggo`，生成命令加 `--overridesFile docs/.swaggo`；顺带修正网页控制台会话创建响应码（200→201）与 SAML 元数据下载（改文件响应）两处注解失真
+
+## [0.3.1]
+
+### Added
+
+- AI 网关流式断流自动降级：Messages 与 Chat Completions 流式请求在客户端尚未收到任何输出时遭遇上游断流，自动降级为非流式重做，结果按标准事件 / chunk 序列一次推送，调用日志记 `retries=1` 与降级标记。实测 OCI 兼容面对 `instructions` 与 `tools` 合计超约 64.5KB 的流式请求会静默断连（无错误事件，非流式正常，消息正文不计入），Claude Code 等大体量系统提示客户端极易触发；Responses 直通因初始事件已转发无法透明降级，日志记「上游流提前终止」；README 增补「已知上游限制：大 system 区流式断流」小节
+
+### Changed
+
+- Messages 的 `max_tokens` 改为可缺省：缺省或 ≤0 时按默认值 8192 放行（此前返回 400；部分客户端将该字段视为选填）
+
+### Fixed
+
+- AI 网关流式假成功：上游 error / `response.failed` 事件此前被吞掉，流提前 EOF 也被伪装成正常结束，调用日志呈现 200 · 0/0 且无错误信息；现 Messages 将上游失败转为 Anthropic `error` 事件，三个流式端点日志均记录上游错误消息或「上游流提前终止，未返回终态事件」
+
+## [0.3.0]
+
+### Added
+
+- 恢复 OpenAI Chat Completions 兼容端点 `/ai/v1/chat/completions`：请求经 OCI Responses 接口转换转发，支持非流式与 SSE、文本与图片、function 工具调用、结构化输出、推理力度、模型白名单、缓存 token 用量和调用日志
+- 代理侧批量关联租户接口 `PUT /api/v1/proxies/{id}/tenants`：一次调用设置关联该代理的租户全集——列表内租户改挂到本代理（含从其他代理改挂），列表外已关联的解除，事务生效并返回最新关联计数；含无效租户 ID 返回 400，代理不存在返回 404
+- `/api/v1/about` 扩展运行时信息：进程启动时刻 `startedAt`、运行秒数 `uptimeSeconds` 与资源占用 `resources`——CPU 自启动均值（占单核百分比）、CPU 核数、goroutine 数、Go 堆内存与向系统申请内存、数据库引擎与磁盘占用（SQLite 统计主文件与 -wal / -shm；MySQL / PostgreSQL 外部库返回 -1 表示不可度量）
+- README 重构部署、配置、鉴权、反向代理、AI 网关与升级发布说明，并新增 Responses、Chat Completions、Embeddings、Anthropic Messages 与标准接口的字段兼容矩阵，明确直通、转换、降级、忽略和拒绝边界
+
+### Changed
+
+- 云端关键事件通知补充操作者、来源 IP、执行结果和说明信息；支持从 OCI Audit 与 IDCS 登录事件提取成功 / 失败状态、策略描述和登录失败原因，缺失字段统一显示占位符
+
+### Removed
+
+- **移除自定义审计告警规则**及其阈值、窗口和冷却匹配能力；`/api/v1/log-events/alert-rules` 与 `/api/v1/log-events/alert-rules/{ruleId}` 管理接口、`audit_alert` 通知模板不再提供，已有规则升级后不再执行
+
+## [0.2.0]
+
+### Added
+
+- AI 网关模型黑名单：面板集中拉黑模型，全渠道剔除（模型列表 / 路由 / 探测候选不再出现，同步不入库），移除条目并重新同步后恢复；新表 `ai_model_blacklists`（启动自动迁移），管理接口 `/api/v1/ai-blacklist`
+- AI 密钥模型白名单：密钥可限定放行的模型列表（空 = 不限），白名单外调用 404，模型列表只返回交集
+- 推理力度（effort）直通：Responses `reasoning.effort` 与 Anthropic Messages `output_config.effort` 原样透传上游、不做档位校验；grok-4.3 五档（`none` 完全关闭思考）、gpt-oss `minimal`-`high`、multi-agent `xhigh`（控制并行 agent 数）实测生效
+- xAI Grok 服务端工具：`/ai/v1/responses` 的 `tools` 支持 `{"type":"web_search"}` / `{"type":"x_search"}`（官方 Agent Tools 格式，可与 `function` 混用，仅非流式），响应保留 `web_search_call` 输出项与引用
+
+### Changed
+
+- **AI 网关对话链路整体切换到 OCI OpenAI 兼容面**（`/actions/v1/responses`），typed chat 面全链路移除：`/ai/v1/responses` 请求体原样直通（仅强制 `store:false`），流式为上游原生事件流，gpt-oss 推理增量直达客户端；`/ai/v1/messages` 转换为 Responses 请求直通，流式经事件桥转回标准 Anthropic 事件序列，`stop_sequences` / `top_k` 等无兼容面对应物的字段改为静默忽略，思考输出不转 `thinking` 块；渠道探测改用兼容面最小请求；流式建立后绑定渠道，中断不重试不计熔断
+- 后台任务「立即执行」改异步触发：接口立即返回 202，执行结果经任务日志轮询呈现；同一任务在途时重复触发返回 409「任务正在执行中」，与 cron 重叠触发静默跳过（此前同步阻塞数十秒且同一任务可并发重复执行）
+- 渠道探测候选跨厂商分散（上限 8，voice 等不可对话形态排除）：单一厂商在区域内全为微调基座时不再拖垮整个渠道的探测结论
+- 网关调用遇「模型不可按需调用」类错误（微调基座 400 / 实体不存在 404）自动换渠道重试且不计入渠道熔断，错误信息带模型名便于加入黑名单；仅鉴权类错误才判定租户无配额
+- 同名模型多条目去重优先保留非微调基座条目，降低缓存到不可调用 OCID 的概率
+
+### Removed
+
+- **`/ai/v1/chat/completions` 端点移除**（404）：OpenAI 官方新能力已集中到 Responses 且本面板真实流量为零，请迁移 `/ai/v1/responses` 或 `/ai/v1/messages`
+- google / cohere 对话模型不再供给（OCI OpenAI 兼容面不承接，上游 400；cohere embed 向量模型不受影响），模型目录按 `xai.` / `meta.` / `openai.` 厂商前缀过滤
+
+### Fixed
+
+- 渠道「同步模型成功但探测报错不可用」：法兰克福等区域的微调基座模型占满探测候选导致的误报（探测状态与真实可用性不符的根因）
+- 租户删除：告警命中清理改子查询，日志事件数万条时不再超出 SQL 绑定变量上限导致删除失败；无法解析的任务 payload 记警告跳过并保留原任务，不再永久阻断租户删除
+
+## [0.1.0]
+
+### Added
+
+- 通知渠道扩展：Webhook（Slack / 飞书 / 钉钉 / 企微预设）、ntfy、Bark、SMTP 邮件与 Telegram 五渠道并存，逐渠道推送互不影响，密钥加密落库
+- 审计告警规则：按租户 / 事件 / 资源 / 来源 IP（CIDR in|notin）组合匹配，阈值窗口计数与冷却，命中即推送模板通知
+- 撤销全部会话：`POST /auth/revoke-sessions` 使全部旧令牌立即失效，响应携带操作者的新令牌
+- 租户暂停状态识别：测活经账户 capabilities 判定 `suspended`，独立于存活 / 失效展示
+- 抢机成功通知补全实例明细：租户别名、区域别名、类型与配置、镜像、公网 IPv4 与 root 密码
+
+### Changed
+
+- 敏感身份操作（修改凭据、TOTP 开关、外部身份绑定 / 解绑、密码登录开关）成功后令牌版本递增：全部旧 JWT 立即失效，响应返回新令牌无感续期（原 204 改为 200 + token）
+- 管理员徽标与管理员组识别支持双候选组名（Administrators / OCI_Administrators）
+- Release 流程：前端产物版本由 `DASH_VERSION` 显式固定并做 sha256 校验，二进制附 `SHA256SUMS`；docker-compose 示例默认只绑定 `127.0.0.1`
+- README 新增「反向代理」小节（Caddy / nginx / Traefik 示例与路由前缀清单）
+
+### Fixed
+
+- 删除租户改为单事务清理全部关联本地数据（任务及日志、测活 / 成本快照、回传事件、告警规则、AI 渠道、Webhook 密钥），等待在途任务执行结束，迟到写入不再残留孤儿数据
+
+### Security
+
+- 实例 root 密码收敛：实例详情默认掩码、点击查看，串口控制台不再提供「复制 root 密码」
+- JWT 令牌版本机制：凭据 / TOTP / 外部身份 / 登录策略变更即失效全部旧令牌
+- HTTP 入口加固：显式读写超时与优雅关闭、请求体上限（面板 1MB / AI 网关 10MB）、登录字段长度上限、登录守卫条目数有界
+- 出错响应脱敏：内部错误只返回固定文案 + requestId，完整原因写服务端日志；GORM SQL 日志参数化输出，不再记录实参
+
+## [0.0.1]

 首个版本：自托管 OCI 多租户管理面板后端。

@@ -0,0 +1 @@
+v0.5.0
@@ -4,16 +4,21 @@

 # OCI Portal

-**Oracle Cloud Infrastructure 多租户管理面板**
+**自托管的 OCI 多租户管理面板与 GenAI 兼容网关**

-![Go](https://img.shields.io/badge/Go-1.26-00ADD8?logo=go&logoColor=white)
-![License](https://img.shields.io/badge/License-MIT-blue)
+![Release](https://img.shields.io/github/v/release/wangdefaa/oci-portal?display_name=tag)
+![Go](https://img.shields.io/badge/Go-1.26.5-00ADD8?logo=go&logoColor=white)
 ![Docker](https://img.shields.io/badge/Docker-amd64%20%7C%20arm64-2496ED?logo=docker&logoColor=white)
+![License](https://img.shields.io/badge/License-MIT-blue)

-本仓库为后端；前端工程见 [oci-portal-dash](https://github.com/wangdefaa/oci-portal-dash)（构建产物嵌入本服务成单文件）
+[快速开始](#快速开始) · [核心能力](#核心能力) · [生产部署](#生产部署) · [AI 网关](#ai-网关) · [开发](#开发)

 </div>

+OCI Portal 将多份 OCI API Key、云资源、自动化任务、审计事件和 OCI GenAI 渠道集中到一个管理界面。Vue 前端通过 `go:embed` 嵌入 Go 服务，Release 以单个二进制和多架构容器镜像交付。
+
+本仓库为后端与发行仓库；前端源码位于 [oci-portal-dash](https://github.com/wangdefaa/oci-portal-dash)。
+
 ## 界面预览

 | 总览 | 登录 |
@@ -28,78 +33,240 @@
 | --- | --- |
 | ![AI 网关](docs/assets/screenshot-ai-gateway.png) | ![通知设置](docs/assets/screenshot-settings-notify.png) |

-## 特性
+## 核心能力

- **多租户管理**：多份 OCI API Key 配置集中管理，私钥/口令 AES-256-GCM 加密落库；分组、批量测活、账户类型与订阅信息识别
- **资源管理**：实例（含创建/电源操作/更换公网 IP/IPv6/VNIC/引导卷）、VCN/子网/安全列表、块存储、限额与成本查询，多区域/多区间支持
- **抢机任务**：cron 周期尝试创建实例直到成功，支持熔断与通知
- **网页控制台**：实例串行控制台（xterm）与 VNC（noVNC），两跳 SSH 隧道
- **AI 网关**：OpenAI / Anthropic 兼容端点转发 OCI GenAI（对话/Responses/Embeddings），号池渠道加权负载均衡、熔断探测、密钥管理与用量日志
- **日志回传**：OCI Audit 事件经 Connector Hub → Notifications HTTPS 订阅回传入库，一键创建链路
- **租户治理**：IAM 用户/MFA/API Key 管理、密码策略、身份提供商（SAML）、通知收件人
- **安全**：JWT + bcrypt、TOTP 两步验证、OIDC/GitHub 外部登录、登录锁定、IP 限速、真实 IP 头可配、系统操作审计
- **通知**：Telegram 模板化推送（测活/抢机/成本/锁定等事件）
+- **租户与区域**：集中管理多份 OCI API Key，支持分组、批量测活、账户画像、订阅区域缓存与区域切换；私钥和口令使用 AES-256-GCM 加密落库
+- **计算、网络与存储**：实例创建与电源操作、公网 IP、IPv6、VNIC、串行控制台连接，VCN / 子网 / 安全列表，引导卷与块存储挂载，限额和成本查询
+- **自动化任务**：抢机、租户测活、成本同步、AI 渠道探测四类 cron 任务，提供执行日志、重叠执行防护、熔断与结果通知
+- **网页控制台**：浏览器内使用 xterm 串行终端和 noVNC，通过 OCI 控制台连接建立两跳 SSH 隧道
+- **身份与审计**：IAM 用户、MFA、API Key、密码策略、SAML 身份提供商、通知收件人和多 Identity Domain 管理；OCI Audit 事件可经 Service Connector Hub 与 Notifications 回传，关键事件按类别通过「云端事件」通知推送
+- **通知与安全**：Telegram、Webhook、ntfy、Bark、SMTP 五类渠道；JWT、bcrypt、TOTP、OIDC / GitHub 登录、登录锁定、IP 限速、会话撤销和系统操作审计
+- **AI 网关**：提供 OpenAI Responses、Chat Completions、Embeddings 与 Anthropic Messages 兼容接口，支持渠道分组、加权路由、熔断探测、模型黑白名单、密钥管理和调用日志
+
+## 运行形态
+
+```text
+浏览器 / API 客户端
+        │
+        ▼
+┌──────────────────── OCI Portal 单个 Go 进程 ────────────────────┐
+│  Vue 3 静态资源（go:embed）                                     │
+│  /api/v1  → Gin → Service → OCI SDK                            │
+│  /ai/v1   → 兼容转换与路由 → OCI Generative AI                 │
+│  GORM      → SQLite（默认）/ MySQL / PostgreSQL（experimental） │
+└─────────────────────────────────────────────────────────────────┘
+```
+
+默认推荐 SQLite 单实例部署。MySQL 和 PostgreSQL 适配仍属 experimental，不应据此推断服务支持多副本并发运行。

 ## 快速开始

+### Docker Compose（推荐）
+
+前置条件：Docker、Docker Compose v2、OpenSSL。
+
+1. 克隆仓库并生成一份需要长期保存的 `.env`：
+
+   ```bash
+   git clone https://github.com/wangdefaa/oci-portal.git
+   cd oci-portal
+
+   umask 077
+   printf 'DATA_KEY=%s\nJWT_SECRET=%s\nADMIN_PASSWORD=%s\n' \
+     "$(openssl rand -hex 32)" \
+     "$(openssl rand -hex 32)" \
+     "$(openssl rand -base64 24)" > .env
+   chmod 600 .env
+   ```
+
+2. 准备数据目录并启动：
+
+   ```bash
+   mkdir -p data
+
+   # Linux bind mount 需要让镜像内的 nonroot 用户（uid 65532）可写。
+   sudo chown 65532:65532 data
+
+   docker compose up -d
+   docker compose ps
+   ```
+
+3. 查看初始管理员密码并登录：
+
+   ```bash
+   grep '^ADMIN_PASSWORD=' .env
+   ```
+
+   访问 `http://127.0.0.1:18888`，默认用户名为 `admin`。
+
+> `DATA_KEY` 用于解密数据库中的 OCI 私钥、口令和渠道凭据。它只能生成一次并持续复用；丢失或更换后，已有密文无法恢复。请将 `.env` 与 `data/oci-portal.db` 一起备份。
+
 ### 二进制运行

-从 Release 下载对应架构的二进制后：
+Release 提供 Linux amd64 / arm64 二进制。以下以 amd64 为例；arm64 主机将文件名中的 `amd64` 替换为 `arm64`：

 ```bash
-DATA_KEY=$(openssl rand -hex 32) JWT_SECRET=$(openssl rand -hex 32) ADMIN_PASSWORD=<初始密码> ./oci-portal-server
+curl -fLO https://github.com/wangdefaa/oci-portal/releases/latest/download/oci-portal-server-linux-amd64
+chmod +x oci-portal-server-linux-amd64
+
+# 复用上文生成并妥善保存的 .env。
+set -a
+. ./.env
+set +a
+
+./oci-portal-server-linux-amd64
 ```

-访问 `http://localhost:8080`，用 `admin` 与初始密码登录。
+默认访问地址为 `http://localhost:8080`。`ADMIN_PASSWORD` 只在数据库中没有用户时创建初始管理员，后续启动不会用它重置密码。

-### Docker Compose
+### 源码构建
+
+源码构建要求 Go 1.26.5、`curl`、`unzip` 和 `sha256sum`。仓库只保留前端占位页，编译完整单文件前应下载 `DASH_VERSION` 指定的前端产物并校验：

 ```bash
-DATA_KEY=$(openssl rand -hex 32) JWT_SECRET=$(openssl rand -hex 32) ADMIN_PASSWORD=<初始密码> docker compose up -d
+DASH_TAG="$(tr -d '\r\n' < DASH_VERSION)"
+DASH_BASE="https://github.com/wangdefaa/oci-portal-dash/releases/download/${DASH_TAG}"
+
+curl -fL -o dist.zip "${DASH_BASE}/dist.zip"
+curl -fL -o dist.zip.sha256 "${DASH_BASE}/dist.zip.sha256"
+sha256sum -c dist.zip.sha256
+
+rm -rf internal/webui/dist
+mkdir -p internal/webui/dist
+unzip -q dist.zip -d internal/webui/dist
+
+CGO_ENABLED=0 go build -trimpath -o bin/oci-portal-server ./cmd/server
 ```

-### 源码构建（单文件，含前端）
+本地构建显示 `dev` 版本；Release 工作流会注入正式版本和构建时间。`internal/webui/dist` 中的真实前端产物不应提交到仓库。

-```bash
-# 1. 获取前端产物:从前端仓库 Release 下载 dist.zip(或本地 npm run build 后拷入)
-curl -fL -o dist.zip https://github.com/wangdefaa/oci-portal-dash/releases/latest/download/dist.zip
-rm -rf internal/webui/dist && mkdir -p internal/webui/dist && unzip -q dist.zip -d internal/webui/dist
-# 2. 编译(免 CGO,可交叉编译;-X 两项注入「设置·关于」页的版本与构建时间,可省略,省略则显示 dev)
-CGO_ENABLED=0 go build -trimpath \
-  -ldflags "-s -w -X oci-portal/internal/api.buildVersion=v0.0.1 -X oci-portal/internal/api.buildTime=$(date -u +%Y-%m-%dT%H:%M:%SZ)" \
-  -o bin/oci-portal-server ./cmd/server
+## 生产部署
+
+服务本身只提供 HTTP。除本机试用外，应保持服务仅监听回环地址或容器内部网络，并由 TLS 反向代理提供 HTTPS；否则管理员密码、JWT、AI 密钥和租户凭据会经明文连接传输。
+
+### Caddy
+
+```caddyfile
+portal.example.com {
+    reverse_proxy 127.0.0.1:18888
+}
 ```

-> 注意：上述解压会覆盖仓库占位文件 `internal/webui/dist/index.html`，提交代码前勿把真实产物带入版本库。
+Caddy 会自动处理 WebSocket。使用 nginx、Traefik 或其他反向代理时，还需要满足：

-## 环境变量
+- 为串行终端和 VNC 转发 WebSocket Upgrade 头
+- 为 AI 流式响应和控制台连接设置较长的读写超时，建议 `1h`
+- 整站请求体上限至少为 `10MB`；后端会继续限制面板 API 为 `1MB`、AI 网关为 `10MB`
+- 将面板内「设置 → 安全 → 真实 IP 请求头」配置为反向代理实际写入的头；系统审计、登录锁定和 IP 限速依赖它
+- 配置 `PUBLIC_URL` 或面板地址，供 OAuth 回调和 OCI 日志回传链路使用
+
+<details>
+<summary>nginx 最小示例</summary>
+
+```nginx
+map $http_upgrade $connection_upgrade {
+    default upgrade;
+    ""      close;
+}
+
+server {
+    listen 443 ssl;
+    server_name portal.example.com;
+
+    ssl_certificate     /etc/nginx/certs/portal.example.com.crt;
+    ssl_certificate_key /etc/nginx/certs/portal.example.com.key;
+
+    client_max_body_size 10m;
+
+    location / {
+        proxy_pass http://127.0.0.1:18888;
+        proxy_http_version 1.1;
+        proxy_set_header Upgrade $http_upgrade;
+        proxy_set_header Connection $connection_upgrade;
+        proxy_set_header Host $host;
+        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
+        proxy_read_timeout 1h;
+        proxy_send_timeout 1h;
+    }
+}
+```
+
+</details>
+
+若前端静态文件与后端分离部署，`/api/*` 和 `/ai/*` 必须代理到后端，其余路径由 SPA 静态服务处理并回退到 `index.html`。Traefik 通过容器网络连接时应直达容器端口 `8080`，无需暴露宿主机端口。
+
+## AI 网关
+
+面板内置 OpenAI / Anthropic 兼容的 GenAI 网关：独立密钥鉴权（`Authorization: Bearer sk-...` 或 `x-api-key`），支持渠道分组、加权路由、熔断探测、模型黑白名单、内容日志与调用日志。
+
+| 端点 | 定位 | 流式 |
+| --- | --- | --- |
+| `POST /ai/v1/responses` | OpenAI Responses，无状态主接口（xAI 服务端工具 / MCP） | SSE |
+| `POST /ai/v1/chat/completions` | OpenAI Chat Completions 兼容层 | SSE |
+| `POST /ai/v1/messages` | Anthropic Messages 转换层 | SSE |
+| `POST /ai/v1/embeddings` | OpenAI Embeddings | 否 |
+| `POST /ai/v1/audio/speech` | 文本转语音（xAI Voice） | 否 |
+| `POST /ai/v1/tts` | 文本转语音（xAI 官方格式） | 否 |
+| `POST /ai/v1/rerank` | 文档重排（Cohere Rerank） | 否 |
+| `POST /ai/v1/moderations` | 内容审核（OCI Guardrails） | 否 |
+| `GET /ai/v1/models` | 当前密钥可见的模型列表 | 否 |
+
+协议兼容边界、已知上游限制与逐字段兼容矩阵见 **[AI 网关文档](./docs/AI网关.md)**。
+
+## API 与配置
+
+### API 文档
+
+| 路径 | 鉴权 | 用途 |
+| --- | --- | --- |
+| `/api/v1/*` | 登录返回的 JWT Bearer Token | 面板管理 API |
+| `/ai/v1/*` | AI 网关密钥 | OpenAI / Anthropic 兼容接口 |
+| `/api/v1/webhooks/oci-logs/:secret` | URL 中的独立回传密钥 | OCI Notifications 日志回传 |
+
+OpenAPI 文件随仓库维护：[`docs/swagger.yaml`](docs/swagger.yaml) · [`docs/swagger.json`](docs/swagger.json)。
+
+运行进程时设置 `SWAGGER=1` 可开放 `/swagger/index.html`。Swagger 默认关闭，生产环境建议仅在受控网络内按需开启。接口注释变更后必须重新生成 OpenAPI。
+
+### 环境变量

 | 变量 | 必填 | 默认值 | 说明 |
-| --- | --- | --- | --- |
-| `DATA_KEY` | 是 | 无 | 敏感字段加密主密钥（更换后已入库密文无法解密） |
-| `JWT_SECRET` | 是 | 无 | 登录令牌签名密钥 |
+| --- | :---: | --- | --- |
+| `DATA_KEY` | 是 | — | 敏感字段加密主密钥；必须持久保存，不能随意轮换 |
+| `JWT_SECRET` | 是 | — | JWT 签名密钥；更换会使已有登录令牌失效 |
 | `ADMIN_USERNAME` | 否 | `admin` | 初始管理员用户名 |
-| `ADMIN_PASSWORD` | 首次启动是 | 无 | 仅在用户不存在时创建；已存在不重置 |
+| `ADMIN_PASSWORD` | 首次启动 | — | 仅在数据库无用户时创建管理员，不会重置已有密码 |
 | `ADDR` | 否 | `:8080` | HTTP 监听地址 |
-| `DB_DRIVER` | 否 | `sqlite` | `sqlite` / `mysql` / `postgres`（后两者 experimental） |
-| `DB_DSN` | 外部库时是 | 无 | MySQL 需 `parseTime=True`；PostgreSQL 标准 DSN |
-| `DB_PATH` | 否 | `oci-portal.db` | SQLite 文件路径 |
-| `PUBLIC_URL` | 否 | 无 | 面板公网基址，日志回传一键创建链路用 |
-| `HTTPS_PROXY` | 否 | 无 | 出站代理（如 Telegram 通知走代理） |
-| `TZ` | 否 | 系统 | cron 表达式解释时区（容器内建议显式设置，二进制已嵌 tzdata） |
-| `SWAGGER` | 否 | 关 | `1` 时开放 `/swagger/index.html` API 文档（生产建议按需临时开启） |
+| `DB_DRIVER` | 否 | `sqlite` | `sqlite` / `mysql` / `postgres`；后两者为 experimental |
+| `DB_PATH` | SQLite | `oci-portal.db` | SQLite 文件路径 |
+| `DB_DSN` | 外部数据库 | — | MySQL 需 `parseTime=True`；不要在日志或文档中暴露凭据 |
+| `PUBLIC_URL` | 否 | — | 面板公网基址，作为 OAuth 回调和日志回传引导的回退值 |
+| `TZ` | 否 | 系统时区 | cron 表达式的解释时区；容器示例使用 `Asia/Shanghai` |
+| `HTTP_PROXY` / `HTTPS_PROXY` / `NO_PROXY` | 否 | — | Go 标准出站代理变量；面板内显式代理配置优先用于对应业务 |
+| `SWAGGER` | 否 | 关闭 | 设为 `1` 时开放 Swagger UI |
+| `GIN_MODE` | 否 | `release` | `debug` / `release` |
+
+## 升级与备份
+
+- 升级前同时备份 `.env` 和数据库；SQLite Compose 部署的数据文件为 `data/oci-portal.db`
+- 保持 `DATA_KEY` 不变；只恢复数据库而没有原密钥，敏感字段无法解密
+- 服务启动时会自动执行数据库迁移；跨版本升级前先阅读 [CHANGELOG](CHANGELOG.md)
+- Compose 部署使用 `docker compose pull && docker compose up -d` 更新镜像
+- OCI API Key 应遵循最小权限原则；生产环境保持 Swagger 关闭并限制管理面访问来源

 ## 开发

 ```bash
-go test ./...            # 全量测试
-go vet ./... && gofmt -l .
-go tool swag init -g cmd/server/main.go -o docs --parseInternal --parseDependency   # 接口注释变更后重新生成 OpenAPI
+gofmt -l .
+go vet ./...
+go test ./...
+
+# Handler 注释变更后重新生成唯一的对外 API 文档。
+go tool swag init -g cmd/server/main.go -o docs --parseInternal --parseDependency --overridesFile docs/.swaggo
 ```

-API 文档：全部接口带 swaggo 注释，`SWAGGER=1` 启动后访问 `/swagger/index.html`；spec 文件在 `docs/swagger.json|yaml`。
-
-编码规范与项目约定见 [AGENTS.md](AGENTS.md) 与 [.trellis/spec/](.trellis/spec/)。
+- Go 后端规范与目录约定：[`AGENTS.md`](AGENTS.md) · [`.trellis/spec/backend/`](.trellis/spec/backend/)
+- 前端开发与构建：[oci-portal-dash](https://github.com/wangdefaa/oci-portal-dash)
+- 版本变更：[`CHANGELOG.md`](CHANGELOG.md)

 ## License

@@ -9,12 +9,15 @@ import (
 	"net"
 	"net/http"
 	"os"
+	"os/signal"
+	"syscall"
 	"time"

 	// 纯 Go 时区数据(+~450KB):distroless/scratch 容器无 tzdata,
 	// 嵌入后 TZ 环境变量即可让 cron 按本地时区解释
 	_ "time/tzdata"

+	"github.com/gin-gonic/gin"
 	"gorm.io/gorm"

 	// swag 生成的 OpenAPI spec(init 时注册,SWAGGER=1 的 /swagger 路由消费)
@@ -35,6 +38,12 @@ import (
 // @in header
 // @name Authorization
 // @description 登录接口返回的 JWT,格式: Bearer <token>
+func init() {
+	if os.Getenv("GIN_MODE") == "" {
+		gin.SetMode(gin.ReleaseMode)
+	}
+}
+
 func main() {
 	healthcheck := flag.Bool("healthcheck", false, "探活本机服务后退出:健康 0,异常 1(容器 HEALTHCHECK 用)")
 	flag.Parse()
@@ -126,6 +135,7 @@ func run() error {
 	defer aiGateway.Wait()
 	defer stopCleanup()
 	tasks := service.NewTaskService(db, ociConfigs, notifier, settings)
+	ociConfigs.SetTenantCleanupDeps(tasks)
 	tasks.AttachAiGateway(aiGateway)
 	aiGateway.SetOnChannelsChanged(tasks.SyncAiProbeTask)
 	if err := tasks.Start(); err != nil {
@@ -137,5 +147,36 @@ func run() error {
 	console := service.NewConsoleService(ociConfigs)
 	proxies := service.NewProxyService(db, cipher)
 	defer proxies.Wait()
-	return api.NewRouter(auth, oauth, ociConfigs, tasks, console, settings, notifier, systemLogs, logEvents, proxies, aiGateway).Run(cfg.Addr)
+	// 「关于」页存储指标需知数据库形态
+	api.SetAboutRuntime(cfg.DBDriver, cfg.DBPath)
+	router := api.NewRouter(auth, oauth, ociConfigs, tasks, console, settings, notifier, systemLogs, logEvents, proxies, aiGateway)
+	return serveHTTP(cfg.Addr, router)
+}
+
+// serveHTTP 以显式超时与请求头上限启动 HTTP 服务,并在 SIGINT/SIGTERM 时
+// 优雅关闭(10 秒排空);防慢连接(Slowloris)与超大头耗尽资源。
+// WriteTimeout 置 0:AI 网关流式响应无固定上界,WebSocket 为 hijack 连接
+// 本就不受 net/http 超时管控;读侧超时已覆盖慢请求风险。
+func serveHTTP(addr string, handler http.Handler) error {
+	srv := &http.Server{
+		Addr:              addr,
+		Handler:           handler,
+		ReadHeaderTimeout: 5 * time.Second,
+		ReadTimeout:       60 * time.Second,
+		IdleTimeout:       2 * time.Minute,
+		MaxHeaderBytes:    64 << 10,
+	}
+	ctx, stop := signal.NotifyContext(context.Background(), os.Interrupt, syscall.SIGTERM)
+	defer stop()
+	errCh := make(chan error, 1)
+	go func() { errCh <- srv.ListenAndServe() }()
+	select {
+	case err := <-errCh:
+		return err
+	case <-ctx.Done():
+		log.Println("[info] shutting down http server")
+		shutdownCtx, cancel := context.WithTimeout(context.Background(), 10*time.Second)
+		defer cancel()
+		return srv.Shutdown(shutdownCtx)
+	}
 }
@@ -1,9 +1,10 @@
 services:
  oci-portal:
    image: ghcr.io/wangdefaa/oci-portal:latest
+    container_name: oci-portal
    restart: unless-stopped
    ports:
-      - "18888:8080"
+      - "127.0.0.1:18888:8080"
    environment:
      DB_PATH: /data/oci-portal.db
      DATA_KEY: ${DATA_KEY:?DATA_KEY is required}
@@ -11,7 +12,4 @@ services:
      ADMIN_PASSWORD: ${ADMIN_PASSWORD:-}
      TZ: Asia/Shanghai
    volumes:
-      - oci-portal-data:/data
-
-volumes:
-  oci-portal-data:
+      - ./data:/data
@@ -0,0 +1,6 @@
+// swag 全局类型替换:原始 JSON 与联合类型统一渲染为 AnyJSON(任意 JSON 值),
+// 避免 json.RawMessage 被误渲染成 byte 数组、联合类型暴露内部字段。
+replace encoding/json.RawMessage oci-portal/internal/aiwire.AnyJSON
+replace oci-portal/internal/aiwire.RespInput oci-portal/internal/aiwire.AnyJSON
+replace oci-portal/internal/aiwire.Content oci-portal/internal/aiwire.AnyJSON
+replace oci-portal/internal/aiwire.StringList oci-portal/internal/aiwire.AnyJSON
@@ -0,0 +1,294 @@
+# AI 网关
+
+> 本文是 OCI Portal AI 网关的完整使用与兼容性文档：端点定位、协议兼容边界、已知上游限制与字段兼容矩阵。
+> 路由与鉴权的机器可读定义以 [Swagger YAML](swagger.yaml) 或运行时 Swagger UI 为准；无法由 OpenAPI 表达的兼容边界以本文为准。
+
+AI 网关使用面板创建的独立密钥鉴权，支持 `Authorization: Bearer sk-...` 和 `x-api-key: sk-...`。密钥可绑定渠道分组和模型白名单；全局模型黑名单会从模型列表、路由和探测候选中同时排除目标模型。
+
+| 端点 | 定位 | 流式 |
+| --- | --- | --- |
+| `POST /ai/v1/responses` | OpenAI Responses，无状态主接口 | SSE |
+| `POST /ai/v1/chat/completions` | OpenAI Chat Completions，存量客户端兼容层 | SSE |
+| `POST /ai/v1/messages` | Anthropic Messages 转换层 | SSE |
+| `POST /ai/v1/embeddings` | OpenAI Embeddings | 否 |
+| `POST /ai/v1/audio/speech` | OpenAI Audio Speech，文本转语音（xAI Voice） | 否 |
+| `POST /ai/v1/tts` | xAI 官方格式文本转语音（同一上游，网关转换） | 否 |
+| `POST /ai/v1/rerank` | 文档重排（Cohere Rerank，Jina 风格协议） | 否 |
+| `POST /ai/v1/moderations` | 内容审核（OCI Guardrails：内容审核 / PII / 提示注入） | 否 |
+| `GET /ai/v1/models` | 当前密钥可见的模型列表 | 否 |
+
+兼容边界：
+
+- 对话请求统一转发 OCI OpenAI 兼容面，当前供给以 `xai.`、`meta.`、`openai.` 前缀模型为主；Cohere Embeddings 不受该对话模型范围影响
+- 网关不保存会话历史，客户端需要携带完整上下文；Responses 拒绝非空 `previous_response_id`、非 `null` `conversation` 和 `background:true`
+- Responses 支持 Oracle 文档化的 xAI 服务端工具 `web_search` / `x_search` / `code_interpreter` 与远程 `mcp` 工具（非流式与流式均可），工具参数与限制遵循 [xAI 规格](https://docs.oracle.com/en-us/iaas/Content/generative-ai/get-started-agents.htm#xai-compatible-tools)；`code_interpreter` 的命名容器管理（containers API）与 File Search 不提供
+- Responses 的 `reasoning.effort`、Messages 的 `output_config.effort` 和 Chat Completions 的 `reasoning_effort` 会传给上游，实际档位和效果由模型决定
+- Chat Completions 只承担协议转换与兼容修复；新能力优先在 Responses 和 Messages 提供
+- 单次请求最多尝试三个渠道；可重试错误会切换渠道，流式响应建立后不会换渠道重试
+- Audio Speech 直通 OCI 兼容面（模型 `xai.grok-tts`，voice 取 xAI Grok Voice 列表：`ara`/`eve`/`leo`/`rex`/`sal`）；上游把 `language` 当必填，缺省时网关自动注入 `"auto"`，xAI 专属参数（`output_format` 等）可平铺在请求体透传；仅单请求返回音频，不提供 WebSocket 流式
+- `/ai/v1/tts` 为 xAI 官方 TTS 格式(`text`/`language` 必填、`voice_id`、`output_format` 对象)的转换端点:网关转换为 OpenAI 兼容形态后走同一上游与渠道调度,`model` 为网关扩展字段(缺省 `xai.grok-tts`);实测 `output_format`(codec/sample_rate/bit_rate)与 `speed` 透传生效
+- Rerank 走 OCI typed 面（`cohere.rerank-v4.0-pro` / `-fast`），请求 `{model, query, documents[], top_n?, return_documents?}`，响应 `results[].index` 指向入参下标；无 token 用量口径，调用日志只记时延
+- Moderations 是 OpenAI moderations 外壳映射 OCI Guardrails：`input` 为字符串或字符串数组（至多 8 条），categories 用 OCI 原生维度 `overall` / `blocklist` / `prompt_injection`（阈值 0.5 判定 `flagged`），PII 命中放扩展字段 `results[].pii`（不参与 flagged）；`model` 字段接受但忽略，无模型白名单维度；实测中文人名/手机号识别较弱，英文 PII 识别正常
+- Responses 的 `input_file` 内容块（file_url / file_data）实测被上游拒绝：`File content is currently unsupported for ZDR customers`——网关强制 `store:false` 属 ZDR 形态，该能力在上游侧不可用
+
+### 已知上游限制：大 system 区流式断流
+
+实测（2026-07-13）OCI 兼容面对 `instructions` 与 `tools` 合计超约 64.5KB 的**流式**请求会在发出少量事件后静默断开连接（无任何错误事件；同请求非流式正常），与模型、字符集、消息正文大小均无关——消息正文（`input`）不计入该限制。Chat Completions 与 Messages 的 system/developer 提示会转换为 `instructions`，因此 Claude Code 等自带大体量系统提示与工具定义的客户端极易触发。
+
+网关侧应对：
+
+- Messages 与 Chat Completions 的流式请求在客户端尚未收到任何输出时遭遇上游断流，会自动降级为非流式重做，并按标准事件/chunk 序列一次推送；调用日志记 `retries=1` 与降级标记
+- Responses 直通因初始事件已转发、协议上无法透明降级，调用日志记「上游流提前终止」，客户端需自行回退非流式
+- 应急规避：将超长 system 内容移入首条 user 消息正文可绕过该限制（正文不计入），但语义有别，根治有待上游修复
+
+这里提供的是兼容接口而非 OpenAI / Anthropic 协议的完整实现。OCI OpenAI 兼容面的部分行为来自实测，未见 Oracle 文档承诺，可能随上游调整。路由与鉴权定义以 [Swagger YAML](swagger.yaml) 或运行时 Swagger UI 为准；无法由 OpenAPI 完整表达的兼容边界列于上方。
+
+### 字段兼容矩阵
+
+以下矩阵以 2026-07-13 的 [OpenAI Responses](https://developers.openai.com/api/reference/resources/responses/methods/create)、[Chat Completions](https://developers.openai.com/api/reference/resources/chat/subresources/completions/methods/create)、[Embeddings](https://developers.openai.com/api/reference/resources/embeddings/methods/create)、[Audio Speech](https://developers.openai.com/api/reference/resources/audio/methods/speech)、[Moderations](https://developers.openai.com/api/reference/resources/moderations/methods/create) 和 [Anthropic Messages](https://platform.claude.com/docs/en/api/messages/create) 为标准基线，并与当前实现逐项核对；Rerank 无 OpenAI 对应端点，基线取 [Cohere Rerank](https://docs.cohere.com/reference/rerank) 协议。
+
+这是一份兼容性快照，不替代 Swagger。标准接口和 OCI 上游都可能变化，最终行为以当前版本代码与实测为准。
+
+矩阵只列网关支持的字段；不支持（本地拒绝）与被忽略的字段不入表，在各端点段落末尾以文字简述。
+
+| 标记 | 含义 |
+| :---: | --- |
+| ✅ | 网关直接支持 |
+| ➡️ | 网关原样透传；是否生效由 OCI 上游决定 |
+| 🔄 | 网关进行字段或协议转换后支持 |
+| ◐ | 部分支持、存在前置条件或语义降级 |
+
+<details>
+<summary><code>POST /ai/v1/responses</code> 对比 OpenAI Responses</summary>
+
+Responses 是“原始 JSON 直通 + 本地门禁”。除 `store` 外，网关不会重建请求体；未知顶层字段也会保留并送往 OCI。
+
+| 标准字段 | 状态 | 网关行为 |
+| --- | :---: | --- |
+| `model` | ◐ | 必须非空，还要通过密钥模型白名单并匹配可用渠道；随后原样透传 |
+| `input` | ➡️ | 支持标准的字符串或 item 数组，原始内容透传；网关不逐项保证 OCI 能处理所有 item 类型 |
+| `instructions`、`max_output_tokens` | ➡️ | 类型可解析后原样透传，不做范围或模型能力校验 |
+| `temperature`、`top_p`、`parallel_tool_calls` | ➡️ | 原样透传，不做取值范围校验 |
+| `text` / `text.format` / `text.verbosity` | ➡️ | 整个原始对象透传；结构化输出是否可用由 OCI 模型决定 |
+| `reasoning` | ➡️ | 整个原始对象透传；`effort` 不校验档位，`summary` 等字段不会被网关删除 |
+| `tool_choice` | ➡️ | 任意 JSON 原样透传，本地不校验枚举或结构 |
+| `store` | 🔄 | 无论客户端传什么，上游请求都强制改写为 `false` |
+| `stream` | ✅ | 普通请求、`function` 与服务端工具均支持 SSE |
+| `tools[].type=function` | ➡️ | 非流式和流式均可，工具对象原样透传 |
+| `tools[].type=web_search` / `x_search` / `code_interpreter` | ◐ | Oracle 文档化的 xAI 服务端工具，参数与限制遵循 xAI 规格（如 `allowed_domains` 上限 10、`container` 支持 `{"type":"auto"}`），非流式与流式均实测可用；`x_search` 不是 OpenAI 标准工具 |
+| `tools[].type=mcp` | ➡️ | 远程 MCP 服务由上游直连调用（`server_url` / `require_approval` / `authorization` 等原样透传），非流式与流式均实测可用 |
+| 其余标准与未知顶层字段 | ➡️ | `context_management`、`include`、`metadata`、`prompt`、`prompt_cache_key`、`service_tier`、`truncation`、`user` 等未建模字段一律保留在原始请求中，由 OCI 决定是否接受 |
+
+不支持（请求到达上游前返回 400）：非空 `previous_response_id`、非 `null` 的 `conversation`、`background:true`——网关无状态，不保存历史响应；以及 `function` / xAI 服务端工具 / `mcp` 之外的工具类型（`web_search_preview`、`file_search`、`computer`、`image_generation`、`shell`、`custom` 等）。
+
+响应边界：
+
+- 非流式成功响应不做转换，OCI JSON 原样返回；usage 解析只用于面板调用日志
+- SSE 事件逐行原样转发，不补 `data: [DONE]`，推理增量也不会被网关过滤
+- 流建立前最多切换三个渠道；流建立后中断不重试，客户端可能只收到部分事件
+- 未知模型返回 404、无渠道返回 503；上游错误会套入 OpenAI 风格错误体，不保证与标准 OpenAI 错误字段完全相同
+
+实现依据：[`airesponses.go`](../internal/service/airesponses.go) · [`responses.go`](../internal/aiwire/responses.go) · [`aigateway.go`](../internal/api/aigateway.go)
+
+</details>
+
+<details>
+<summary><code>POST /ai/v1/chat/completions</code> 对比 OpenAI Chat Completions</summary>
+
+Chat Completions 会先转换为 Responses 请求，再把 OCI Responses 响应桥接回 Chat Completions 形态。转换器只搬运下表字段，其余字段不进入上游请求。
+
+| 标准字段 | 状态 | 网关行为 |
+| --- | :---: | --- |
+| `model` | ◐ | 必填，受密钥白名单和可用渠道限制；模型名保留到上游请求 |
+| `messages` | 🔄 | 必填并转换为 Responses `input` / `instructions` |
+| `system` / `developer` 消息 | ◐ | 文本按出现顺序合并为 `instructions`；块数组中的非文本内容被忽略 |
+| `user` / `assistant` 文本内容 | 🔄 | 字符串及 `text` 块分别转为 `input_text` / `output_text` |
+| `image_url` 内容块 | ◐ | URL 或 data URI 转为 `input_image`；`image_url.detail` 被忽略 |
+| assistant `tool_calls` / `role=tool` | 🔄 | 转为 `function_call` / `function_call_output`，保留调用 ID、函数名和参数 |
+| `max_completion_tokens` | 🔄 | 转为 `max_output_tokens`，优先于 `max_tokens` |
+| `max_tokens` | 🔄 | 未提供 `max_completion_tokens` 时转为 `max_output_tokens` |
+| `temperature`、`top_p`、`parallel_tool_calls` | ◐ | 原值写入 Responses 请求，但不校验范围或模型能力 |
+| `stream` | 🔄 | OCI Responses SSE 桥接为 `chat.completion.chunk`，末尾补 `data: [DONE]`；上游断流且尚无输出时自动降级非流式重做，结果按 chunk 序列一次推送 |
+| `stream_options.include_usage` | 🔄 | 控制网关在终块后追加 `choices: []` 的 usage 块 |
+| `tools[].type=function` | ◐ | `name`、`description`、`parameters` 支持；`function.strict` 被忽略 |
+| `tool_choice` | ◐ | 支持 `auto` / `none` / `required` 和具名 function；非法或未知值被静默忽略 |
+| `response_format` | ◐ | `json_object`、`json_schema` 转为 Responses `text.format`；未知类型交给 OCI 处理 |
+| `reasoning_effort` | ◐ | 转小写后映射为 `reasoning.effort`，不校验模型或档位 |
+| `store` | 🔄 | 客户端取值被忽略，转换后的上游请求始终使用 `store:false` |
+
+不支持（返回 400）：`input_audio`、`file`、`refusal` 等消息内容块；`function` 以外的工具类型（含 `custom`）。
+
+静默忽略（转换后的上游请求不包含）：消息级 `name` / `refusal` / `audio` / 旧式 `function_call`；采样与输出控制类 `stop`、`seed`、`n`（恒返回单个 choice）、`frequency_penalty`、`presence_penalty`、`logprobs`、`top_logprobs`、`logit_bias`；平台类 `user`、`audio`、`modalities`、`prediction`、`metadata`、`moderation`、`prompt_cache_key`、`safety_identifier`、`service_tier`、`verbosity`、`web_search_options`、`stream_options.include_obfuscation` 及其他未知字段。
+
+响应边界：
+
+- 非流式固定生成一个 `choices[0]`；文本会合并，函数调用转为 `tool_calls`
+- `finish_reason` 只生成 `stop`、`tool_calls`、`length`；其他上游终止原因不保留
+- reasoning 输出、logprobs、refusal、annotations、audio、service tier 和 system fingerprint 不返回
+- usage 保留 `prompt_tokens`、`completion_tokens`、`total_tokens` 和 `prompt_tokens_details.cached_tokens`
+- 已建立的流中断或上游 `response.failed` 不会转换成标准 SSE 错误事件
+
+实现依据：[`chatresponses.go`](../internal/service/chatresponses.go) · [`openai.go`](../internal/aiwire/openai.go) · [`aigateway.go`](../internal/api/aigateway.go)
+
+</details>
+
+<details>
+<summary><code>POST /ai/v1/embeddings</code> 对比 OpenAI Embeddings</summary>
+
+| 标准字段 | 状态 | 网关行为 |
+| --- | :---: | --- |
+| `model` | ◐ | 必填，受密钥白名单限制，并且必须存在具有 `EMBEDDING` 能力的渠道 |
+| `input` 为字符串 | ✅ | 包装为单个输入后调用 OCI |
+| `input` 为字符串数组 | ✅ | 按原顺序调用 OCI；空字符串不在本地拒绝，由 OCI 决定 |
+| `dimensions` | ◐ | 映射为 OCI 输出维度，不做范围或模型能力校验 |
+| `encoding_format=float` | ✅ | 返回 float 数组；省略时行为相同 |
+
+不支持（返回 400）：token ID 数组（一维或二维）形态的 `input`、空数组、`null`，以及 `encoding_format=base64`。静默忽略：`user` 及其他未知字段。
+
+响应使用标准的 `object:"list"`、`data[].object:"embedding"`、`index`、`model` 和可选 `usage` 外壳；向量为 `float32` 数组，不支持流式。
+
+实现依据：[`embeddings.go`](../internal/aiwire/embeddings.go) · [`aigateway_chat.go`](../internal/service/aigateway_chat.go) · [`aigateway.go`](../internal/api/aigateway.go)
+
+</details>
+
+<details>
+<summary><code>POST /ai/v1/messages</code> 对比 Anthropic Messages</summary>
+
+网关接受 `Authorization: Bearer` 或 `x-api-key`，但不会校验或使用标准 Anthropic `anthropic-version`、`anthropic-beta` 请求头。
+
+| 标准字段 | 状态 | 网关行为 |
+| --- | :---: | --- |
+| `model` | ◐ | 用于模型与渠道选择，但空字符串不会在 handler 中按参数错误拒绝，通常最终返回模型不存在 |
+| `max_tokens` | 🔄 | 可缺省（缺省或 ≤0 时按默认值 8192），转换为 `max_output_tokens` |
+| `messages` | ◐ | 必须非空；角色、交替顺序和空内容不做完整校验 |
+| `system` | ◐ | 支持字符串或 text 块数组；多个文本块直接拼接，`cache_control` 等附加字段被忽略 |
+| `temperature`、`top_p` | ◐ | 写入 Responses 请求，不做取值范围或模型能力校验 |
+| `stream` | 🔄 | Responses SSE 桥接为 Anthropic 事件序列；上游断流且尚无输出时自动降级非流式重做，结果按事件序列一次推送 |
+| `tools` | ◐ | 每个工具都转换成 Responses `function`；自定义客户端工具可用，Anthropic 服务端工具类型不保留原语义 |
+| `tool_choice` | ◐ | 支持 `auto`、`any`、`none`、具名 `tool`；`disable_parallel_tool_use` 等附加字段被忽略 |
+| `output_config.effort` | 🔄 | 转小写后映射为 Responses `reasoning.effort` |
+
+顶层静默忽略（能解析但不传上游）：`top_k`、`stop_sequences`（响应 `stop_sequence` 恒为 `null`）、`metadata`、`thinking`（不控制上游思考预算）、`output_config` 其余子字段、`cache_control`、`container`、`inference_geo`、`service_tier` 及其他未知顶层字段。
+
+`messages[].content`：
+
+| 标准内容块 | 状态 | 网关行为 |
+| --- | :---: | --- |
+| 字符串 / `text` | 🔄 | user 转 `input_text`，assistant 历史转 `output_text` |
+| `image` | ◐ | 仅支持 `base64` 和 `url` source；缺字段或其他 source 类型返回 400 |
+| `tool_use` | 🔄 | 转为 `function_call`，保留 ID、名称和输入 |
+| `tool_result` | ◐ | 转为 `function_call_output`；块数组只拼接 text，`is_error` 和非文本结果丢失 |
+| `null` / 空块数组 | ◐ | 该消息可能从上游 input 中消失，不返回参数错误 |
+
+不支持的内容块（返回 400）：`document`、服务端工具结果及其他未知块。历史 `thinking` / `redacted_thinking` 块会被删除，不进入上游上下文。
+
+响应边界：
+
+- 非流式只把 Responses `output_text` 转成 `text`、`function_call` 转成 `tool_use`
+- `stop_reason` 只生成 `end_turn`、`tool_use`、`max_tokens`；`stop_sequence` 恒为 `null`
+- reasoning 不会生成 Anthropic `thinking` / `redacted_thinking` 块，也没有 signature
+- usage 只保留 `input_tokens`、`output_tokens` 和 `cache_read_input_tokens`，不提供 `cache_creation_input_tokens`
+- 流式输出标准事件骨架，但不生成 `thinking_delta` 和 `signature_delta`；上游错误事件与无终态断流会转成 Anthropic `error` 事件
+
+实现依据：[`anthresponses.go`](../internal/service/anthresponses.go) · [`anthropic.go`](../internal/aiwire/anthropic.go) · [`aigateway.go`](../internal/api/aigateway.go)
+
+</details>
+
+<details>
+<summary><code>POST /ai/v1/audio/speech</code> 对比 OpenAI Audio Speech</summary>
+
+Audio Speech 与 Responses 同为“原始 JSON 直通 + 本地门禁”：除缺省注入 `language` 外不重建请求体，未知字段原样透传。
+
+| 标准字段 | 状态 | 网关行为 |
+| --- | :---: | --- |
+| `model` | ◐ | 必填，受密钥白名单限制，并且必须存在具有 `TTS` 能力的渠道（当前上游仅 `xai.grok-tts`） |
+| `input` | ◐ | 必填非空，随后原样透传 |
+| `voice` | ➡️ | 原样透传；取 xAI Grok Voice 音色（`ara` / `eve` / `leo` / `rex` / `sal`），OpenAI 标准音色名不可用 |
+| `response_format` | ➡️ | 原样透传；实测 `mp3` 可用，其余格式由上游决定 |
+| `language`（扩展字段） | 🔄 | 上游必填；客户端缺省时网关自动注入 `"auto"` |
+| `speed`、`instructions` 及其他未知字段 | ➡️ | 原样透传（含 xAI 专属参数如 `output_format`），是否生效由上游决定 |
+
+不支持：流式音频（`stream_format` 等流式选项无效，响应恒为一次性完整音频）与 WebSocket 语音会话。
+
+响应边界：
+
+- 成功响应为音频字节，`Content-Type` 透传上游（缺省 `audio/mpeg`）
+- 无 token 用量口径，调用日志只记时延与渠道
+
+实现依据：[`genai_speech.go`](../internal/oci/genai_speech.go) · [`service/aigateway_extras.go`](../internal/service/aigateway_extras.go) · [`api/aigateway_extras.go`](../internal/api/aigateway_extras.go)
+
+</details>
+
+<details>
+<summary><code>POST /ai/v1/tts</code> 对比 xAI TTS</summary>
+
+`/ai/v1/tts` 接受 [xAI 官方 TTS 格式](https://docs.x.ai/developers/model-capabilities/audio/text-to-speech)(OCI 无 HTTP 版 xAI 面,网关转换为 OpenAI 兼容形态后与 `/ai/v1/audio/speech` 走同一上游与调度)。
+
+| 标准字段 | 状态 | 网关行为 |
+| --- | :---: | --- |
+| `text` | 🔄 | 必填非空,转换为上游 `input` |
+| `language` | ◐ | 必填(对齐 xAI 官方;接受 BCP-47 或 `auto`),原样透传 |
+| `voice_id` | 🔄 | 转换为上游 `voice`;缺省交上游默认(`eve`) |
+| `output_format` | ➡️ | `{codec, sample_rate, bit_rate}` 对象原样透传,实测生效(44.1kHz/192kbps 验证) |
+| `speed` | ➡️ | 原样透传,实测接受 |
+| `model`(网关扩展) | ◐ | xAI 官方无此字段;缺省注入 `xai.grok-tts`,可显式覆盖,受密钥白名单限制 |
+| `optimize_streaming_latency`、`text_normalization`、`with_timestamps` 及其他未知字段 | ➡️ | 原样透传,是否生效由上游决定 |
+
+不支持:流式音频输出(xAI 官方 `optimize_streaming_latency` 面向流式场景,本端点响应恒为一次性完整音频);WebSocket 流式(OCI 另有 `wss://…/xai/v1/tts` 私有协议面,网关未代理)。
+
+响应边界:
+
+- 成功响应为音频字节,`Content-Type` 透传上游(缺省 `audio/mpeg`)
+- 无 token 用量口径,调用日志只记时延与渠道(端点名 `tts`)
+
+实现依据:[`service/aigateway_extras.go`](../internal/service/aigateway_extras.go) · [`api/aigateway_extras.go`](../internal/api/aigateway_extras.go)
+
+</details>
+
+<details>
+<summary><code>POST /ai/v1/rerank</code> 对比 Cohere Rerank</summary>
+
+Rerank 无 OpenAI 对应端点，协议取 Jina / Cohere 通行风格，上游走 OCI typed 面（`cohere.rerank-v4.0-pro` / `-fast`）。
+
+| 字段 | 状态 | 网关行为 |
+| --- | :---: | --- |
+| `model` | ◐ | 必填，受密钥白名单限制，并且必须存在具有 `RERANK` 能力的渠道 |
+| `query` | ✅ | 必填非空 |
+| `documents` | ◐ | 必填，仅接受字符串数组；Cohere 旧版 `{"text": ...}` 对象数组形态返回 400 |
+| `top_n` | ✅ | 可选，传给上游限制返回条数；缺省返回全部文档的重排结果 |
+| `return_documents` | ✅ | 可选，`true` 时 `results[].document.text` 回带原文 |
+
+静默忽略：`max_tokens_per_doc` 等 Cohere 专属参数及其他未知字段。
+
+响应边界：
+
+- `results[]` 按相关度降序，`index` 指向入参 `documents` 下标，`relevance_score` 为 0–1 浮点；响应 `model` 回显请求值
+- 无 token 用量口径，调用日志只记时延与渠道
+
+实现依据：[`genai_guard.go`](../internal/oci/genai_guard.go) · [`rerank.go`](../internal/aiwire/rerank.go) · [`service/aigateway_extras.go`](../internal/service/aigateway_extras.go)
+
+</details>
+
+<details>
+<summary><code>POST /ai/v1/moderations</code> 对比 OpenAI Moderations</summary>
+
+Moderations 是 OpenAI moderations 外壳映射 OCI Guardrails（内容审核 / PII / 提示注入）。上游是服务级 API，没有模型与白名单维度；渠道按分组直接挑选。
+
+| 标准字段 | 状态 | 网关行为 |
+| --- | :---: | --- |
+| `input` 为字符串 | ✅ | 单条审核 |
+| `input` 为字符串数组 | ✅ | 逐条审核，单次 1~8 条 |
+
+不支持（返回 400）：多模态 input（图片等对象数组形态）、空字符串条目、空数组或超过 8 条的数组。静默忽略：`model`（接受任意值，不校验白名单）及其他未知字段。
+
+响应边界：
+
+- `categories` / `category_scores` 用 OCI 原生维度 `overall` / `blocklist` / `prompt_injection`，而非 OpenAI 标准类目（`hate` / `violence` 等）；任一维度得分 ≥ 0.5 判定 `flagged`
+- PII 命中放扩展字段 `results[].pii`（`text` / `label` / `score` / `offset` / `length`），不参与 `flagged` 判定；实测中文人名 / 手机号识别较弱，英文 PII 识别正常
+- 响应 `model` 恒为 `oci-guardrails`
+
+实现依据：[`genai_guard.go`](../internal/oci/genai_guard.go) · [`moderations.go`](../internal/aiwire/moderations.go) · [`service/aigateway_extras.go`](../internal/service/aigateway_extras.go)
+
+</details>
+
+`GET /ai/v1/models` 使用 OpenAI Models 列表外壳（`object`、`data[].id/object/created/owned_by`），但只返回当前渠道目录中通过分组、全局黑名单和密钥白名单筛选后的模型；网关不提供标准的单模型检索端点。面板「过滤弃用模型」开关开启时，OCI 已宣布弃用（即使未到退役日）的模型同时从模型列表与路由中排除，关闭后恢复。
@@ -20,6 +20,13 @@ type MessagesRequest struct {
 	ToolChoice    json.RawMessage `json:"tool_choice,omitempty"`
 	Metadata      json.RawMessage `json:"metadata,omitempty"`
 	Thinking      json.RawMessage `json:"thinking,omitempty"`
+	// OutputConfig 是 Anthropic 官方输出控制(effort 等);网关映射 effort 到 OCI reasoningEffort。
+	OutputConfig *AnthOutputConfig `json:"output_config,omitempty"`
+}
+
+// AnthOutputConfig 是 Anthropic Messages 顶层 output_config 对象。
+type AnthOutputConfig struct {
+	Effort string `json:"effort,omitempty"`
 }

 // SystemText 把顶层 system(string 或块数组)拍平为文本。
@@ -0,0 +1,5 @@
+package aiwire
+
+// AnyJSON 仅供 swagger 文档渲染:经 .swaggo 全局替换,代表任意 JSON 值
+// (json.RawMessage 与 string/数组联合类型),运行时代码不使用。
+type AnyJSON struct{}
@@ -0,0 +1,63 @@
+package aiwire
+
+import "encoding/json"
+
+// ModerationsRequest 是 /ai/v1/moderations 请求体;input 兼容 OpenAI 的
+// string 与 []string 两种形态,model 字段接受但忽略(上游为服务级 API)。
+type ModerationsRequest struct {
+	Model string          `json:"model,omitempty"`
+	Input json.RawMessage `json:"input"`
+}
+
+// ModerationPii 是一处 PII 命中(OCI 扩展,OpenAI 协议无对应物)。
+type ModerationPii struct {
+	Text   string  `json:"text"`
+	Label  string  `json:"label"`
+	Score  float64 `json:"score"`
+	Offset int     `json:"offset"`
+	Length int     `json:"length"`
+}
+
+// ModerationResult 是一条输入的审核结果:categories / category_scores 用
+// OCI 原生维度(overall / blocklist / prompt_injection),pii 为扩展字段。
+type ModerationResult struct {
+	Flagged        bool               `json:"flagged"`
+	Categories     map[string]bool    `json:"categories"`
+	CategoryScores map[string]float64 `json:"category_scores"`
+	Pii            []ModerationPii    `json:"pii,omitempty"`
+}
+
+// ModerationsResponse 是 /ai/v1/moderations 响应体(OpenAI moderations 外壳)。
+type ModerationsResponse struct {
+	ID      string             `json:"id"`
+	Model   string             `json:"model"`
+	Results []ModerationResult `json:"results"`
+}
+
+// SpeechRequest 描述 /ai/v1/audio/speech 请求体的已知字段(文档用途)。
+// 直通端点实际按原样透传,未列字段与 xAI 专属参数(如 output_format)同样保留。
+type SpeechRequest struct {
+	Model          string `json:"model"`
+	Input          string `json:"input"`
+	Voice          string `json:"voice,omitempty"`
+	ResponseFormat string `json:"response_format,omitempty"`
+	Language       string `json:"language,omitempty"`
+}
+
+// TtsRequest 描述 /ai/v1/tts 请求体的已知字段(xAI 官方 TTS 格式,文档用途)。
+// model 为网关扩展字段(缺省 xai.grok-tts);未列字段原样透传上游。
+type TtsRequest struct {
+	Model        string           `json:"model,omitempty"`
+	Text         string           `json:"text"`
+	Language     string           `json:"language"`
+	VoiceID      string           `json:"voice_id,omitempty"`
+	OutputFormat *TtsOutputFormat `json:"output_format,omitempty"`
+	Speed        float64          `json:"speed,omitempty"`
+}
+
+// TtsOutputFormat 是 xAI TTS 输出格式配置(缺省 MP3 24kHz/128kbps)。
+type TtsOutputFormat struct {
+	Codec      string `json:"codec,omitempty"`
+	SampleRate int    `json:"sample_rate,omitempty"`
+	BitRate    int    `json:"bit_rate,omitempty"`
+}
@@ -1,6 +1,6 @@
-// Package aiwire 定义 AI 网关的线格式类型:OpenAI Chat Completions(兼作网关内部
-// 中间表示 IR,因 OCI GENERIC 格式即其镜像)与 Anthropic Messages。零内部依赖,
-// oci 层与 service 层共同引用。
+// Package aiwire 定义 AI 网关的线格式类型:OpenAI Responses / Chat Completions /
+// Anthropic Messages / Embeddings 与通用记账、错误、模型列表结构。上游统一为
+// OCI OpenAI 兼容面。
 package aiwire

 import (
@@ -8,28 +8,6 @@ import (
 	"strings"
 )

-// ChatRequest 是 OpenAI /v1/chat/completions 请求体;TopK 为 OCI/Anthropic 扩展字段。
-type ChatRequest struct {
-	Model               string          `json:"model"`
-	Messages            []ChatMessage   `json:"messages"`
-	MaxTokens           *int            `json:"max_tokens,omitempty"`
-	MaxCompletionTokens *int            `json:"max_completion_tokens,omitempty"`
-	Temperature         *float64        `json:"temperature,omitempty"`
-	TopP                *float64        `json:"top_p,omitempty"`
-	TopK                *int            `json:"top_k,omitempty"`
-	FrequencyPenalty    *float64        `json:"frequency_penalty,omitempty"`
-	PresencePenalty     *float64        `json:"presence_penalty,omitempty"`
-	Stop                StringList      `json:"stop,omitempty"`
-	Seed                *int            `json:"seed,omitempty"`
-	N                   *int            `json:"n,omitempty"`
-	Stream              bool            `json:"stream,omitempty"`
-	StreamOptions       *StreamOptions  `json:"stream_options,omitempty"`
-	Tools               []Tool          `json:"tools,omitempty"`
-	ToolChoice          json.RawMessage `json:"tool_choice,omitempty"`
-	ResponseFormat      *ResponseFormat `json:"response_format,omitempty"`
-	ReasoningEffort     string          `json:"reasoning_effort,omitempty"`
-}
-
 // StringList 兼容 OpenAI stop 的 string 与 []string 两种线格式。
 type StringList []string

@@ -50,7 +28,75 @@ func (s *StringList) UnmarshalJSON(b []byte) error {
 	return nil
 }

-// StreamOptions 控制流式末尾是否附带 usage 事件。
+// Usage 是 token 用量。
+type Usage struct {
+	PromptTokens     int `json:"prompt_tokens"`
+	CompletionTokens int `json:"completion_tokens"`
+	TotalTokens      int `json:"total_tokens"`
+	// PromptTokensDetails 携带缓存命中细分;OCI 为隐式 prompt 缓存,只有读取计数
+	PromptTokensDetails *PromptTokensDetails `json:"prompt_tokens_details,omitempty"`
+}
+
+// PromptTokensDetails 是输入 token 细分(OpenAI prompt_tokens_details 同构)。
+type PromptTokensDetails struct {
+	CachedTokens int `json:"cached_tokens"`
+}
+
+// CachedTokens 返回缓存命中读取的 token 数;无细分时为 0。
+func (u *Usage) CachedTokens() int {
+	if u == nil || u.PromptTokensDetails == nil {
+		return 0
+	}
+	return u.PromptTokensDetails.CachedTokens
+}
+
+// Model 是 /v1/models 列表项。
+type Model struct {
+	ID      string `json:"id"`
+	Object  string `json:"object"`
+	Created int64  `json:"created"`
+	OwnedBy string `json:"owned_by"`
+}
+
+// ModelList 是 /v1/models 响应体。
+type ModelList struct {
+	Object string  `json:"object"`
+	Data   []Model `json:"data"`
+}
+
+// ErrorBody 是 OpenAI 格式错误响应。
+type ErrorBody struct {
+	Error ErrorDetail `json:"error"`
+}
+
+// ErrorDetail 是错误明细。
+type ErrorDetail struct {
+	Message string `json:"message"`
+	Type    string `json:"type"`
+	Code    string `json:"code,omitempty"`
+}
+
+// ---- Chat Completions 线格式(/ai/v1/chat/completions,Tier 2 兼容层)----
+
+// ChatRequest 是 OpenAI /v1/chat/completions 请求体;只声明网关读取的字段,
+// 无对应物的字段(stop / seed / n / penalty 等)由 JSON 解码天然忽略。
+type ChatRequest struct {
+	Model               string          `json:"model"`
+	Messages            []ChatMessage   `json:"messages"`
+	MaxTokens           *int            `json:"max_tokens,omitempty"`
+	MaxCompletionTokens *int            `json:"max_completion_tokens,omitempty"`
+	Temperature         *float64        `json:"temperature,omitempty"`
+	TopP                *float64        `json:"top_p,omitempty"`
+	ParallelToolCalls   *bool           `json:"parallel_tool_calls,omitempty"`
+	Stream              bool            `json:"stream,omitempty"`
+	StreamOptions       *StreamOptions  `json:"stream_options,omitempty"`
+	Tools               []Tool          `json:"tools,omitempty"`
+	ToolChoice          json.RawMessage `json:"tool_choice,omitempty"`
+	ResponseFormat      *ResponseFormat `json:"response_format,omitempty"`
+	ReasoningEffort     string          `json:"reasoning_effort,omitempty"`
+}
+
+// StreamOptions 控制流式末尾是否附带 usage 块。
 type StreamOptions struct {
 	IncludeUsage bool `json:"include_usage"`
 }
@@ -59,48 +105,31 @@ type StreamOptions struct {
 type ChatMessage struct {
 	Role       string     `json:"role"`
 	Content    Content    `json:"content"`
-	Name       string     `json:"name,omitempty"`
 	ToolCalls  []ToolCall `json:"tool_calls,omitempty"`
 	ToolCallID string     `json:"tool_call_id,omitempty"`
 }

-// Content 兼容 string 与多模态块数组两种线格式;一期仅透传文本。
+// Content 兼容 string 与多模态块数组两种线格式,序列化时保形。
 type Content struct {
-	// Text 在原始值为字符串时填充
 	Text    string
-	// Parts 在原始值为数组时填充
 	Parts   []ContentPart
-	// isArray 记录原始形态,序列化时保形
-	isArray bool
-}
-
-// ContentPart 是块数组中的一段;text 与 image_url 被网关处理,其余类型拒绝。
-type ContentPart struct {
-	Type     string    `json:"type"`
-	Text     string    `json:"text,omitempty"`
-	ImageURL *ImageURL `json:"image_url,omitempty"`
-}
-
-// ImageURL 是 image_url 块负载;url 为 data URI(base64)或公网地址,与 OCI ImageUrl 同构。
-type ImageURL struct {
-	URL    string `json:"url"`
-	Detail string `json:"detail,omitempty"`
+	IsArray bool
 }

 func (c *Content) UnmarshalJSON(b []byte) error {
-	trimmed := strings.TrimSpace(string(b))
-	if trimmed == "null" {
+	t := strings.TrimSpace(string(b))
+	if t == "null" {
 		return nil
 	}
-	if len(trimmed) > 0 && trimmed[0] == '[' {
-		c.isArray = true
+	if strings.HasPrefix(t, "[") {
+		c.IsArray = true
 		return json.Unmarshal(b, &c.Parts)
 	}
 	return json.Unmarshal(b, &c.Text)
 }

 func (c Content) MarshalJSON() ([]byte, error) {
-	if c.isArray {
+	if c.IsArray {
 		return json.Marshal(c.Parts)
 	}
 	return json.Marshal(c.Text)
@@ -111,14 +140,9 @@ func NewTextContent(text string) Content {
 	return Content{Text: text}
 }

-// NewPartsContent 构造块数组形态的内容(多模态消息用)。
-func NewPartsContent(parts []ContentPart) Content {
-	return Content{Parts: parts, isArray: true}
-}
-
-// JoinText 返回全部文本内容(数组形态拼接 text 块)。
+// JoinText 拍平内容为纯文本(数组形态拼接 text 块)。
 func (c Content) JoinText() string {
-	if !c.isArray {
+	if !c.IsArray {
 		return c.Text
 	}
 	var sb strings.Builder
@@ -130,23 +154,20 @@ func (c Content) JoinText() string {
 	return sb.String()
 }

-// HasUnsupported 报告是否含网关无法承接的内容块(text / image_url 之外的类型)。
-func (c Content) HasUnsupported() bool {
-	for _, p := range c.Parts {
-		switch p.Type {
-		case "", "text":
-		case "image_url":
-			if p.ImageURL == nil || p.ImageURL.URL == "" {
-				return true
-			}
-		default:
-			return true
-		}
-	}
-	return false
+// ContentPart 是块数组中的一段;text 与 image_url 被网关承接,其余类型拒绝。
+type ContentPart struct {
+	Type     string    `json:"type"`
+	Text     string    `json:"text,omitempty"`
+	ImageURL *ImageURL `json:"image_url,omitempty"`
 }

-// Tool 是 function 工具定义。
+// ImageURL 是 image_url 块负载;url 为 data URI(base64)或公网地址。
+type ImageURL struct {
+	URL    string `json:"url"`
+	Detail string `json:"detail,omitempty"`
+}
+
+// Tool 是嵌套形态的 function 工具定义。
 type Tool struct {
 	Type     string      `json:"type"`
 	Function FunctionDef `json:"function"`
@@ -178,7 +199,7 @@ type ResponseFormat struct {
 	JSONSchema json.RawMessage `json:"json_schema,omitempty"`
 }

-// ChatResponse 是非流式响应体。
+// ChatResponse 是非流式响应体(chat.completion)。
 type ChatResponse struct {
 	ID      string   `json:"id"`
 	Object  string   `json:"object"`
@@ -195,28 +216,6 @@ type Choice struct {
 	FinishReason string      `json:"finish_reason"`
 }

-// Usage 是 token 用量。
-type Usage struct {
-	PromptTokens     int `json:"prompt_tokens"`
-	CompletionTokens int `json:"completion_tokens"`
-	TotalTokens      int `json:"total_tokens"`
-	// PromptTokensDetails 携带缓存命中细分;OCI 为隐式 prompt 缓存,只有读取计数
-	PromptTokensDetails *PromptTokensDetails `json:"prompt_tokens_details,omitempty"`
-}
-
-// PromptTokensDetails 是输入 token 细分(OpenAI prompt_tokens_details 同构)。
-type PromptTokensDetails struct {
-	CachedTokens int `json:"cached_tokens"`
-}
-
-// CachedTokens 返回缓存命中读取的 token 数;无细分时为 0。
-func (u *Usage) CachedTokens() int {
-	if u == nil || u.PromptTokensDetails == nil {
-		return 0
-	}
-	return u.PromptTokensDetails.CachedTokens
-}
-
 // ChatChunk 是流式增量事件体(chat.completion.chunk)。
 type ChatChunk struct {
 	ID      string        `json:"id"`
@@ -254,29 +253,3 @@ type FunctionCallDelta struct {
 	Name      string `json:"name,omitempty"`
 	Arguments string `json:"arguments,omitempty"`
 }
-
-// Model 是 /v1/models 列表项。
-type Model struct {
-	ID      string `json:"id"`
-	Object  string `json:"object"`
-	Created int64  `json:"created"`
-	OwnedBy string `json:"owned_by"`
-}
-
-// ModelList 是 /v1/models 响应体。
-type ModelList struct {
-	Object string  `json:"object"`
-	Data   []Model `json:"data"`
-}
-
-// ErrorBody 是 OpenAI 格式错误响应。
-type ErrorBody struct {
-	Error ErrorDetail `json:"error"`
-}
-
-// ErrorDetail 是错误明细。
-type ErrorDetail struct {
-	Message string `json:"message"`
-	Type    string `json:"type"`
-	Code    string `json:"code,omitempty"`
-}
@@ -0,0 +1,28 @@
+package aiwire
+
+// RerankRequest 是 /ai/v1/rerank 请求体(Jina / Cohere 通行风格)。
+type RerankRequest struct {
+	Model           string   `json:"model"`
+	Query           string   `json:"query"`
+	Documents       []string `json:"documents"`
+	TopN            *int     `json:"top_n,omitempty"`
+	ReturnDocuments *bool    `json:"return_documents,omitempty"`
+}
+
+// RerankDocument 包装原文,仅在 return_documents 时返回。
+type RerankDocument struct {
+	Text string `json:"text"`
+}
+
+// RerankResult 是一条重排结果:index 指向请求 documents 下标。
+type RerankResult struct {
+	Index          int             `json:"index"`
+	RelevanceScore float64         `json:"relevance_score"`
+	Document       *RerankDocument `json:"document,omitempty"`
+}
+
+// RerankResponse 是 /ai/v1/rerank 响应体。
+type RerankResponse struct {
+	Model   string         `json:"model"`
+	Results []RerankResult `json:"results"`
+}
@@ -156,39 +156,6 @@ type RespReasoning struct {
 	Effort string `json:"effort,omitempty"`
 }

-// Response 是 /responses 响应对象。
-type Response struct {
-	ID                string          `json:"id"`
-	Object            string          `json:"object"`
-	CreatedAt         int64           `json:"created_at"`
-	Status            string          `json:"status"`
-	Model             string          `json:"model"`
-	Output            []RespOutItem   `json:"output"`
-	Usage             *RespUsage      `json:"usage,omitempty"`
-	Error             *RespError      `json:"error"`
-	IncompleteDetails *RespIncomplete `json:"incomplete_details"`
-	Store             bool            `json:"store"`
-}
-
-// RespOutItem 是 output 数组项:message 或 function_call。
-type RespOutItem struct {
-	Type      string        `json:"type"`
-	ID        string        `json:"id"`
-	Status    string        `json:"status,omitempty"`
-	Role      string        `json:"role,omitempty"`
-	Content   []RespOutPart `json:"content,omitempty"`
-	CallID    string        `json:"call_id,omitempty"`
-	Name      string        `json:"name,omitempty"`
-	Arguments string        `json:"arguments,omitempty"`
-}
-
-// RespOutPart 是 message item 的内容部件(output_text)。
-type RespOutPart struct {
-	Type        string `json:"type"`
-	Text        string `json:"text"`
-	Annotations []any  `json:"annotations"`
-}
-
 // RespUsage 是 Responses 口径的用量(input/output 命名)。
 type RespUsage struct {
 	InputTokens         int            `json:"input_tokens"`
@@ -217,3 +184,16 @@ type RespError struct {
 type RespIncomplete struct {
 	Reason string `json:"reason"`
 }
+
+// RespResponse 描述 /ai/v1/responses 非流式响应的常用顶层字段(文档用途)。
+// 直通端点原样返回上游 JSON,未列字段(reasoning、incomplete_details 等)同样保留。
+type RespResponse struct {
+	ID        string          `json:"id"`
+	Object    string          `json:"object"`
+	CreatedAt int64           `json:"created_at"`
+	Status    string          `json:"status"`
+	Model     string          `json:"model"`
+	Output    json.RawMessage `json:"output"`
+	Usage     *RespUsage      `json:"usage,omitempty"`
+	Error     *RespError      `json:"error,omitempty"`
+}
@@ -1,8 +1,11 @@
 package api

 import (
+	"math"
 	"net/http"
+	"os"
 	"runtime"
+	"time"

 	"github.com/gin-gonic/gin"
 )
@@ -12,13 +15,22 @@ import (
 var (
 	buildVersion = "dev"
 	buildTime    = ""
+	// startedAt 为进程启动时刻,运行时长与 CPU 均值的计算基准
+	startedAt = time.Now()
+	// aboutDB 由 main 启动时注入,存储指标据此定位数据库文件
+	aboutDB struct{ driver, path string }
 )

+// SetAboutRuntime 注入数据库形态,「关于」页存储指标展示用;启动时调用一次。
+func SetAboutRuntime(dbDriver, dbPath string) {
+	aboutDB.driver, aboutDB.path = dbDriver, dbPath
+}
+
 // about 返回构建与运行环境信息,「设置 · 关于」页展示。
 //
-// @Summary  返回构建与运行环境信息,「设置 · 关于」页展示
+// @Summary  返回构建、运行时长与资源占用信息,「设置 · 关于」页展示
 // @Tags     设置
-// @Success  200 {object} map[string]any
+// @Success  200 {object} aboutResponse
 // @Security BearerAuth
 // @Router   /api/v1/about [get]
 func about(c *gin.Context) {
@@ -27,5 +39,43 @@ func about(c *gin.Context) {
 		"buildTime":     buildTime,
 		"goVersion":     runtime.Version(),
 		"platform":      runtime.GOOS + "/" + runtime.GOARCH,
+		"startedAt":     startedAt.UTC().Format(time.RFC3339),
+		"uptimeSeconds": int64(time.Since(startedAt).Seconds()),
+		"resources":     resourcesSnapshot(),
 	})
 }
+
+// resourcesSnapshot 采集进程资源占用:CPU 自启动均值、内存、goroutine 与存储。
+func resourcesSnapshot() gin.H {
+	var ms runtime.MemStats
+	runtime.ReadMemStats(&ms)
+	pct := 0.0
+	if up := time.Since(startedAt).Seconds(); up > 0 {
+		// 多核并行时均值可超 100%,口径为「占单核百分比」
+		pct = processCPUTime().Seconds() / up * 100
+	}
+	return gin.H{
+		"cpuAvgPercent": math.Round(pct*100) / 100,
+		"numCpu":        runtime.NumCPU(),
+		"goroutines":    runtime.NumGoroutine(),
+		"memHeapBytes":  ms.HeapAlloc,
+		"memSysBytes":   ms.Sys,
+		"dbEngine":      aboutDB.driver,
+		"dbBytes":       dbFileBytes(aboutDB.driver, aboutDB.path),
+	}
+}
+
+// dbFileBytes 统计数据库磁盘占用:sqlite 累加主文件与 -wal / -shm,
+// 外部数据库(mysql / postgres)不可从本机度量,返回 -1 表示不可用。
+func dbFileBytes(driver, path string) int64 {
+	if driver != "sqlite" || path == "" {
+		return -1
+	}
+	var total int64
+	for _, p := range []string{path, path + "-wal", path + "-shm"} {
+		if fi, err := os.Stat(p); err == nil {
+			total += fi.Size()
+		}
+	}
+	return total
+}
@@ -0,0 +1,8 @@
+//go:build !unix
+
+package api
+
+import "time"
+
+// processCPUTime 非 unix 平台无 getrusage,CPU 均值退化为 0。
+func processCPUTime() time.Duration { return 0 }
@@ -0,0 +1,57 @@
+package api
+
+import (
+	"os"
+	"path/filepath"
+	"testing"
+)
+
+func TestDBFileBytes(t *testing.T) {
+	dir := t.TempDir()
+	dbPath := filepath.Join(dir, "app.db")
+	mustWrite := func(p string, n int) {
+		t.Helper()
+		if err := os.WriteFile(p, make([]byte, n), 0o600); err != nil {
+			t.Fatalf("write %s: %v", p, err)
+		}
+	}
+	mustWrite(dbPath, 100)
+	mustWrite(dbPath+"-wal", 30)
+
+	cases := []struct {
+		name   string
+		driver string
+		path   string
+		want   int64
+	}{
+		{name: "sqlite 主文件加 wal", driver: "sqlite", path: dbPath, want: 130},
+		{name: "sqlite 文件不存在计 0", driver: "sqlite", path: filepath.Join(dir, "none.db"), want: 0},
+		{name: "外部数据库不可度量", driver: "postgres", path: "", want: -1},
+		{name: "路径为空不可度量", driver: "sqlite", path: "", want: -1},
+	}
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			if got := dbFileBytes(tc.driver, tc.path); got != tc.want {
+				t.Fatalf("dbFileBytes(%q, %q) = %d, want %d", tc.driver, tc.path, got, tc.want)
+			}
+		})
+	}
+}
+
+func TestResourcesSnapshot(t *testing.T) {
+	SetAboutRuntime("sqlite", filepath.Join(t.TempDir(), "absent.db"))
+	got := resourcesSnapshot()
+
+	if got["numCpu"].(int) < 1 || got["goroutines"].(int) < 1 {
+		t.Fatalf("numCpu / goroutines 应为正数: %+v", got)
+	}
+	if got["memHeapBytes"].(uint64) == 0 || got["memSysBytes"].(uint64) == 0 {
+		t.Fatalf("内存指标应为正数: %+v", got)
+	}
+	if got["cpuAvgPercent"].(float64) < 0 {
+		t.Fatalf("cpuAvgPercent 不应为负: %+v", got)
+	}
+	if got["dbEngine"].(string) != "sqlite" || got["dbBytes"].(int64) != 0 {
+		t.Fatalf("存储指标不符: %+v", got)
+	}
+}
@@ -0,0 +1,18 @@
+//go:build unix
+
+package api
+
+import (
+	"syscall"
+	"time"
+)
+
+// processCPUTime 返回进程自启动累计的 CPU 时间(用户态 + 内核态);
+// 取不到时返回 0,「关于」页 CPU 均值显示为 0 而非报错。
+func processCPUTime() time.Duration {
+	var ru syscall.Rusage
+	if err := syscall.Getrusage(syscall.RUSAGE_SELF, &ru); err != nil {
+		return 0
+	}
+	return time.Duration(ru.Utime.Nano() + ru.Stime.Nano())
+}
@@ -6,6 +6,9 @@ import (

 	"github.com/gin-gonic/gin"

+	_ "oci-portal/internal/aiwire" // swagger 注解引用
+	_ "oci-portal/internal/model"  // swagger 注解引用
+
 	"oci-portal/internal/service"
 )

@@ -18,7 +21,7 @@ type aiAdminHandler struct {

 // @Summary  ---- 密钥 ----
 // @Tags     AI 管理
-// @Success  200 {object} map[string]any
+// @Success  200 {object} itemsResponse[model.AiKey]
 // @Security BearerAuth
 // @Router   /api/v1/ai-keys [get]
 func (h *aiAdminHandler) listKeys(c *gin.Context) {
@@ -35,7 +38,7 @@ func (h *aiAdminHandler) listKeys(c *gin.Context) {
 // @Summary  生成密钥
 // @Tags     AI 管理
 // @Param    body body object true "请求体(见接口说明)"
-// @Success  201 {object} map[string]any
+// @Success  201 {object} aiKeyCreateResponse "key 为明文密钥,仅本次返回"
 // @Security BearerAuth
 // @Router   /api/v1/ai-keys [post]
 func (h *aiAdminHandler) createKey(c *gin.Context) {
@@ -43,12 +46,13 @@ func (h *aiAdminHandler) createKey(c *gin.Context) {
 		Name   string   `json:"name" binding:"required"`
 		Value  string   `json:"value"`
 		Group  string   `json:"group"`
+		Models []string `json:"models"`
 	}
 	if err := c.ShouldBindJSON(&req); err != nil {
 		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
 		return
 	}
-	plaintext, key, err := h.gw.CreateKey(c.Request.Context(), req.Name, req.Value, req.Group)
+	plaintext, key, err := h.gw.CreateKey(c.Request.Context(), req.Name, req.Value, req.Group, req.Models)
 	if err != nil {
 		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
 		return
@@ -72,12 +76,13 @@ func (h *aiAdminHandler) updateKey(c *gin.Context) {
 		Name    string    `json:"name"`
 		Enabled *bool     `json:"enabled"`
 		Group   *string   `json:"group"`
+		Models  *[]string `json:"models"`
 	}
 	if err := c.ShouldBindJSON(&req); err != nil {
 		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
 		return
 	}
-	if err := h.gw.UpdateKey(c.Request.Context(), id, req.Name, req.Enabled, req.Group); err != nil {
+	if err := h.gw.UpdateKey(c.Request.Context(), id, req.Name, req.Enabled, req.Group, req.Models); err != nil {
 		respondError(c, err)
 		return
 	}
@@ -108,7 +113,7 @@ func (h *aiAdminHandler) deleteKey(c *gin.Context) {
 // @Tags     AI 管理
 // @Param    id path int true "配置 ID"
 // @Param    body body object true "请求体(见接口说明)"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} model.AiKey
 // @Security BearerAuth
 // @Router   /api/v1/ai-keys/{id}/content-log [put]
 func (h *aiAdminHandler) updateKeyContentLog(c *gin.Context) {
@@ -135,7 +140,7 @@ func (h *aiAdminHandler) updateKeyContentLog(c *gin.Context) {
 //
 // @Summary  分页查询内容日志
 // @Tags     AI 管理
-// @Success  200 {object} map[string]any
+// @Success  200 {object} pagedResponse[model.AiCallLog]
 // @Security BearerAuth
 // @Router   /api/v1/ai-content-logs [get]
 func (h *aiAdminHandler) listContentLogs(c *gin.Context) {
@@ -155,7 +160,7 @@ func (h *aiAdminHandler) listContentLogs(c *gin.Context) {

 // @Summary  ---- 渠道 ----
 // @Tags     AI 管理
-// @Success  200 {object} map[string]any
+// @Success  200 {object} itemsResponse[model.AiChannel]
 // @Security BearerAuth
 // @Router   /api/v1/ai-channels [get]
 func (h *aiAdminHandler) listChannels(c *gin.Context) {
@@ -170,7 +175,7 @@ func (h *aiAdminHandler) listChannels(c *gin.Context) {
 // @Summary  创建 AI 渠道
 // @Tags     AI 管理
 // @Param    body body service.ChannelInput true "请求体"
-// @Success  201 {object} map[string]any
+// @Success  201 {object} model.AiChannel
 // @Security BearerAuth
 // @Router   /api/v1/ai-channels [post]
 func (h *aiAdminHandler) createChannel(c *gin.Context) {
@@ -234,7 +239,7 @@ func (h *aiAdminHandler) deleteChannel(c *gin.Context) {
 // @Summary  触发探测:服务可见性 + 模型同步 + 配额试调,返回更新后的渠道
 // @Tags     AI 管理
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} model.AiChannel
 // @Security BearerAuth
 // @Router   /api/v1/ai-channels/{id}/probe [post]
 func (h *aiAdminHandler) probeChannel(c *gin.Context) {
@@ -253,7 +258,7 @@ func (h *aiAdminHandler) probeChannel(c *gin.Context) {
 // @Summary  同步渠道模型缓存
 // @Tags     AI 管理
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} itemsResponse[model.AiModelCache]
 // @Security BearerAuth
 // @Router   /api/v1/ai-channels/{id}/sync-models [post]
 func (h *aiAdminHandler) syncChannelModels(c *gin.Context) {
@@ -273,7 +278,7 @@ func (h *aiAdminHandler) syncChannelModels(c *gin.Context) {

 // @Summary  ---- 聚合模型与调用日志 ----
 // @Tags     AI 管理
-// @Success  200 {object} map[string]any
+// @Success  200 {object} itemsResponse[aiwire.Model]
 // @Security BearerAuth
 // @Router   /api/v1/ai-models [get]
 func (h *aiAdminHandler) gatewayModels(c *gin.Context) {
@@ -285,9 +290,67 @@ func (h *aiAdminHandler) gatewayModels(c *gin.Context) {
 	c.JSON(http.StatusOK, gin.H{"items": list.Data})
 }

+// ---- 模型黑名单 ----
+
+// @Summary  模型黑名单列表
+// @Tags     AI 管理
+// @Success  200 {object} itemsResponse[model.AiModelBlacklist]
+// @Security BearerAuth
+// @Router   /api/v1/ai-blacklist [get]
+func (h *aiAdminHandler) listBlacklist(c *gin.Context) {
+	items, err := h.gw.Blacklist(c.Request.Context())
+	if err != nil {
+		respondError(c, err)
+		return
+	}
+	c.JSON(http.StatusOK, gin.H{"items": items})
+}
+
+// addBlacklist 拉黑模型:删除全部渠道缓存中的同名条目,后续同步 / 探测均过滤。
+//
+// @Summary  添加模型黑名单
+// @Tags     AI 管理
+// @Param    body body object true "请求体:{name: 模型名}"
+// @Success  201 {object} itemResponse[model.AiModelBlacklist]
+// @Security BearerAuth
+// @Router   /api/v1/ai-blacklist [post]
+func (h *aiAdminHandler) addBlacklist(c *gin.Context) {
+	var req struct {
+		Name string `json:"name" binding:"required"`
+	}
+	if err := c.ShouldBindJSON(&req); err != nil {
+		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
+		return
+	}
+	item, err := h.gw.AddBlacklist(c.Request.Context(), req.Name)
+	if err != nil {
+		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
+		return
+	}
+	c.JSON(http.StatusCreated, gin.H{"item": item})
+}
+
+// @Summary  移除模型黑名单(重新同步后模型恢复入池)
+// @Tags     AI 管理
+// @Param    id path int true "黑名单条目 ID"
+// @Success  204 "无内容"
+// @Security BearerAuth
+// @Router   /api/v1/ai-blacklist/{id} [delete]
+func (h *aiAdminHandler) removeBlacklist(c *gin.Context) {
+	id, ok := aiPathID(c)
+	if !ok {
+		return
+	}
+	if err := h.gw.RemoveBlacklist(c.Request.Context(), id); err != nil {
+		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
+		return
+	}
+	c.Status(http.StatusNoContent)
+}
+
 // @Summary  AI 调用日志列表
 // @Tags     AI 管理
-// @Success  200 {object} map[string]any
+// @Success  200 {object} pagedResponse[model.AiContentLog]
 // @Security BearerAuth
 // @Router   /api/v1/ai-logs [get]
 func (h *aiAdminHandler) listLogs(c *gin.Context) {
@@ -309,3 +372,40 @@ func aiPathID(c *gin.Context) (uint, bool) {
 	}
 	return uint(id), true
 }
+
+// aiSettingsResponse 是 AI 网关全局设置(文档与响应共用)。
+type aiSettingsResponse struct {
+	FilterDeprecated bool `json:"filterDeprecated"`
+}
+
+// aiSettings 返回 AI 网关全局设置。
+//
+// @Summary  AI 网关全局设置
+// @Tags     AI 管理
+// @Success  200 {object} aiSettingsResponse
+// @Security BearerAuth
+// @Router   /api/v1/ai-settings [get]
+func (h *aiAdminHandler) aiSettings(c *gin.Context) {
+	c.JSON(http.StatusOK, aiSettingsResponse{FilterDeprecated: h.gw.FilterDeprecated()})
+}
+
+// updateAiSettings 更新 AI 网关全局设置(当前仅「过滤弃用模型」开关)。
+//
+// @Summary  更新 AI 网关全局设置
+// @Tags     AI 管理
+// @Param    body body aiSettingsResponse true "开启后已宣布弃用(即使未退役)的模型从列表与路由中排除"
+// @Success  200 {object} aiSettingsResponse
+// @Security BearerAuth
+// @Router   /api/v1/ai-settings [put]
+func (h *aiAdminHandler) updateAiSettings(c *gin.Context) {
+	var req aiSettingsResponse
+	if err := c.ShouldBindJSON(&req); err != nil {
+		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
+		return
+	}
+	if err := h.gw.SetFilterDeprecated(c.Request.Context(), req.FilterDeprecated); err != nil {
+		respondError(c, err)
+		return
+	}
+	c.JSON(http.StatusOK, aiSettingsResponse{FilterDeprecated: h.gw.FilterDeprecated()})
+}
@@ -1,13 +1,15 @@
 package api

 import (
+	"bufio"
+	"bytes"
 	"crypto/rand"
 	"encoding/hex"
 	"encoding/json"
 	"errors"
-	"fmt"
 	"io"
 	"net/http"
+	"slices"
 	"strings"
 	"time"

@@ -86,6 +88,25 @@ func keyGroup(c *gin.Context) string {
 	return ""
 }

+// keyModels 取当前请求密钥的模型白名单(空 = 不限模型)。
+func keyModels(c *gin.Context) []string {
+	if key, ok := c.Get(aiKeyCtx); ok {
+		return key.(*model.AiKey).Models
+	}
+	return nil
+}
+
+// checkKeyModel 校验请求模型是否在密钥白名单内;拒绝时按端点协议返回
+// 404 model_not_found(与未知模型同口径,不泄露密钥配置细节)并返回 false。
+func checkKeyModel(c *gin.Context, modelName string) bool {
+	allowed := keyModels(c)
+	if len(allowed) == 0 || slices.Contains(allowed, modelName) {
+		return true
+	}
+	aiError(c, http.StatusNotFound, "model_not_found", "模型不存在或无权访问: "+modelName)
+	return false
+}
+
 // logEntry 组装调用日志骨架;usage / status 由调用方补齐。
 func (h *aiGatewayHandler) logEntry(c *gin.Context, endpoint, modelName string, stream bool, meta service.ChatMeta, start time.Time) model.AiCallLog {
 	entry := model.AiCallLog{
@@ -102,21 +123,6 @@ func (h *aiGatewayHandler) logEntry(c *gin.Context, endpoint, modelName string,
 }

 // validateIR 做协议无关的入参检查(模型名、消息、不支持的内容块)。
-func validateIR(ir aiwire.ChatRequest) error {
-	if strings.TrimSpace(ir.Model) == "" {
-		return fmt.Errorf("model 不能为空")
-	}
-	if len(ir.Messages) == 0 {
-		return fmt.Errorf("messages 不能为空")
-	}
-	for _, m := range ir.Messages {
-		if m.Content.HasUnsupported() {
-			return service.ErrAiUnsupportedBlock
-		}
-	}
-	return nil
-}
-
 // maybeLogContent 在调用密钥显式开启内容日志且未过期时写正文(红线例外);
 // respBody 为 nil 时只记请求(流式响应与向量结果不记录);callLogID 关联同次调用日志。
 func (h *aiGatewayHandler) maybeLogContent(c *gin.Context, callLogID uint, endpoint, modelName string, stream bool, reqBody, respBody any) {
@@ -147,47 +153,6 @@ func (h *aiGatewayHandler) logFailure(c *gin.Context, entry model.AiCallLog, req
 	h.maybeLogContent(c, callID, entry.Endpoint, entry.Model, entry.Stream, reqBody, nil)
 }

-// chatCompletions 是 OpenAI /ai/v1/chat/completions 端点。
-//
-// @Summary  OpenAI 兼容对话补全
-// @Tags     AI 网关
-// @Param    body body object true "OpenAI chat/completions 请求体(支持 stream)"
-// @Success  200 {object} map[string]any "OpenAI 兼容响应(流式为 SSE)"
-// @Router   /ai/v1/chat/completions [post]
-func (h *aiGatewayHandler) chatCompletions(c *gin.Context) {
-	var ir aiwire.ChatRequest
-	if err := c.ShouldBindJSON(&ir); err != nil {
-		aiError(c, http.StatusBadRequest, "invalid_request_error", err.Error())
-		return
-	}
-	if err := validateIR(ir); err != nil {
-		aiError(c, http.StatusBadRequest, "invalid_request_error", err.Error())
-		return
-	}
-	if ir.Stream {
-		h.streamOpenAI(c, ir)
-		return
-	}
-	start := time.Now()
-	resp, meta, err := h.gw.Chat(c.Request.Context(), ir, keyGroup(c))
-	entry := h.logEntry(c, "openai", ir.Model, false, meta, start)
-	if err != nil {
-		upstreamError(c, err)
-		entry.ErrMsg = err.Error()
-		h.logFailure(c, entry, ir)
-		return
-	}
-	resp.ID = aiRandID("chatcmpl-")
-	if resp.Created == 0 {
-		resp.Created = time.Now().Unix()
-	}
-	entry.Status = http.StatusOK
-	fillUsage(&entry, resp.Usage)
-	callID := h.gw.LogCall(entry)
-	h.maybeLogContent(c, callID, "openai", ir.Model, false, ir, resp)
-	c.JSON(http.StatusOK, resp)
-}
-
 func fillUsage(entry *model.AiCallLog, u *aiwire.Usage) {
 	if u == nil {
 		return
@@ -200,8 +165,8 @@ func fillUsage(entry *model.AiCallLog, u *aiwire.Usage) {
 //
 // @Summary  OpenAI 兼容向量嵌入
 // @Tags     AI 网关
-// @Param    body body object true "OpenAI embeddings 请求体"
-// @Success  200 {object} map[string]any "OpenAI 兼容响应"
+// @Param    body body aiwire.EmbeddingsRequest true "OpenAI embeddings 请求体"
+// @Success  200 {object} aiwire.EmbeddingsResponse "OpenAI 兼容响应"
 // @Router   /ai/v1/embeddings [post]
 func (h *aiGatewayHandler) embeddings(c *gin.Context) {
 	var req aiwire.EmbeddingsRequest
@@ -217,6 +182,9 @@ func (h *aiGatewayHandler) embeddings(c *gin.Context) {
 		aiError(c, http.StatusBadRequest, "invalid_request_error", "仅支持 encoding_format=float")
 		return
 	}
+	if !checkKeyModel(c, req.Model) {
+		return
+	}
 	start := time.Now()
 	resp, meta, err := h.gw.Embeddings(c.Request.Context(), req, keyGroup(c))
 	entry := h.logEntry(c, "embeddings", req.Model, false, meta, start)
@@ -235,44 +203,6 @@ func (h *aiGatewayHandler) embeddings(c *gin.Context) {
 	c.JSON(http.StatusOK, resp)
 }

-// streamOpenAI 以 OpenAI SSE 直通流式响应,末尾发 [DONE]。
-func (h *aiGatewayHandler) streamOpenAI(c *gin.Context, ir aiwire.ChatRequest) {
-	start := time.Now()
-	stream, meta, err := h.gw.OpenStream(c.Request.Context(), ir, keyGroup(c))
-	entry := h.logEntry(c, "openai", ir.Model, true, meta, start)
-	if err != nil {
-		upstreamError(c, err)
-		entry.ErrMsg = err.Error()
-		h.logFailure(c, entry, ir)
-		return
-	}
-	defer stream.Close()
-	sseHeaders(c)
-	id, created := aiRandID("chatcmpl-"), time.Now().Unix()
-	var usage *aiwire.Usage
-	for {
-		chunk, err := stream.Next()
-		if err != nil {
-			if !errors.Is(err, io.EOF) {
-				entry.ErrMsg = err.Error()
-			}
-			break
-		}
-		chunk.ID, chunk.Created = id, created
-		if chunk.Usage != nil {
-			usage = chunk.Usage
-		}
-		writeSSEData(c, chunk)
-	}
-	c.Writer.WriteString("data: [DONE]\n\n")
-	c.Writer.Flush()
-	entry.Status = http.StatusOK
-	entry.LatencyMs = time.Since(start).Milliseconds()
-	fillUsage(&entry, usage)
-	callID := h.gw.LogCall(entry)
-	h.maybeLogContent(c, callID, "openai", ir.Model, true, ir, nil)
-}
-
 func sseHeaders(c *gin.Context) {
 	c.Header("Content-Type", "text/event-stream")
 	c.Header("Cache-Control", "no-cache")
@@ -280,23 +210,16 @@ func sseHeaders(c *gin.Context) {
 	c.Writer.Flush()
 }

-func writeSSEData(c *gin.Context, v any) {
-	b, err := json.Marshal(v)
-	if err != nil {
-		return
-	}
-	c.Writer.WriteString("data: ")
-	c.Writer.Write(b)
-	c.Writer.WriteString("\n\n")
-	c.Writer.Flush()
-}
+// anthDefaultMaxTokens 是 max_tokens 缺省时的默认输出上限:Anthropic 协议
+// 该字段必填,但部分客户端当选填不传,按默认值放行而非 400 拒绝。
+const anthDefaultMaxTokens = 8192

 // messages 是 Anthropic /ai/v1/messages 端点。
 //
 // @Summary  Anthropic Messages 兼容端点
 // @Tags     AI 网关
-// @Param    body body object true "Anthropic messages 请求体(支持 stream)"
-// @Success  200 {object} map[string]any "Anthropic 兼容响应(流式为 SSE)"
+// @Param    body body aiwire.MessagesRequest true "Anthropic messages 请求体(支持 stream;经 OCI OpenAI 兼容面直通;max_tokens 可缺省,默认 8192)"
+// @Success  200 {object} aiwire.MessagesResponse "Anthropic 兼容响应(非流式;流式为 SSE 事件序列)"
 // @Router   /ai/v1/messages [post]
 func (h *aiGatewayHandler) messages(c *gin.Context) {
 	var req aiwire.MessagesRequest
@@ -305,72 +228,139 @@ func (h *aiGatewayHandler) messages(c *gin.Context) {
 		return
 	}
 	if req.MaxTokens <= 0 {
-		aiError(c, http.StatusBadRequest, "invalid_request_error", "max_tokens 必填且需大于 0")
+		req.MaxTokens = anthDefaultMaxTokens
+	}
+	if len(req.Messages) == 0 {
+		aiError(c, http.StatusBadRequest, "invalid_request_error", "messages 不能为空")
 		return
 	}
-	ir, err := service.AnthropicToIR(req)
+	if !checkKeyModel(c, req.Model) {
+		return
+	}
+	body, err := service.AnthropicToResponsesBody(req)
 	if err != nil {
 		aiError(c, http.StatusBadRequest, "invalid_request_error", err.Error())
 		return
 	}
-	if err := validateIR(ir); err != nil {
-		aiError(c, http.StatusBadRequest, "invalid_request_error", err.Error())
-		return
-	}
 	if req.Stream {
-		h.streamAnthropic(c, ir)
+		h.streamAnthropic(c, body, req)
 		return
 	}
 	start := time.Now()
-	resp, meta, err := h.gw.Chat(c.Request.Context(), ir, keyGroup(c))
-	entry := h.logEntry(c, "anthropic", ir.Model, false, meta, start)
+	payload, meta, err := h.gw.RespPassthrough(c.Request.Context(), body, req.Model, keyGroup(c))
+	entry := h.logEntry(c, "anthropic", req.Model, false, meta, start)
 	if err != nil {
 		upstreamError(c, err)
 		entry.ErrMsg = err.Error()
 		h.logFailure(c, entry, req)
 		return
 	}
+	out, err := service.ResponsesToAnthropic(payload, aiRandID("msg_"))
+	if err != nil {
+		aiError(c, http.StatusBadGateway, "api_error", err.Error())
+		entry.ErrMsg = err.Error()
+		h.logFailure(c, entry, req)
+		return
+	}
 	entry.Status = http.StatusOK
-	fillUsage(&entry, resp.Usage)
+	fillUsage(&entry, service.RespPassthroughUsage(payload))
 	callID := h.gw.LogCall(entry)
-	out := service.IRRespToAnthropic(resp, aiRandID("msg_"))
-	h.maybeLogContent(c, callID, "anthropic", ir.Model, false, req, out)
+	h.maybeLogContent(c, callID, "anthropic", req.Model, false, req, out)
 	c.JSON(http.StatusOK, out)
 }

-// streamAnthropic 经状态机把 IR chunk 流聚合为 Anthropic SSE 事件流。
-func (h *aiGatewayHandler) streamAnthropic(c *gin.Context, ir aiwire.ChatRequest) {
+// streamAnthropic 流式直通并桥接:上游 Responses SSE 事件经 AnthRespBridge
+// 转为 Anthropic 事件序列逐块写出;usage 从 completed 事件记账。
+func (h *aiGatewayHandler) streamAnthropic(c *gin.Context, body []byte, req aiwire.MessagesRequest) {
 	start := time.Now()
-	stream, meta, err := h.gw.OpenStream(c.Request.Context(), ir, keyGroup(c))
-	entry := h.logEntry(c, "anthropic", ir.Model, true, meta, start)
+	upstream, meta, err := h.gw.RespPassthroughStream(c.Request.Context(), body, req.Model, keyGroup(c))
+	entry := h.logEntry(c, "anthropic", req.Model, true, meta, start)
 	if err != nil {
 		upstreamError(c, err)
 		entry.ErrMsg = err.Error()
-		h.logFailure(c, entry, ir)
+		h.logFailure(c, entry, req)
 		return
 	}
-	defer stream.Close()
+	defer upstream.Close()
 	sseHeaders(c)
-	st := service.NewAnthStream(aiRandID("msg_"), ir.Model)
-	for {
-		chunk, err := stream.Next()
-		if err != nil {
-			if !errors.Is(err, io.EOF) {
+	bridge := service.NewAnthRespBridge(aiRandID("msg_"), req.Model)
+	emitted := 0
+	if err := forwardSSEData(upstream, func(data []byte) {
+		evs := bridge.Feed(data)
+		emitted += len(evs)
+		writeAnthEvents(c, evs)
+	}); err != nil {
 		entry.ErrMsg = err.Error()
 	}
-			break
-		}
-		writeAnthEvents(c, st.Feed(chunk))
-	}
-	writeAnthEvents(c, st.Finish())
+	// 上游断流且客户端尚未收到任何事件(OCI 兼容面对大请求 + 推理模型的
+	// 流式通道会在 reasoning 阶段掐断):降级非流式重做,结果按事件序列推送
+	if emitted == 0 && !bridge.SawTerminal() && c.Request.Context().Err() == nil {
+		if h.anthFallback(c, &entry, req) {
 			entry.Status = http.StatusOK
 			entry.LatencyMs = time.Since(start).Milliseconds()
-	usage := st.Usage()
+			callID := h.gw.LogCall(entry)
+			h.maybeLogContent(c, callID, "anthropic", req.Model, true, req, nil)
+			return
+		}
+	}
+	writeAnthEvents(c, bridge.Finish())
+	// 上游错误事件 / 提前终止:SSE 头已发出维持 200,错误落日志可查
+	if msg := bridge.Err(); msg != "" && entry.ErrMsg == "" {
+		entry.ErrMsg = msg
+	}
+	entry.Status = http.StatusOK
+	entry.LatencyMs = time.Since(start).Milliseconds()
+	usage := bridge.Usage()
 	entry.PromptTokens, entry.CompletionTokens = usage.InputTokens, usage.OutputTokens
 	entry.TotalTokens = usage.InputTokens + usage.OutputTokens
 	entry.CachedTokens = usage.CacheReadInputTokens
 	callID := h.gw.LogCall(entry)
-	h.maybeLogContent(c, callID, "anthropic", ir.Model, true, ir, nil)
+	h.maybeLogContent(c, callID, "anthropic", req.Model, true, req, nil)
+}
+
+// anthFallback 用非流式重做同一请求并把完整结果按事件序列推送;
+// 成功返回 true 并把渠道 / 用量 / 降级标注写入日志条目。
+func (h *aiGatewayHandler) anthFallback(c *gin.Context, entry *model.AiCallLog, req aiwire.MessagesRequest) bool {
+	req.Stream = false
+	body, err := service.AnthropicToResponsesBody(req)
+	if err != nil {
+		return false
+	}
+	payload, meta, err := h.gw.RespPassthrough(c.Request.Context(), body, req.Model, keyGroup(c))
+	if err != nil {
+		return false
+	}
+	out, err := service.ResponsesToAnthropic(payload, aiRandID("msg_"))
+	if err != nil {
+		return false
+	}
+	writeAnthEvents(c, service.AnthMessageEvents(out))
+	entry.ChannelID, entry.ChannelName = meta.ChannelID, meta.ChannelName
+	entry.Retries++
+	entry.ErrMsg = "流式上游断流,已降级非流式完成"
+	fillUsage(entry, service.RespPassthroughUsage(payload))
+	return true
+}
+
+// forwardSSEData 逐行读上游 SSE,把 data 行交给 emit 即时转换写出;
+// Anthropic 与 Chat Completions 两条流式桥共用。
+func forwardSSEData(upstream io.Reader, emit func([]byte)) error {
+	reader := bufio.NewReader(upstream)
+	for {
+		line, err := reader.ReadBytes('\n')
+		if len(line) > 0 {
+			trimmed := bytes.TrimSpace(line)
+			if data, ok := bytes.CutPrefix(trimmed, []byte("data: ")); ok {
+				emit(data)
+			}
+		}
+		if err != nil {
+			if errors.Is(err, io.EOF) {
+				return nil
+			}
+			return err
+		}
+	}
 }

 func writeAnthEvents(c *gin.Context, events []service.AnthEvent) {
@@ -386,11 +376,159 @@ func writeAnthEvents(c *gin.Context, events []service.AnthEvent) {
 	c.Writer.Flush()
 }

+// chatCompletions 是 OpenAI /ai/v1/chat/completions 端点(Tier 2 兼容层,
+// 承接只会说 Chat Completions 的存量客户端)。
+//
+// @Summary  OpenAI Chat Completions 兼容端点
+// @Tags     AI 网关
+// @Param    body body aiwire.ChatRequest true "OpenAI chat/completions 请求体(支持 stream;经 Responses 转换直通)"
+// @Success  200 {object} aiwire.ChatResponse "OpenAI 兼容响应(非流式;流式为 SSE,末尾 data: [DONE])"
+// @Router   /ai/v1/chat/completions [post]
+func (h *aiGatewayHandler) chatCompletions(c *gin.Context) {
+	var req aiwire.ChatRequest
+	if err := c.ShouldBindJSON(&req); err != nil {
+		aiError(c, http.StatusBadRequest, "invalid_request_error", err.Error())
+		return
+	}
+	if strings.TrimSpace(req.Model) == "" || len(req.Messages) == 0 {
+		aiError(c, http.StatusBadRequest, "invalid_request_error", "model 与 messages 不能为空")
+		return
+	}
+	if !checkKeyModel(c, req.Model) {
+		return
+	}
+	body, err := service.ChatToResponsesBody(req)
+	if err != nil {
+		aiError(c, http.StatusBadRequest, "invalid_request_error", err.Error())
+		return
+	}
+	if req.Stream {
+		h.streamChat(c, body, req)
+		return
+	}
+	h.chatOnce(c, body, req)
+}
+
+// chatOnce 处理非流式:直通上游 → 转回 chat.completion → 记账。
+func (h *aiGatewayHandler) chatOnce(c *gin.Context, body []byte, req aiwire.ChatRequest) {
+	start := time.Now()
+	payload, meta, err := h.gw.RespPassthrough(c.Request.Context(), body, req.Model, keyGroup(c))
+	entry := h.logEntry(c, "openai", req.Model, false, meta, start)
+	if err != nil {
+		upstreamError(c, err)
+		entry.ErrMsg = err.Error()
+		h.logFailure(c, entry, req)
+		return
+	}
+	out, err := service.ResponsesToChat(payload, aiRandID("chatcmpl-"), time.Now().Unix())
+	if err != nil {
+		aiError(c, http.StatusBadGateway, "api_error", err.Error())
+		entry.ErrMsg = err.Error()
+		h.logFailure(c, entry, req)
+		return
+	}
+	entry.Status = http.StatusOK
+	fillUsage(&entry, out.Usage)
+	callID := h.gw.LogCall(entry)
+	h.maybeLogContent(c, callID, "openai", req.Model, false, req, out)
+	c.JSON(http.StatusOK, out)
+}
+
+// streamChat 流式直通并桥接:上游 Responses SSE 经 ChatRespBridge 转为
+// chat.completion.chunk 序列逐块写出,末尾发 [DONE];usage 从 completed 事件记账。
+func (h *aiGatewayHandler) streamChat(c *gin.Context, body []byte, req aiwire.ChatRequest) {
+	start := time.Now()
+	upstream, meta, err := h.gw.RespPassthroughStream(c.Request.Context(), body, req.Model, keyGroup(c))
+	entry := h.logEntry(c, "openai", req.Model, true, meta, start)
+	if err != nil {
+		upstreamError(c, err)
+		entry.ErrMsg = err.Error()
+		h.logFailure(c, entry, req)
+		return
+	}
+	defer upstream.Close()
+	sseHeaders(c)
+	includeUsage := req.StreamOptions != nil && req.StreamOptions.IncludeUsage
+	bridge := service.NewChatRespBridge(aiRandID("chatcmpl-"), req.Model, time.Now().Unix(), includeUsage)
+	emitted := 0
+	if err := forwardSSEData(upstream, func(data []byte) {
+		chunks := bridge.Feed(data)
+		emitted += len(chunks)
+		writeChatChunks(c, chunks)
+	}); err != nil {
+		entry.ErrMsg = err.Error()
+	}
+	// 上游断流且客户端尚未收到任何块:降级非流式重做(成因同 messages 端点)
+	if emitted == 0 && bridge.Usage() == nil && c.Request.Context().Err() == nil {
+		if h.chatFallback(c, &entry, req, includeUsage) {
+			entry.Status = http.StatusOK
+			entry.LatencyMs = time.Since(start).Milliseconds()
+			callID := h.gw.LogCall(entry)
+			h.maybeLogContent(c, callID, "openai", req.Model, true, req, nil)
+			return
+		}
+	}
+	writeChatChunks(c, bridge.Finish())
+	c.Writer.WriteString("data: [DONE]\n\n")
+	c.Writer.Flush()
+	// 未见 completed 即结束:usage 缺失说明上游流异常提前终止,落日志可查
+	if bridge.Usage() == nil && entry.ErrMsg == "" {
+		entry.ErrMsg = "上游流提前终止,未返回终态事件"
+	}
+	entry.Status = http.StatusOK
+	entry.LatencyMs = time.Since(start).Milliseconds()
+	fillUsage(&entry, bridge.Usage())
+	callID := h.gw.LogCall(entry)
+	h.maybeLogContent(c, callID, "openai", req.Model, true, req, nil)
+}
+
+// chatFallback 用非流式重做同一请求并把完整结果按 chunk 序列推送;
+// 成功返回 true 并把渠道 / 用量 / 降级标注写入日志条目。
+func (h *aiGatewayHandler) chatFallback(c *gin.Context, entry *model.AiCallLog, req aiwire.ChatRequest, includeUsage bool) bool {
+	req.Stream = false
+	body, err := service.ChatToResponsesBody(req)
+	if err != nil {
+		return false
+	}
+	payload, meta, err := h.gw.RespPassthrough(c.Request.Context(), body, req.Model, keyGroup(c))
+	if err != nil {
+		return false
+	}
+	out, err := service.ResponsesToChat(payload, aiRandID("chatcmpl-"), time.Now().Unix())
+	if err != nil {
+		return false
+	}
+	writeChatChunks(c, service.ChatResponseChunks(out, includeUsage))
+	c.Writer.WriteString("data: [DONE]\n\n")
+	c.Writer.Flush()
+	entry.ChannelID, entry.ChannelName = meta.ChannelID, meta.ChannelName
+	entry.Retries++
+	entry.ErrMsg = "流式上游断流,已降级非流式完成"
+	fillUsage(entry, service.RespPassthroughUsage(payload))
+	return true
+}
+
+// writeChatChunks 逐块写出 chunk 的 SSE data 行并 flush。
+func writeChatChunks(c *gin.Context, chunks []aiwire.ChatChunk) {
+	for _, ch := range chunks {
+		b, err := json.Marshal(ch)
+		if err != nil {
+			continue
+		}
+		c.Writer.WriteString("data: ")
+		c.Writer.Write(b)
+		c.Writer.WriteString("\n\n")
+	}
+	if len(chunks) > 0 {
+		c.Writer.Flush()
+	}
+}
+
 // listModels 是 /ai/v1/models 端点(OpenAI 格式,从启用渠道的模型缓存聚合)。
 //
 // @Summary  可用模型列表
 // @Tags     AI 网关
-// @Success  200 {object} map[string]any "OpenAI 兼容 models 列表"
+// @Success  200 {object} aiwire.ModelList "OpenAI 兼容 models 列表"
 // @Router   /ai/v1/models [get]
 func (h *aiGatewayHandler) listModels(c *gin.Context) {
 	list, err := h.gw.GatewayModels(c.Request.Context(), keyGroup(c))
@@ -398,6 +536,15 @@ func (h *aiGatewayHandler) listModels(c *gin.Context) {
 		aiError(c, http.StatusInternalServerError, "api_error", "查询模型列表失败")
 		return
 	}
+	if allowed := keyModels(c); len(allowed) > 0 {
+		kept := make([]aiwire.Model, 0, len(list.Data))
+		for _, m := range list.Data {
+			if slices.Contains(allowed, m.ID) {
+				kept = append(kept, m)
+			}
+		}
+		list.Data = kept
+	}
 	c.JSON(http.StatusOK, list)
 }

@@ -405,31 +552,45 @@ func (h *aiGatewayHandler) listModels(c *gin.Context) {
 //
 // @Summary  OpenAI Responses 兼容端点
 // @Tags     AI 网关
-// @Param    body body object true "OpenAI responses 请求体(支持 stream)"
-// @Success  200 {object} map[string]any "OpenAI 兼容响应(流式为 SSE)"
+// @Param    body body aiwire.RespRequest true "OpenAI responses 请求体(支持 stream;服务端工具 web_search/x_search/code_interpreter/mcp 含流式;未列字段原样透传上游)"
+// @Success  200 {object} aiwire.RespResponse "OpenAI 兼容响应(非流式;流式为 SSE);直通仅建模常用字段,未列字段原样返回"
 // @Router   /ai/v1/responses [post]
 func (h *aiGatewayHandler) responses(c *gin.Context) {
-	var req aiwire.RespRequest
-	if err := c.ShouldBindJSON(&req); err != nil {
-		aiError(c, http.StatusBadRequest, "invalid_request_error", err.Error())
-		return
-	}
-	ir, err := service.ResponsesToIR(req)
+	raw, err := c.GetRawData()
 	if err != nil {
 		aiError(c, http.StatusBadRequest, "invalid_request_error", err.Error())
 		return
 	}
-	if err := validateIR(ir); err != nil {
+	var req aiwire.RespRequest
+	if err := json.Unmarshal(raw, &req); err != nil {
+		aiError(c, http.StatusBadRequest, "invalid_request_error", err.Error())
+		return
+	}
+	if !checkKeyModel(c, req.Model) {
+		return
+	}
+	h.responsesPassthrough(c, raw, req)
+}
+
+// responsesPassthrough 直通链路(唯一上游):原始 body 改写后直达
+// OCI /actions/v1/responses,响应原样透传。
+func (h *aiGatewayHandler) responsesPassthrough(c *gin.Context, raw []byte, req aiwire.RespRequest) {
+	if err := service.RespPassthroughValidate(req); err != nil {
+		aiError(c, http.StatusBadRequest, "invalid_request_error", err.Error())
+		return
+	}
+	body, err := service.RespPassthroughBody(raw)
+	if err != nil {
 		aiError(c, http.StatusBadRequest, "invalid_request_error", err.Error())
 		return
 	}
 	if req.Stream {
-		h.streamResponses(c, ir)
+		h.responsesPassthroughStream(c, body, req)
 		return
 	}
 	start := time.Now()
-	resp, meta, err := h.gw.Chat(c.Request.Context(), ir, keyGroup(c))
-	entry := h.logEntry(c, "responses", ir.Model, false, meta, start)
+	payload, meta, err := h.gw.RespPassthrough(c.Request.Context(), body, req.Model, keyGroup(c))
+	entry := h.logEntry(c, "responses", req.Model, false, meta, start)
 	if err != nil {
 		upstreamError(c, err)
 		entry.ErrMsg = err.Error()
@@ -437,54 +598,70 @@ func (h *aiGatewayHandler) responses(c *gin.Context) {
 		return
 	}
 	entry.Status = http.StatusOK
-	fillUsage(&entry, resp.Usage)
+	fillUsage(&entry, service.RespPassthroughUsage(payload))
 	callID := h.gw.LogCall(entry)
-	out := service.IRRespToResponses(resp, aiRandID("resp_"), time.Now().Unix())
-	h.maybeLogContent(c, callID, "responses", ir.Model, false, req, out)
-	c.JSON(http.StatusOK, out)
+	h.maybeLogContent(c, callID, "responses", req.Model, false, req, json.RawMessage(payload))
+	c.Data(http.StatusOK, "application/json; charset=utf-8", payload)
 }

-// streamResponses 经状态机把 IR chunk 流聚合为 Responses 语义事件流。
-func (h *aiGatewayHandler) streamResponses(c *gin.Context, ir aiwire.ChatRequest) {
+// responsesPassthroughStream 流式直通:SSE 事件原样转发(推理增量等直达客户端),
+// 逐行扫描 completed 事件提取 usage 记账。
+func (h *aiGatewayHandler) responsesPassthroughStream(c *gin.Context, body []byte, req aiwire.RespRequest) {
 	start := time.Now()
-	stream, meta, err := h.gw.OpenStream(c.Request.Context(), ir, keyGroup(c))
-	entry := h.logEntry(c, "responses", ir.Model, true, meta, start)
+	upstream, meta, err := h.gw.RespPassthroughStream(c.Request.Context(), body, req.Model, keyGroup(c))
+	entry := h.logEntry(c, "responses", req.Model, true, meta, start)
 	if err != nil {
 		upstreamError(c, err)
 		entry.ErrMsg = err.Error()
-		h.logFailure(c, entry, ir)
+		h.logFailure(c, entry, req)
 		return
 	}
-	defer stream.Close()
+	defer upstream.Close()
 	sseHeaders(c)
-	st := service.NewRespStream(aiRandID("resp_"), ir.Model, time.Now().Unix())
-	for {
-		chunk, err := stream.Next()
+	usage, upErr, err := forwardSSE(c, upstream)
 	if err != nil {
-			if !errors.Is(err, io.EOF) {
 		entry.ErrMsg = err.Error()
+	} else if upErr != "" {
+		// 上游错误事件已原样转发给客户端,这里落日志可查
+		entry.ErrMsg = upErr
+	} else if usage == nil {
+		entry.ErrMsg = "上游流提前终止,未返回终态事件"
 	}
-			break
-		}
-		writeRespEvents(c, st.Feed(chunk))
-	}
-	writeRespEvents(c, st.Finish())
 	entry.Status = http.StatusOK
 	entry.LatencyMs = time.Since(start).Milliseconds()
-	fillUsage(&entry, st.Usage())
+	fillUsage(&entry, usage)
 	callID := h.gw.LogCall(entry)
-	h.maybeLogContent(c, callID, "responses", ir.Model, true, ir, nil)
+	h.maybeLogContent(c, callID, "responses", req.Model, true, req, nil)
 }

-func writeRespEvents(c *gin.Context, events []service.RespEvent) {
-	for _, ev := range events {
-		b, err := json.Marshal(ev.Data)
-		if err != nil {
-			continue
-		}
-		c.Writer.WriteString("event: " + ev.Event + "\ndata: ")
-		c.Writer.Write(b)
-		c.Writer.WriteString("\n\n")
-	}
+// forwardSSE 把上游 SSE 逐行转发给客户端,空行(事件边界)即 flush;
+// 顺带从 data 行提取 response.completed 的 usage 与错误事件消息。
+func forwardSSE(c *gin.Context, upstream io.Reader) (*aiwire.Usage, string, error) {
+	reader := bufio.NewReader(upstream)
+	var usage *aiwire.Usage
+	var upErr string
+	for {
+		line, err := reader.ReadBytes('\n')
+		if len(line) > 0 {
+			c.Writer.Write(line)
+			trimmed := bytes.TrimSpace(line)
+			if len(trimmed) == 0 {
 				c.Writer.Flush()
+			} else if data, ok := bytes.CutPrefix(trimmed, []byte("data: ")); ok {
+				if u := service.RespStreamCompletedUsage(data); u != nil {
+					usage = u
+				}
+				if m := service.RespStreamErrorMsg(data); m != "" && upErr == "" {
+					upErr = m
+				}
+			}
+		}
+		if err != nil {
+			c.Writer.Flush()
+			if errors.Is(err, io.EOF) {
+				return usage, upErr, nil
+			}
+			return usage, upErr, err
+		}
+	}
 }
@@ -0,0 +1,145 @@
+package api
+
+import (
+	"net/http"
+	"strings"
+	"time"
+
+	"github.com/gin-gonic/gin"
+
+	"oci-portal/internal/aiwire"
+	"oci-portal/internal/service"
+)
+
+// audioSpeech 是 OpenAI /ai/v1/audio/speech 端点(TTS,直通兼容面)。
+//
+// @Summary  OpenAI Audio Speech 兼容端点(文本转语音)
+// @Tags     AI 网关
+// @Param    body body aiwire.SpeechRequest true "OpenAI audio speech 请求体(model/input 必填,voice 见 xAI Grok Voice 列表,language 缺省 auto;未列字段原样透传)"
+// @Success  200 {file} binary "音频字节(Content-Type 透传上游,默认 audio/mpeg)"
+// @Router   /ai/v1/audio/speech [post]
+func (h *aiGatewayHandler) audioSpeech(c *gin.Context) {
+	raw, err := c.GetRawData()
+	if err != nil {
+		aiError(c, http.StatusBadRequest, "invalid_request_error", err.Error())
+		return
+	}
+	modelName, body, err := service.SpeechBodyNormalize(raw)
+	if err != nil {
+		aiError(c, http.StatusBadRequest, "invalid_request_error", err.Error())
+		return
+	}
+	h.speechRespond(c, "speech", modelName, raw, body)
+}
+
+// tts 是 xAI 官方格式 TTS 端点(/ai/v1/tts),转换为上游 OpenAI 兼容形态后复用 Speech 编排。
+//
+// @Summary  xAI 官方格式文本转语音端点
+// @Tags     AI 网关
+// @Param    body body aiwire.TtsRequest true "xAI TTS 请求体(text/language 必填,voice_id 缺省 eve;model 为网关扩展,缺省 xai.grok-tts;未列字段原样透传)"
+// @Success  200 {file} binary "音频字节(Content-Type 透传上游,默认 audio/mpeg)"
+// @Router   /ai/v1/tts [post]
+func (h *aiGatewayHandler) tts(c *gin.Context) {
+	raw, err := c.GetRawData()
+	if err != nil {
+		aiError(c, http.StatusBadRequest, "invalid_request_error", err.Error())
+		return
+	}
+	modelName, body, err := service.TtsBodyConvert(raw)
+	if err != nil {
+		aiError(c, http.StatusBadRequest, "invalid_request_error", err.Error())
+		return
+	}
+	h.speechRespond(c, "tts", modelName, raw, body)
+}
+
+// speechRespond 是 audioSpeech / tts 的公共主体:白名单校验、上游调用、日志与音频响应。
+func (h *aiGatewayHandler) speechRespond(c *gin.Context, endpoint, modelName string, raw, body []byte) {
+	if !checkKeyModel(c, modelName) {
+		return
+	}
+	start := time.Now()
+	audio, contentType, meta, err := h.gw.Speech(c.Request.Context(), modelName, body, keyGroup(c))
+	entry := h.logEntry(c, endpoint, modelName, false, meta, start)
+	if err != nil {
+		upstreamError(c, err)
+		entry.ErrMsg = err.Error()
+		h.logFailure(c, entry, string(raw))
+		return
+	}
+	entry.Status = http.StatusOK
+	callID := h.gw.LogCall(entry)
+	h.maybeLogContent(c, callID, endpoint, modelName, false, string(raw), nil)
+	if contentType == "" {
+		contentType = "audio/mpeg"
+	}
+	c.Data(http.StatusOK, contentType, audio)
+}
+
+// rerank 是 /ai/v1/rerank 端点(Jina / Cohere 风格文档重排)。
+//
+// @Summary  文档重排端点(Cohere Rerank)
+// @Tags     AI 网关
+// @Param    body body aiwire.RerankRequest true "重排请求体(model/query/documents 必填,可选 top_n/return_documents)"
+// @Success  200 {object} aiwire.RerankResponse "重排结果(results[].index 指向入参下标)"
+// @Router   /ai/v1/rerank [post]
+func (h *aiGatewayHandler) rerank(c *gin.Context) {
+	var req aiwire.RerankRequest
+	if err := c.ShouldBindJSON(&req); err != nil {
+		aiError(c, http.StatusBadRequest, "invalid_request_error", err.Error())
+		return
+	}
+	if strings.TrimSpace(req.Model) == "" || strings.TrimSpace(req.Query) == "" || len(req.Documents) == 0 {
+		aiError(c, http.StatusBadRequest, "invalid_request_error", "model、query 与 documents 不能为空")
+		return
+	}
+	if !checkKeyModel(c, req.Model) {
+		return
+	}
+	start := time.Now()
+	resp, meta, err := h.gw.Rerank(c.Request.Context(), req, keyGroup(c))
+	entry := h.logEntry(c, "rerank", req.Model, false, meta, start)
+	if err != nil {
+		upstreamError(c, err)
+		entry.ErrMsg = err.Error()
+		h.logFailure(c, entry, req)
+		return
+	}
+	entry.Status = http.StatusOK
+	callID := h.gw.LogCall(entry)
+	h.maybeLogContent(c, callID, "rerank", req.Model, false, req, resp)
+	c.JSON(http.StatusOK, resp)
+}
+
+// moderations 是 /ai/v1/moderations 端点(OpenAI 外壳映射 OCI Guardrails)。
+//
+// @Summary  内容审核端点(OCI Guardrails)
+// @Tags     AI 网关
+// @Param    body body aiwire.ModerationsRequest true "审核请求体(input 为字符串或字符串数组,单次至多 8 条;model 接受但忽略)"
+// @Success  200 {object} aiwire.ModerationsResponse "审核结果(categories/category_scores 为 overall/blocklist/prompt_injection,pii 为扩展字段)"
+// @Router   /ai/v1/moderations [post]
+func (h *aiGatewayHandler) moderations(c *gin.Context) {
+	var req aiwire.ModerationsRequest
+	if err := c.ShouldBindJSON(&req); err != nil {
+		aiError(c, http.StatusBadRequest, "invalid_request_error", err.Error())
+		return
+	}
+	inputs, err := service.ModerationInputs(req.Input)
+	if err != nil {
+		aiError(c, http.StatusBadRequest, "invalid_request_error", err.Error())
+		return
+	}
+	start := time.Now()
+	resp, meta, err := h.gw.Moderations(c.Request.Context(), aiRandID("modr_"), inputs, keyGroup(c))
+	entry := h.logEntry(c, "moderations", "oci-guardrails", false, meta, start)
+	if err != nil {
+		upstreamError(c, err)
+		entry.ErrMsg = err.Error()
+		h.logFailure(c, entry, req)
+		return
+	}
+	entry.Status = http.StatusOK
+	callID := h.gw.LogCall(entry)
+	h.maybeLogContent(c, callID, "moderations", "oci-guardrails", false, req, resp)
+	c.JSON(http.StatusOK, resp)
+}
@@ -0,0 +1,230 @@
+package api
+
+import (
+	"context"
+	"encoding/json"
+	"net/http"
+	"strconv"
+	"strings"
+	"testing"
+	"time"
+
+	"gorm.io/gorm"
+
+	"oci-portal/internal/model"
+)
+
+// seedGatewayModel 直插启用渠道与模型缓存,绕过云同步。
+func seedGatewayModel(t *testing.T, db *gorm.DB, name string) {
+	t.Helper()
+	ch := &model.AiChannel{Name: "t-" + name, OciConfigID: 1, Region: "r-" + name, Enabled: true, Priority: 1, Weight: 1}
+	if err := db.Create(ch).Error; err != nil {
+		t.Fatalf("seed channel: %v", err)
+	}
+	cache := &model.AiModelCache{ChannelID: ch.ID, ModelOcid: "ocid1.." + name, Name: name, Vendor: "v", SyncedAt: time.Now()}
+	if err := db.Create(cache).Error; err != nil {
+		t.Fatalf("seed model cache: %v", err)
+	}
+}
+
+func TestAiGatewayKeyModelRestrict(t *testing.T) {
+	r, auth, _, db := newTestRouterDB(t)
+	token, _, err := auth.Login(context.Background(), "admin", "pass123", "127.0.0.1", "")
+	if err != nil {
+		t.Fatalf("login: %v", err)
+	}
+
+	// 受限密钥:白名单含脏输入,入库应规范化为 2 项(cohere + ghost)
+	w := doRequest(t, r, http.MethodPost, "/api/v1/ai-keys", token,
+		`{"name":"limited","value":"limited-key-1234","models":[" cohere.command-a-03-2025 ","cohere.command-a-03-2025","","ghost-model"]}`)
+	if w.Code != http.StatusCreated {
+		t.Fatalf("创建受限密钥 status = %d, body %s", w.Code, w.Body.String())
+	}
+	var created struct {
+		Item model.AiKey `json:"item"`
+	}
+	if err := json.Unmarshal(w.Body.Bytes(), &created); err != nil {
+		t.Fatalf("decode: %v", err)
+	}
+	if len(created.Item.Models) != 2 || created.Item.Models[0] != "cohere.command-a-03-2025" {
+		t.Fatalf("创建后 models 未规范化: %v", created.Item.Models)
+	}
+	// 不限密钥对照
+	w = doRequest(t, r, http.MethodPost, "/api/v1/ai-keys", token, `{"name":"open","value":"open-key-12345"}`)
+	if w.Code != http.StatusCreated {
+		t.Fatalf("创建不限密钥 status = %d, body %s", w.Code, w.Body.String())
+	}
+
+	seedGatewayModel(t, db, "cohere.command-a-03-2025")
+	seedGatewayModel(t, db, "meta.llama-3.3-70b-instruct")
+
+	deny := []string{"model_not_found", "无权访问"}
+	pass := []string{"未知模型"} // 穿过白名单闸门,由编排层以「未知模型」拒绝(池内无 ghost-model)
+	tests := []struct {
+		name     string
+		key      string
+		path     string
+		body     string
+		wantCode int
+		wantSub  []string
+	}{
+		{"responses 流式同样拦截", "limited-key-1234", "/ai/v1/responses",
+			`{"model":"meta.llama-3.3-70b-instruct","stream":true,"input":"hi"}`, 404, deny},
+		{"responses 白名单内穿透", "limited-key-1234", "/ai/v1/responses",
+			`{"model":"ghost-model","input":"hi"}`, 404, pass},
+		{"embeddings 白名单外拦截", "limited-key-1234", "/ai/v1/embeddings",
+			`{"model":"meta.llama-3.3-70b-instruct","input":["hi"]}`, 404, deny},
+		{"messages 白名单外拦截为 Anthropic 体", "limited-key-1234", "/ai/v1/messages",
+			`{"model":"meta.llama-3.3-70b-instruct","max_tokens":16,"messages":[{"role":"user","content":"hi"}]}`, 404,
+			[]string{`"type":"error"`, "model_not_found", "无权访问"}},
+		{"responses 白名单外拦截", "limited-key-1234", "/ai/v1/responses",
+			`{"model":"meta.llama-3.3-70b-instruct","input":"hi"}`, 404, deny},
+		{"不限密钥穿透闸门", "open-key-12345", "/ai/v1/responses",
+			`{"model":"ghost-model","input":"hi"}`, 404, pass},
+		{"chat completions 白名单外拦截", "limited-key-1234", "/ai/v1/chat/completions",
+			`{"model":"meta.llama-3.3-70b-instruct","messages":[{"role":"user","content":"hi"}]}`, 404, deny},
+		{"chat completions 白名单内穿透", "limited-key-1234", "/ai/v1/chat/completions",
+			`{"model":"ghost-model","messages":[{"role":"user","content":"hi"}]}`, 404, pass},
+		{"messages 缺 max_tokens 按默认值放行", "open-key-12345", "/ai/v1/messages",
+			`{"model":"ghost-model","messages":[{"role":"user","content":"hi"}]}`, 404, pass},
+		{"chat completions 缺 messages 拒绝", "open-key-12345", "/ai/v1/chat/completions",
+			`{"model":"ghost-model"}`, 400, []string{"invalid_request_error"}},
+		{"chat completions 非 function 工具拒绝", "open-key-12345", "/ai/v1/chat/completions",
+			`{"model":"ghost-model","messages":[{"role":"user","content":"hi"}],"tools":[{"type":"web_search","function":{}}]}`,
+			400, []string{"仅支持 function"}},
+		{"chat completions 不支持内容块拒绝", "open-key-12345", "/ai/v1/chat/completions",
+			`{"model":"ghost-model","messages":[{"role":"user","content":[{"type":"input_audio"}]}]}`,
+			400, []string{"invalid_request_error"}},
+		{"audio speech 白名单外拦截", "limited-key-1234", "/ai/v1/audio/speech",
+			`{"model":"meta.llama-3.3-70b-instruct","input":"你好"}`, 404, deny},
+		{"rerank 白名单外拦截", "limited-key-1234", "/ai/v1/rerank",
+			`{"model":"meta.llama-3.3-70b-instruct","query":"q","documents":["d"]}`, 404, deny},
+		{"rerank 缺 documents 拒绝", "open-key-12345", "/ai/v1/rerank",
+			`{"model":"ghost-model","query":"q"}`, 400, []string{"invalid_request_error"}},
+		{"moderations 空 input 拒绝", "open-key-12345", "/ai/v1/moderations",
+			`{"input":[]}`, 400, []string{"invalid_request_error"}},
+		{"tts 缺 language 拒绝", "open-key-12345", "/ai/v1/tts",
+			`{"text":"你好"}`, 400, []string{"invalid_request_error"}},
+		{"tts 白名单外拦截", "limited-key-1234", "/ai/v1/tts",
+			`{"model":"meta.llama-3.3-70b-instruct","text":"你好","language":"zh"}`, 404, deny},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			w := doRequest(t, r, http.MethodPost, tt.path, tt.key, tt.body)
+			if w.Code != tt.wantCode {
+				t.Fatalf("status = %d, want %d, body %s", w.Code, tt.wantCode, w.Body.String())
+			}
+			for _, sub := range tt.wantSub {
+				if !strings.Contains(w.Body.String(), sub) {
+					t.Errorf("body 缺少 %q: %s", sub, w.Body.String())
+				}
+			}
+		})
+	}
+}
+
+func TestAiGatewayModelsFilteredByKey(t *testing.T) {
+	r, auth, _, db := newTestRouterDB(t)
+	token, _, err := auth.Login(context.Background(), "admin", "pass123", "127.0.0.1", "")
+	if err != nil {
+		t.Fatalf("login: %v", err)
+	}
+	seedGatewayModel(t, db, "cohere.command-a-03-2025")
+	seedGatewayModel(t, db, "meta.llama-3.3-70b-instruct")
+	w := doRequest(t, r, http.MethodPost, "/api/v1/ai-keys", token,
+		`{"name":"limited","value":"limited-key-1234","models":["cohere.command-a-03-2025"]}`)
+	if w.Code != http.StatusCreated {
+		t.Fatalf("创建密钥 status = %d", w.Code)
+	}
+	w = doRequest(t, r, http.MethodPost, "/api/v1/ai-keys", token, `{"name":"open","value":"open-key-12345"}`)
+	if w.Code != http.StatusCreated {
+		t.Fatalf("创建密钥 status = %d", w.Code)
+	}
+
+	tests := []struct {
+		name    string
+		key     string
+		want    []string
+		notWant []string
+	}{
+		{"受限密钥仅见交集", "limited-key-1234", []string{"cohere.command-a-03-2025"}, []string{"meta.llama-3.3-70b-instruct"}},
+		{"不限密钥全量可见", "open-key-12345", []string{"cohere.command-a-03-2025", "meta.llama-3.3-70b-instruct"}, nil},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			w := doRequest(t, r, http.MethodGet, "/ai/v1/models", tt.key, "")
+			if w.Code != http.StatusOK {
+				t.Fatalf("status = %d, body %s", w.Code, w.Body.String())
+			}
+			for _, sub := range tt.want {
+				if !strings.Contains(w.Body.String(), sub) {
+					t.Errorf("应包含 %q: %s", sub, w.Body.String())
+				}
+			}
+			for _, sub := range tt.notWant {
+				if strings.Contains(w.Body.String(), sub) {
+					t.Errorf("不应包含 %q: %s", sub, w.Body.String())
+				}
+			}
+		})
+	}
+}
+
+func TestAiKeyModelsUpdateRoundTrip(t *testing.T) {
+	r, auth, _, _ := newTestRouterDB(t)
+	token, _, err := auth.Login(context.Background(), "admin", "pass123", "127.0.0.1", "")
+	if err != nil {
+		t.Fatalf("login: %v", err)
+	}
+	w := doRequest(t, r, http.MethodPost, "/api/v1/ai-keys", token, `{"name":"k","value":"round-key-1234"}`)
+	if w.Code != http.StatusCreated {
+		t.Fatalf("创建 status = %d", w.Code)
+	}
+	var created struct {
+		Item model.AiKey `json:"item"`
+	}
+	_ = json.Unmarshal(w.Body.Bytes(), &created)
+	id := created.Item.ID
+
+	listModels := func() []string {
+		w := doRequest(t, r, http.MethodGet, "/api/v1/ai-keys", token, "")
+		var resp struct {
+			Items []model.AiKey `json:"items"`
+		}
+		_ = json.Unmarshal(w.Body.Bytes(), &resp)
+		for _, k := range resp.Items {
+			if k.ID == id {
+				return k.Models
+			}
+		}
+		t.Fatalf("密钥 %d 不在列表中", id)
+		return nil
+	}
+
+	tests := []struct {
+		name string
+		body string
+		want []string
+	}{
+		{"设置白名单", `{"models":[" a-model ","a-model","b-model"]}`, []string{"a-model", "b-model"}},
+		{"不传 models 保持不变", `{"name":"k2"}`, []string{"a-model", "b-model"}},
+		{"空数组清空恢复不限", `{"models":[]}`, nil},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			w := doRequest(t, r, http.MethodPut, "/api/v1/ai-keys/"+strconv.Itoa(int(id)), token, tt.body)
+			if w.Code != http.StatusNoContent {
+				t.Fatalf("update status = %d, body %s", w.Code, w.Body.String())
+			}
+			got := listModels()
+			if len(got) != len(tt.want) {
+				t.Fatalf("models = %v, want %v", got, tt.want)
+			}
+			for i := range tt.want {
+				if got[i] != tt.want[i] {
+					t.Fatalf("models = %v, want %v", got, tt.want)
+				}
+			}
+		})
+	}
+}
@@ -15,7 +15,7 @@ import (
 // @Param    id path int true "配置 ID"
 // @Param    instanceId path string true "instanceId"
 // @Param    body body object true "请求体(见接口说明)"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} publicIpResponse
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/instances/{instanceId}/change-public-ip [post]
 func (h *ociConfigHandler) changePublicIP(c *gin.Context) {
@@ -40,7 +40,7 @@ func (h *ociConfigHandler) changePublicIP(c *gin.Context) {
 // @Param    id path int true "配置 ID"
 // @Param    instanceId path string true "instanceId"
 // @Param    body body object true "请求体(见接口说明)"
-// @Success  201 {object} map[string]any
+// @Success  201 {object} ipv6AddressResponse
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/instances/{instanceId}/ipv6-addresses [post]
 func (h *ociConfigHandler) addInstanceIpv6(c *gin.Context) {
@@ -92,7 +92,7 @@ type attachVnicRequest struct {
 // @Tags     计算
 // @Param    id path int true "配置 ID"
 // @Param    instanceId path string true "instanceId"
-// @Success  200 {object} map[string]any
+// @Success  200 {array} oci.Vnic
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/instances/{instanceId}/vnics [get]
 func (h *ociConfigHandler) listInstanceVnics(c *gin.Context) {
@@ -113,7 +113,7 @@ func (h *ociConfigHandler) listInstanceVnics(c *gin.Context) {
 // @Param    id path int true "配置 ID"
 // @Param    instanceId path string true "instanceId"
 // @Param    body body attachVnicRequest true "请求体"
-// @Success  201 {object} map[string]any
+// @Success  201 {object} oci.Vnic
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/instances/{instanceId}/vnics [post]
 func (h *ociConfigHandler) attachVnic(c *gin.Context) {
@@ -160,7 +160,7 @@ func (h *ociConfigHandler) detachVnic(c *gin.Context) {
 // @Param    id path int true "配置 ID"
 // @Param    vnicId path string true "vnicId"
 // @Param    body body object true "请求体(见接口说明)"
-// @Success  201 {object} map[string]any
+// @Success  201 {object} addressResponse
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/vnics/{vnicId}/ipv6-addresses [post]
 func (h *ociConfigHandler) addVnicIpv6(c *gin.Context) {
@@ -195,7 +195,7 @@ type attachBootVolumeRequest struct {
 // @Tags     存储
 // @Param    id path int true "配置 ID"
 // @Param    instanceId path string true "instanceId"
-// @Success  200 {object} map[string]any
+// @Success  200 {array} oci.BootVolumeAttachment
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/instances/{instanceId}/boot-volume-attachments [get]
 func (h *ociConfigHandler) listBootVolumeAttachments(c *gin.Context) {
@@ -216,7 +216,7 @@ func (h *ociConfigHandler) listBootVolumeAttachments(c *gin.Context) {
 // @Param    id path int true "配置 ID"
 // @Param    instanceId path string true "instanceId"
 // @Param    body body attachBootVolumeRequest true "请求体"
-// @Success  201 {object} map[string]any
+// @Success  201 {object} oci.BootVolumeAttachment
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/instances/{instanceId}/boot-volume-attachments [post]
 func (h *ociConfigHandler) attachBootVolume(c *gin.Context) {
@@ -242,7 +242,7 @@ func (h *ociConfigHandler) attachBootVolume(c *gin.Context) {
 // @Param    id path int true "配置 ID"
 // @Param    instanceId path string true "instanceId"
 // @Param    body body attachBootVolumeRequest true "请求体"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} oci.BootVolumeAttachment
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/instances/{instanceId}/replace-boot-volume [post]
 func (h *ociConfigHandler) replaceBootVolume(c *gin.Context) {
@@ -293,7 +293,7 @@ type attachVolumeRequest struct {
 // @Summary  块存储卷列表
 // @Tags     存储
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {array} oci.BlockVolume
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/volumes [get]
 func (h *ociConfigHandler) listBlockVolumes(c *gin.Context) {
@@ -313,7 +313,7 @@ func (h *ociConfigHandler) listBlockVolumes(c *gin.Context) {
 // @Tags     存储
 // @Param    id path int true "配置 ID"
 // @Param    instanceId path string true "instanceId"
-// @Success  200 {object} map[string]any
+// @Success  200 {array} oci.VolumeAttachment
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/instances/{instanceId}/volume-attachments [get]
 func (h *ociConfigHandler) listVolumeAttachments(c *gin.Context) {
@@ -334,7 +334,7 @@ func (h *ociConfigHandler) listVolumeAttachments(c *gin.Context) {
 // @Param    id path int true "配置 ID"
 // @Param    instanceId path string true "instanceId"
 // @Param    body body attachVolumeRequest true "请求体"
-// @Success  201 {object} map[string]any
+// @Success  201 {object} oci.VolumeAttachment
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/instances/{instanceId}/volume-attachments [post]
 func (h *ociConfigHandler) attachVolume(c *gin.Context) {
@@ -18,10 +18,11 @@ type authHandler struct {
 }

 type loginRequest struct {
-	Username string `json:"username" binding:"required"`
-	Password string `json:"password" binding:"required"`
+	// 字段长度上限防超长输入撑爆登录守卫与 bcrypt(S-03)
+	Username string `json:"username" binding:"required,max=64"`
+	Password string `json:"password" binding:"required,max=128"`
 	// 两步验证码;账号已启用 TOTP 时必填,缺失返回 428
-	TotpCode string `json:"totpCode"`
+	TotpCode string `json:"totpCode" binding:"max=8"`
 }

 // login 校验用户名密码(与可选 TOTP)后签发 JWT。
@@ -31,9 +32,9 @@ type loginRequest struct {
 // @Accept   json
 // @Produce  json
 // @Param    body body loginRequest true "登录凭据"
-// @Success  200 {object} map[string]any "token 与 expiresAt"
-// @Failure  401 {object} map[string]string "凭据错误"
-// @Failure  428 {object} map[string]any "需要两步验证码(totpRequired=true)"
+// @Success  200 {object} tokenResponse "token 与 expiresAt"
+// @Failure  401 {object} errorResponse "凭据错误"
+// @Failure  428 {object} totpRequiredResponse "需要两步验证码(totpRequired=true)"
 // @Router   /api/v1/auth/login [post]
 func (h *authHandler) login(c *gin.Context) {
 	var req loginRequest
@@ -64,7 +65,7 @@ func (h *authHandler) login(c *gin.Context) {
 		return
 	}
 	if err != nil {
-		c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
+		respondError(c, err)
 		return
 	}
 	h.recordLogin(c, req.Username, http.StatusOK, start)
@@ -8,6 +8,8 @@ import (

 	"github.com/gin-gonic/gin"

+	_ "oci-portal/internal/model" // swagger 注解引用
+
 	"oci-portal/internal/service"
 )

@@ -23,7 +25,7 @@ type authxHandler struct {
 //
 // @Summary  两步验证状态
 // @Tags     认证
-// @Success  200 {object} map[string]bool "enabled"
+// @Success  200 {object} enabledResponse "enabled"
 // @Security BearerAuth
 // @Router   /api/v1/auth/totp [get]
 func (h *authxHandler) totpStatus(c *gin.Context) {
@@ -39,8 +41,8 @@ func (h *authxHandler) totpStatus(c *gin.Context) {
 //
 // @Summary  发起两步验证设置
 // @Tags     认证
-// @Success  200 {object} map[string]string "secret 与 otpauthUri"
-// @Failure  409 {object} map[string]string "已启用"
+// @Success  200 {object} totpSetupResponse "secret 与 otpauthUri"
+// @Failure  409 {object} errorResponse "已启用"
 // @Security BearerAuth
 // @Router   /api/v1/auth/totp/setup [post]
 func (h *authxHandler) totpSetup(c *gin.Context) {
@@ -81,7 +83,7 @@ func (h *authxHandler) totpActivate(c *gin.Context) {
 		respondError(c, err)
 		return
 	}
-	c.Status(http.StatusNoContent)
+	h.respondFreshToken(c)
 }

 // totpDisable 停用两步验证;需当前验证码或登录密码任一确认。
@@ -110,7 +112,34 @@ func (h *authxHandler) totpDisable(c *gin.Context) {
 		respondError(c, err)
 		return
 	}
+	h.respondFreshToken(c)
+}
+
+// respondFreshToken 敏感变更后为操作者签发新令牌返回(版本已递增,
+// 旧令牌全部失效);签发失败降级 204,前端按会话失效走重新登录。
+func (h *authxHandler) respondFreshToken(c *gin.Context) {
+	token, expires, err := h.auth.IssueToken(c.Request.Context(), c.GetString(usernameKey))
+	if err != nil {
 		c.Status(http.StatusNoContent)
+		return
+	}
+	c.JSON(http.StatusOK, gin.H{"token": token, "expiresAt": expires})
+}
+
+// revokeSessions 撤销全部会话(令牌版本递增),并为当前操作者重签新令牌。
+//
+// @Summary  撤销全部会话
+// @Tags     认证
+// @Success  200 {object} tokenResponse "新 token 与 expiresAt"
+// @Security BearerAuth
+// @Router   /api/v1/auth/revoke-sessions [post]
+func (h *authxHandler) revokeSessions(c *gin.Context) {
+	token, expires, err := h.auth.RevokeSessions(c.Request.Context(), c.GetString(usernameKey))
+	if err != nil {
+		respondError(c, err)
+		return
+	}
+	c.JSON(http.StatusOK, gin.H{"token": token, "expiresAt": expires})
 }

 // ---- 登录凭据(JWT 组内) ----
@@ -119,7 +148,7 @@ func (h *authxHandler) totpDisable(c *gin.Context) {
 //
 // @Summary  登录凭据摘要
 // @Tags     认证
-// @Success  200 {object} map[string]any "username 与 passwordLoginDisabled"
+// @Success  200 {object} credentialsResponse "username 与 passwordLoginDisabled"
 // @Security BearerAuth
 // @Router   /api/v1/auth/credentials [get]
 func (h *authxHandler) getCredentials(c *gin.Context) {
@@ -140,7 +169,7 @@ func (h *authxHandler) getCredentials(c *gin.Context) {
 // @Tags     认证
 // @Param    body body service.UpdateCredentialsInput true "新凭据(当前密码必验)"
 // @Success  204 "已更新,请重新登录"
-// @Failure  401 {object} map[string]string "当前密码错误"
+// @Failure  401 {object} errorResponse "当前密码错误"
 // @Security BearerAuth
 // @Router   /api/v1/auth/credentials [put]
 func (h *authxHandler) updateCredentials(c *gin.Context) {
@@ -149,7 +178,7 @@ func (h *authxHandler) updateCredentials(c *gin.Context) {
 		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
 		return
 	}
-	err := h.auth.UpdateCredentials(c.Request.Context(), c.GetString(usernameKey), req)
+	finalName, err := h.auth.UpdateCredentials(c.Request.Context(), c.GetString(usernameKey), req)
 	if errors.Is(err, service.ErrCredentialConfirm) {
 		c.JSON(http.StatusUnauthorized, gin.H{"error": err.Error()})
 		return
@@ -162,7 +191,8 @@ func (h *authxHandler) updateCredentials(c *gin.Context) {
 		respondError(c, err)
 		return
 	}
-	c.Status(http.StatusNoContent)
+	c.Set(usernameKey, finalName)
+	h.respondFreshToken(c)
 }

 // updatePasswordLogin 保存密码登录禁用开关;开启需至少绑定一个外部身份。
@@ -171,7 +201,7 @@ func (h *authxHandler) updateCredentials(c *gin.Context) {
 // @Tags     认证
 // @Param    body body object true "{disabled: bool}"
 // @Success  204 "已保存"
-// @Failure  409 {object} map[string]string "未绑定外部身份"
+// @Failure  409 {object} errorResponse "未绑定外部身份"
 // @Security BearerAuth
 // @Router   /api/v1/auth/password-login [put]
 func (h *authxHandler) updatePasswordLogin(c *gin.Context) {
@@ -191,7 +221,7 @@ func (h *authxHandler) updatePasswordLogin(c *gin.Context) {
 		respondError(c, err)
 		return
 	}
-	c.Status(http.StatusNoContent)
+	h.respondFreshToken(c)
 }

 // ---- OAuth ----
@@ -202,7 +232,7 @@ func (h *authxHandler) updatePasswordLogin(c *gin.Context) {
 //
 // @Summary  外部登录 provider 列表
 // @Tags     认证
-// @Success  200 {object} map[string]any "providers 与 passwordLoginDisabled"
+// @Success  200 {object} oauthProvidersResponse "providers 与 passwordLoginDisabled"
 // @Router   /api/v1/auth/oauth/providers [get]
 func (h *authxHandler) oauthProviders(c *gin.Context) {
 	providers := h.oauth.Providers(c.Request.Context())
@@ -219,7 +249,7 @@ func (h *authxHandler) oauthProviders(c *gin.Context) {
 // @Tags     认证
 // @Param    provider path string true "oidc / github"
 // @Param    mode query string false "bind=绑定当前账号(需 Bearer),缺省登录"
-// @Success  200 {object} map[string]string "url"
+// @Success  200 {object} urlResponse "url"
 // @Router   /api/v1/auth/oauth/{provider}/authorize [get]
 func (h *authxHandler) oauthAuthorize(c *gin.Context) {
 	provider := c.Param("provider")
@@ -254,7 +284,7 @@ func (h *authxHandler) bearerUser(c *gin.Context) (string, bool) {
 	if len(header) < 8 || header[:7] != "Bearer " {
 		return "", false
 	}
-	username, err := h.auth.ParseToken(header[7:])
+	username, err := h.auth.ParseToken(c.Request.Context(), header[7:])
 	if err != nil {
 		return "", false
 	}
@@ -282,9 +312,9 @@ func (h *authxHandler) oauthCallback(c *gin.Context) {
 		c.Redirect(http.StatusFound, target+"?oauthError="+url.QueryEscape(oauthErrText(err)))
 		return
 	}
-	if token == "" {
-		// 绑定模式:回设置页安全 tab
-		c.Redirect(http.StatusFound, "/settings?oauth=bound")
+	if mode == "bind" {
+		// 绑定模式:版本已递增,新 token 经 fragment 带回设置页无感换发
+		c.Redirect(http.StatusFound, "/settings?oauth=bound#oauthToken="+url.QueryEscape(token))
 		return
 	}
 	// 登录模式:token 放 fragment,不进服务端日志与 Referer
@@ -305,7 +335,7 @@ func oauthErrText(err error) string {
 //
 // @Summary  已绑定外部身份
 // @Tags     认证
-// @Success  200 {object} map[string]any "items"
+// @Success  200 {object} itemsResponse[model.UserIdentity] "items"
 // @Security BearerAuth
 // @Router   /api/v1/auth/identities [get]
 func (h *authxHandler) identities(c *gin.Context) {
@@ -323,7 +353,7 @@ func (h *authxHandler) identities(c *gin.Context) {
 // @Tags     认证
 // @Param    id path int true "身份 ID"
 // @Success  204 "已解绑"
-// @Failure  409 {object} map[string]string "最后一个身份不可解绑"
+// @Failure  409 {object} errorResponse "最后一个身份不可解绑"
 // @Security BearerAuth
 // @Router   /api/v1/auth/identities/{id} [delete]
 func (h *authxHandler) unbindIdentity(c *gin.Context) {
@@ -340,7 +370,7 @@ func (h *authxHandler) unbindIdentity(c *gin.Context) {
 		respondError(c, err)
 		return
 	}
-	c.Status(http.StatusNoContent)
+	h.respondFreshToken(c)
 }

 // ---- OAuth provider 设置(JWT 组内) ----
@@ -4,6 +4,8 @@ import (
 	"net/http"

 	"github.com/gin-gonic/gin"
+
+	_ "oci-portal/internal/oci" // swagger 注解引用
 )

 // ---- 控制台连接（VNC / 串口） ----
@@ -18,7 +20,7 @@ type createConsoleConnectionRequest struct {
 // @Param    id path int true "配置 ID"
 // @Param    instanceId path string true "instanceId"
 // @Param    body body createConsoleConnectionRequest true "请求体"
-// @Success  201 {object} map[string]any
+// @Success  201 {object} oci.ConsoleConnection
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/instances/{instanceId}/console-connections [post]
 func (h *ociConfigHandler) createConsoleConnection(c *gin.Context) {
@@ -43,7 +45,7 @@ func (h *ociConfigHandler) createConsoleConnection(c *gin.Context) {
 // @Tags     计算
 // @Param    id path int true "配置 ID"
 // @Param    instanceId path string true "instanceId"
-// @Success  200 {object} map[string]any
+// @Success  200 {array} oci.ConsoleConnection
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/instances/{instanceId}/console-connections [get]
 func (h *ociConfigHandler) listConsoleConnections(c *gin.Context) {
@@ -40,7 +40,8 @@ func boolOr(v *bool, def bool) bool {
 // @Summary  身份提供商列表
 // @Tags     租户 IAM
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Param    domainId query string false "身份域 OCID(缺省 Default 域)"
+// @Success  200 {array} oci.IdentityProviderInfo
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/identity-providers [get]
 func (h *ociConfigHandler) listIdentityProviders(c *gin.Context) {
@@ -48,7 +49,7 @@ func (h *ociConfigHandler) listIdentityProviders(c *gin.Context) {
 	if !ok {
 		return
 	}
-	idps, err := h.svc.IdentityProviders(c.Request.Context(), id)
+	idps, err := h.svc.IdentityProviders(c.Request.Context(), id, c.Query("domainId"))
 	if err != nil {
 		respondError(c, err)
 		return
@@ -59,8 +60,9 @@ func (h *ociConfigHandler) listIdentityProviders(c *gin.Context) {
 // @Summary  创建身份提供商(SAML)
 // @Tags     租户 IAM
 // @Param    id path int true "配置 ID"
+// @Param    domainId query string false "身份域 OCID(缺省 Default 域)"
 // @Param    body body createIdpRequest true "请求体"
-// @Success  201 {object} map[string]any
+// @Success  201 {object} oci.IdentityProviderInfo
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/identity-providers [post]
 func (h *ociConfigHandler) createIdentityProvider(c *gin.Context) {
@@ -73,7 +75,7 @@ func (h *ociConfigHandler) createIdentityProvider(c *gin.Context) {
 		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
 		return
 	}
-	idp, err := h.svc.CreateIdentityProvider(c.Request.Context(), id, oci.CreateIdpInput{
+	idp, err := h.svc.CreateIdentityProvider(c.Request.Context(), id, c.Query("domainId"), oci.CreateIdpInput{
 		Name:                req.Name,
 		Metadata:            req.Metadata,
 		Description:         req.Description,
@@ -97,9 +99,10 @@ func (h *ociConfigHandler) createIdentityProvider(c *gin.Context) {
 // @Summary  激活身份提供商
 // @Tags     租户 IAM
 // @Param    id path int true "配置 ID"
+// @Param    domainId query string false "身份域 OCID(缺省 Default 域)"
 // @Param    idpId path string true "idpId"
 // @Param    body body object true "请求体(见接口说明)"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} oci.IdentityProviderInfo
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/identity-providers/{idpId}/activate [post]
 func (h *ociConfigHandler) activateIdentityProvider(c *gin.Context) {
@@ -114,7 +117,7 @@ func (h *ociConfigHandler) activateIdentityProvider(c *gin.Context) {
 		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
 		return
 	}
-	idp, err := h.svc.SetIdentityProviderEnabled(c.Request.Context(), id, c.Param("idpId"), *req.Enabled)
+	idp, err := h.svc.SetIdentityProviderEnabled(c.Request.Context(), id, c.Query("domainId"), c.Param("idpId"), *req.Enabled)
 	if err != nil {
 		respondError(c, err)
 		return
@@ -125,6 +128,7 @@ func (h *ociConfigHandler) activateIdentityProvider(c *gin.Context) {
 // @Summary  删除身份提供商
 // @Tags     租户 IAM
 // @Param    id path int true "配置 ID"
+// @Param    domainId query string false "身份域 OCID(缺省 Default 域)"
 // @Param    idpId path string true "idpId"
 // @Success  204 "无内容"
 // @Security BearerAuth
@@ -134,7 +138,7 @@ func (h *ociConfigHandler) deleteIdentityProvider(c *gin.Context) {
 	if !ok {
 		return
 	}
-	if err := h.svc.DeleteIdentityProvider(c.Request.Context(), id, c.Param("idpId")); err != nil {
+	if err := h.svc.DeleteIdentityProvider(c.Request.Context(), id, c.Query("domainId"), c.Param("idpId")); err != nil {
 		respondError(c, err)
 		return
 	}
@@ -144,7 +148,8 @@ func (h *ociConfigHandler) deleteIdentityProvider(c *gin.Context) {
 // @Summary  下载 SAML 元数据
 // @Tags     租户 IAM
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Param    domainId query string false "身份域 OCID(缺省 Default 域)"
+// @Success  200 {file} file "SAML 元数据 XML(附件下载)"
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/saml-metadata [get]
 func (h *ociConfigHandler) downloadSamlMetadata(c *gin.Context) {
@@ -152,7 +157,7 @@ func (h *ociConfigHandler) downloadSamlMetadata(c *gin.Context) {
 	if !ok {
 		return
 	}
-	metadata, err := h.svc.DomainSamlMetadata(c.Request.Context(), id)
+	metadata, err := h.svc.DomainSamlMetadata(c.Request.Context(), id, c.Query("domainId"))
 	if err != nil {
 		respondError(c, err)
 		return
@@ -164,7 +169,8 @@ func (h *ociConfigHandler) downloadSamlMetadata(c *gin.Context) {
 // @Summary  单点登录规则列表
 // @Tags     租户 IAM
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Param    domainId query string false "身份域 OCID(缺省 Default 域)"
+// @Success  200 {array} oci.SignOnRuleInfo
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/sign-on-rules [get]
 func (h *ociConfigHandler) listSignOnRules(c *gin.Context) {
@@ -172,7 +178,7 @@ func (h *ociConfigHandler) listSignOnRules(c *gin.Context) {
 	if !ok {
 		return
 	}
-	rules, err := h.svc.ConsoleSignOnRules(c.Request.Context(), id)
+	rules, err := h.svc.ConsoleSignOnRules(c.Request.Context(), id, c.Query("domainId"))
 	if err != nil {
 		respondError(c, err)
 		return
@@ -183,8 +189,9 @@ func (h *ociConfigHandler) listSignOnRules(c *gin.Context) {
 // @Summary  创建 MFA 豁免规则
 // @Tags     租户 IAM
 // @Param    id path int true "配置 ID"
+// @Param    domainId query string false "身份域 OCID(缺省 Default 域)"
 // @Param    body body object true "请求体(见接口说明)"
-// @Success  201 {object} map[string]any
+// @Success  201 {object} oci.SignOnRuleInfo
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/sign-on-exemptions [post]
 func (h *ociConfigHandler) createMfaExemption(c *gin.Context) {
@@ -199,7 +206,7 @@ func (h *ociConfigHandler) createMfaExemption(c *gin.Context) {
 		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
 		return
 	}
-	rule, err := h.svc.CreateMfaExemption(c.Request.Context(), id, req.IdentityProviderID)
+	rule, err := h.svc.CreateMfaExemption(c.Request.Context(), id, c.Query("domainId"), req.IdentityProviderID)
 	if err != nil {
 		respondError(c, err)
 		return
@@ -210,6 +217,7 @@ func (h *ociConfigHandler) createMfaExemption(c *gin.Context) {
 // @Summary  删除 MFA 豁免规则
 // @Tags     租户 IAM
 // @Param    id path int true "配置 ID"
+// @Param    domainId query string false "身份域 OCID(缺省 Default 域)"
 // @Param    ruleId path string true "ruleId"
 // @Success  204 "无内容"
 // @Security BearerAuth
@@ -219,7 +227,7 @@ func (h *ociConfigHandler) deleteMfaExemption(c *gin.Context) {
 	if !ok {
 		return
 	}
-	if err := h.svc.DeleteMfaExemption(c.Request.Context(), id, c.Param("ruleId")); err != nil {
+	if err := h.svc.DeleteMfaExemption(c.Request.Context(), id, c.Query("domainId"), c.Param("ruleId")); err != nil {
 		respondError(c, err)
 		return
 	}
@@ -11,7 +11,7 @@ import (
 // @Summary  可用域列表
 // @Tags     租户配置
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {array} string
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/availability-domains [get]
 func (h *ociConfigHandler) availabilityDomains(c *gin.Context) {
@@ -67,7 +67,7 @@ type instanceActionRequest struct {
 // @Summary  实例列表
 // @Tags     计算
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {array} oci.Instance
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/instances [get]
 func (h *ociConfigHandler) listInstances(c *gin.Context) {
@@ -87,7 +87,7 @@ func (h *ociConfigHandler) listInstances(c *gin.Context) {
 // @Tags     计算
 // @Param    id path int true "配置 ID"
 // @Param    body body createInstanceRequest true "请求体"
-// @Success  201 {object} map[string]any
+// @Success  201 {object} createInstancesResponse "部分成功仍 201,errors 为逐台失败信息"
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/instances [post]
 func (h *ociConfigHandler) createInstance(c *gin.Context) {
@@ -147,7 +147,7 @@ func (h *ociConfigHandler) createInstance(c *gin.Context) {
 // @Tags     计算
 // @Param    id path int true "配置 ID"
 // @Param    instanceId path string true "instanceId"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} oci.Instance
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/instances/{instanceId} [get]
 func (h *ociConfigHandler) getInstance(c *gin.Context) {
@@ -168,7 +168,7 @@ func (h *ociConfigHandler) getInstance(c *gin.Context) {
 // @Param    id path int true "配置 ID"
 // @Param    instanceId path string true "instanceId"
 // @Param    body body updateInstanceRequest true "请求体"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} oci.Instance
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/instances/{instanceId} [put]
 func (h *ociConfigHandler) updateInstance(c *gin.Context) {
@@ -220,7 +220,7 @@ func (h *ociConfigHandler) terminateInstance(c *gin.Context) {
 // @Param    id path int true "配置 ID"
 // @Param    instanceId path string true "instanceId"
 // @Param    body body instanceActionRequest true "请求体"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} oci.Instance
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/instances/{instanceId}/action [post]
 func (h *ociConfigHandler) instanceAction(c *gin.Context) {
@@ -253,7 +253,7 @@ type updateBootVolumeRequest struct {
 // @Summary  引导卷列表
 // @Tags     存储
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {array} oci.BootVolume
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/boot-volumes [get]
 func (h *ociConfigHandler) listBootVolumes(c *gin.Context) {
@@ -273,7 +273,7 @@ func (h *ociConfigHandler) listBootVolumes(c *gin.Context) {
 // @Tags     存储
 // @Param    id path int true "配置 ID"
 // @Param    bootVolumeId path string true "bootVolumeId"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} oci.BootVolume
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/boot-volumes/{bootVolumeId} [get]
 func (h *ociConfigHandler) getBootVolume(c *gin.Context) {
@@ -294,7 +294,7 @@ func (h *ociConfigHandler) getBootVolume(c *gin.Context) {
 // @Param    id path int true "配置 ID"
 // @Param    bootVolumeId path string true "bootVolumeId"
 // @Param    body body updateBootVolumeRequest true "请求体"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} oci.BootVolume
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/boot-volumes/{bootVolumeId} [put]
 func (h *ociConfigHandler) updateBootVolume(c *gin.Context) {
@@ -7,6 +7,8 @@ import (

 	"github.com/gin-gonic/gin"

+	_ "oci-portal/internal/model" // swagger 注解引用
+
 	"oci-portal/internal/service"
 )

@@ -20,7 +22,7 @@ type logEventHandler struct {
 // @Summary  查询配置的回调地址
 // @Tags     任务与日志回传
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} logWebhookStatusResponse
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/log-webhook [get]
 func (h *logEventHandler) getWebhook(c *gin.Context) {
@@ -45,7 +47,7 @@ func (h *logEventHandler) getWebhook(c *gin.Context) {
 // @Summary  生成
 // @Tags     任务与日志回传
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} service.LogWebhookInfo
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/log-webhook [post]
 func (h *logEventHandler) ensureWebhook(c *gin.Context) {
@@ -87,7 +89,7 @@ func (h *logEventHandler) revokeWebhook(c *gin.Context) {
 // @Tags     任务与日志回传
 // @Param    configId query int false "按配置过滤"
 // @Param    q query string false "关键字"
-// @Success  200 {object} map[string]any "items 与 total"
+// @Success  200 {object} pagedResponse[model.LogEvent] "items 与 total"
 // @Security BearerAuth
 // @Router   /api/v1/log-events [get]
 func (h *logEventHandler) list(c *gin.Context) {
@@ -111,7 +113,7 @@ func (h *logEventHandler) list(c *gin.Context) {
 // @Summary  查询 OCI 侧链路状态与关键事件清单
 // @Tags     任务与日志回传
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} service.RelayView
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/log-relay [get]
 func (h *logEventHandler) getRelay(c *gin.Context) {
@@ -132,7 +134,7 @@ func (h *logEventHandler) getRelay(c *gin.Context) {
 // @Summary  一键建立回传链路
 // @Tags     任务与日志回传
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} service.RelayView
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/log-relay [post]
 func (h *logEventHandler) setupRelay(c *gin.Context) {
@@ -15,7 +15,8 @@ import (
 // usernameKey 是鉴权通过后写入 gin.Context 的用户名键。
 const usernameKey = "username"

-// RequireAuth 校验 Authorization: Bearer 令牌，通过后把用户名放进上下文。
+// RequireAuth 校验 Authorization: Bearer 令牌（签名、有效期与令牌版本），
+// 通过后把用户名放进上下文。
 func RequireAuth(auth *service.AuthService) gin.HandlerFunc {
 	return func(c *gin.Context) {
 		token, ok := strings.CutPrefix(c.GetHeader("Authorization"), "Bearer ")
@@ -23,7 +24,7 @@ func RequireAuth(auth *service.AuthService) gin.HandlerFunc {
 			c.AbortWithStatusJSON(http.StatusUnauthorized, gin.H{"error": "missing bearer token"})
 			return
 		}
-		username, err := auth.ParseToken(token)
+		username, err := auth.ParseToken(c.Request.Context(), token)
 		if err != nil {
 			c.AbortWithStatusJSON(http.StatusUnauthorized, gin.H{"error": "invalid or expired token"})
 			return
@@ -33,6 +34,17 @@ func RequireAuth(auth *service.AuthService) gin.HandlerFunc {
 	}
 }

+// bodyLimit 给请求体套上限(S-03):超限读取由 MaxBytesReader 截断报错,
+// 绑定失败返回 400;webhook 入口另有独立 256KiB 自限,不经此中间件。
+func bodyLimit(n int64) gin.HandlerFunc {
+	return func(c *gin.Context) {
+		if c.Request.Body != nil {
+			c.Request.Body = http.MaxBytesReader(c.Writer, c.Request.Body, n)
+		}
+		c.Next()
+	}
+}
+
 // systemLogMiddleware 把写请求（POST/PUT/DELETE）异步记入系统日志，只读请求跳过。
 // 只记方法、路径等元数据，绝不读请求体——请求体可能含私钥 / 口令等敏感数据。
 func systemLogMiddleware(logs *service.SystemLogService) gin.HandlerFunc {
@@ -11,7 +11,7 @@ import (
 // @Summary  实例形状列表
 // @Tags     租户配置
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {array} oci.ComputeShape
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/shapes [get]
 func (h *ociConfigHandler) shapes(c *gin.Context) {
@@ -30,7 +30,7 @@ func (h *ociConfigHandler) shapes(c *gin.Context) {
 // @Summary  镜像列表
 // @Tags     租户配置
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {array} oci.Image
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/images [get]
 func (h *ociConfigHandler) images(c *gin.Context) {
@@ -54,7 +54,7 @@ func (h *ociConfigHandler) images(c *gin.Context) {
 // @Tags     租户配置
 // @Param    id path int true "配置 ID"
 // @Param    imageId path string true "imageId"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} oci.Image
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/images/{imageId} [get]
 func (h *ociConfigHandler) getImage(c *gin.Context) {
@@ -89,7 +89,7 @@ type renameRequest struct {
 // @Summary  VCN 列表
 // @Tags     网络
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {array} oci.VCN
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/vcns [get]
 func (h *ociConfigHandler) listVCNs(c *gin.Context) {
@@ -109,7 +109,7 @@ func (h *ociConfigHandler) listVCNs(c *gin.Context) {
 // @Tags     网络
 // @Param    id path int true "配置 ID"
 // @Param    body body createVCNRequest true "请求体"
-// @Success  201 {object} map[string]any
+// @Success  201 {object} oci.VCN
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/vcns [post]
 func (h *ociConfigHandler) createVCN(c *gin.Context) {
@@ -141,7 +141,7 @@ func (h *ociConfigHandler) createVCN(c *gin.Context) {
 // @Tags     网络
 // @Param    id path int true "配置 ID"
 // @Param    vcnId path string true "vcnId"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} oci.VCN
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/vcns/{vcnId} [get]
 func (h *ociConfigHandler) getVCN(c *gin.Context) {
@@ -162,7 +162,7 @@ func (h *ociConfigHandler) getVCN(c *gin.Context) {
 // @Param    id path int true "配置 ID"
 // @Param    vcnId path string true "vcnId"
 // @Param    body body renameRequest true "请求体"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} oci.VCN
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/vcns/{vcnId} [put]
 func (h *ociConfigHandler) updateVCN(c *gin.Context) {
@@ -211,7 +211,7 @@ type enableIPv6Request struct {
 // @Param    id path int true "配置 ID"
 // @Param    vcnId path string true "vcnId"
 // @Param    body body enableIPv6Request true "请求体"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} ipv6StepsResponse
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/vcns/{vcnId}/enable-ipv6 [post]
 func (h *ociConfigHandler) enableVCNIPv6(c *gin.Context) {
@@ -247,7 +247,7 @@ type createSubnetRequest struct {
 // @Summary  子网列表
 // @Tags     网络
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {array} oci.Subnet
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/subnets [get]
 func (h *ociConfigHandler) listSubnets(c *gin.Context) {
@@ -267,7 +267,7 @@ func (h *ociConfigHandler) listSubnets(c *gin.Context) {
 // @Tags     网络
 // @Param    id path int true "配置 ID"
 // @Param    body body createSubnetRequest true "请求体"
-// @Success  201 {object} map[string]any
+// @Success  201 {object} oci.Subnet
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/subnets [post]
 func (h *ociConfigHandler) createSubnet(c *gin.Context) {
@@ -300,7 +300,7 @@ func (h *ociConfigHandler) createSubnet(c *gin.Context) {
 // @Tags     网络
 // @Param    id path int true "配置 ID"
 // @Param    subnetId path string true "subnetId"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} oci.Subnet
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/subnets/{subnetId} [get]
 func (h *ociConfigHandler) getSubnet(c *gin.Context) {
@@ -321,7 +321,7 @@ func (h *ociConfigHandler) getSubnet(c *gin.Context) {
 // @Param    id path int true "配置 ID"
 // @Param    subnetId path string true "subnetId"
 // @Param    body body renameRequest true "请求体"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} oci.Subnet
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/subnets/{subnetId} [put]
 func (h *ociConfigHandler) updateSubnet(c *gin.Context) {
@@ -381,7 +381,7 @@ type updateSecurityListRequest struct {
 // @Summary  安全列表清单
 // @Tags     网络
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {array} oci.SecurityList
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/security-lists [get]
 func (h *ociConfigHandler) listSecurityLists(c *gin.Context) {
@@ -401,7 +401,7 @@ func (h *ociConfigHandler) listSecurityLists(c *gin.Context) {
 // @Tags     网络
 // @Param    id path int true "配置 ID"
 // @Param    body body createSecurityListRequest true "请求体"
-// @Success  201 {object} map[string]any
+// @Success  201 {object} oci.SecurityList
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/security-lists [post]
 func (h *ociConfigHandler) createSecurityList(c *gin.Context) {
@@ -432,7 +432,7 @@ func (h *ociConfigHandler) createSecurityList(c *gin.Context) {
 // @Tags     网络
 // @Param    id path int true "配置 ID"
 // @Param    securityListId path string true "securityListId"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} oci.SecurityList
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/security-lists/{securityListId} [get]
 func (h *ociConfigHandler) getSecurityList(c *gin.Context) {
@@ -453,7 +453,7 @@ func (h *ociConfigHandler) getSecurityList(c *gin.Context) {
 // @Param    id path int true "配置 ID"
 // @Param    securityListId path string true "securityListId"
 // @Param    body body updateSecurityListRequest true "请求体"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} oci.SecurityList
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/security-lists/{securityListId} [put]
 func (h *ociConfigHandler) updateSecurityList(c *gin.Context) {
@@ -1,13 +1,18 @@
 package api

 import (
+	"crypto/rand"
+	"encoding/hex"
 	"errors"
+	"log"
 	"net/http"
 	"strconv"

 	"github.com/gin-gonic/gin"
+
 	"github.com/oracle/oci-go-sdk/v65/common"
 	"gorm.io/gorm"
+	_ "oci-portal/internal/model" // swagger 注解引用

 	"oci-portal/internal/oci"
 	"oci-portal/internal/service"
@@ -36,7 +41,7 @@ type importRequest struct {
 // @Summary  导入租户配置
 // @Tags     租户配置
 // @Param    body body importRequest true "API Key 配置(私钥密文落库)"
-// @Success  201 {object} map[string]any
+// @Success  201 {object} model.OciConfig
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs [post]
 func (h *ociConfigHandler) create(c *gin.Context) {
@@ -68,13 +73,13 @@ func (h *ociConfigHandler) create(c *gin.Context) {

 // @Summary  租户配置列表
 // @Tags     租户配置
-// @Success  200 {object} map[string]any "items"
+// @Success  200 {array} service.ConfigSummary
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs [get]
 func (h *ociConfigHandler) list(c *gin.Context) {
 	items, err := h.svc.List(c.Request.Context())
 	if err != nil {
-		c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
+		respondError(c, err)
 		return
 	}
 	c.JSON(http.StatusOK, items)
@@ -83,7 +88,7 @@ func (h *ociConfigHandler) list(c *gin.Context) {
 // @Summary  租户配置详情
 // @Tags     租户配置
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} model.OciConfig
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id} [get]
 func (h *ociConfigHandler) get(c *gin.Context) {
@@ -102,7 +107,7 @@ func (h *ociConfigHandler) get(c *gin.Context) {
 // @Summary  验证配置连通性并刷新画像
 // @Tags     租户配置
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} configChangesResponse "config 与 changes 字段变更对照"
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/verify [post]
 func (h *ociConfigHandler) verify(c *gin.Context) {
@@ -136,7 +141,7 @@ type updateConfigRequest struct {
 // @Tags     租户配置
 // @Param    id path int true "配置 ID"
 // @Param    body body object true "可更新字段(别名/分组/代理/多区域开关等)"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} configChangesResponse "config 与 changes 字段变更对照"
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id} [put]
 func (h *ociConfigHandler) update(c *gin.Context) {
@@ -168,6 +173,7 @@ func (h *ociConfigHandler) update(c *gin.Context) {
 }

 // @Summary  删除租户配置
+// @Description 删除租户配置及其本地关联任务、快照、回传、告警和 AI 渠道数据，并撤销 Webhook 密钥；不删除 OCI 云端资源。
 // @Tags     租户配置
 // @Param    id path int true "配置 ID"
 // @Success  204 "无内容"
@@ -190,7 +196,7 @@ func (h *ociConfigHandler) remove(c *gin.Context) {
 // @Summary  列出租户下全部 ACTIVE compartment
 // @Tags     租户配置
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {array} oci.Compartment
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/compartments [get]
 func (h *ociConfigHandler) compartments(c *gin.Context) {
@@ -212,7 +218,7 @@ func (h *ociConfigHandler) compartments(c *gin.Context) {
 // @Summary  返回筛选器用区域列表：未开多区域支持只含默认区域
 // @Tags     租户配置
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {array} service.RegionSubscriptionView
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/cached-regions [get]
 func (h *ociConfigHandler) cachedRegions(c *gin.Context) {
@@ -233,7 +239,7 @@ func (h *ociConfigHandler) cachedRegions(c *gin.Context) {
 // @Summary  返回筛选器用区间列表：未开多区间支持返回空数组
 // @Tags     租户配置
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {array} oci.Compartment
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/cached-compartments [get]
 func (h *ociConfigHandler) cachedCompartments(c *gin.Context) {
@@ -258,17 +264,29 @@ func pathID(c *gin.Context) (uint, bool) {
 	return uint(id), true
 }

+// respondError 统一出错响应边界(S-08):OCI 服务错误经脱敏透出,记录缺失
+// 映射 404;其余按内部错误处理——响应只含固定文案与关联 ID,完整原因
+// (可能携带 SQL / DSN / 路径等内部细节)仅写服务端日志,经同一 ID 对应。
 func respondError(c *gin.Context, err error) {
-	status := http.StatusInternalServerError
-	if errors.Is(err, gorm.ErrRecordNotFound) {
-		status = http.StatusNotFound
-	}
 	var svcErr common.ServiceError
 	if errors.As(err, &svcErr) {
 		respondOCIError(c, err, svcErr)
 		return
 	}
-	c.JSON(status, gin.H{"error": err.Error()})
+	if errors.Is(err, gorm.ErrRecordNotFound) {
+		c.JSON(http.StatusNotFound, gin.H{"error": "资源不存在"})
+		return
+	}
+	id := newRequestID()
+	log.Printf("[ERR %s] %s %s: %v", id, c.Request.Method, requestPath(c), err)
+	c.JSON(http.StatusInternalServerError, gin.H{"error": "服务器内部错误", "requestId": id})
+}
+
+// newRequestID 生成错误关联 ID(8 字节随机 hex),响应与服务端日志据此对应。
+func newRequestID() string {
+	b := make([]byte, 8)
+	_, _ = rand.Read(b)
+	return hex.EncodeToString(b)
 }

 // respondOCIError 提炼 OCI 服务错误：透传 HTTP 状态码，
@@ -278,9 +296,10 @@ func respondOCIError(c *gin.Context, err error, svcErr common.ServiceError) {
 	if status < http.StatusBadRequest {
 		status = http.StatusBadGateway
 	}
-	// 面板的 401 专属本地 JWT 失效（前端收到即登出）；
-	// 上游 OCI 的 401 是代理调用被拒，改用 502 透出
-	if status == http.StatusUnauthorized {
+	// 面板的 401 专属本地 JWT 失效（前端收到即登出）、429 专属本地 IP 限速
+	// （前端收到即跳 /blocked）；上游 OCI 的同码是代理调用被拒/云端限流，
+	// 均改用 502 透出，body 保留 ociCode 供辨识
+	if status == http.StatusUnauthorized || status == http.StatusTooManyRequests {
 		status = http.StatusBadGateway
 	}
 	body := gin.H{
@@ -0,0 +1,54 @@
+package api
+
+import (
+	"fmt"
+	"net/http/httptest"
+	"testing"
+
+	"github.com/gin-gonic/gin"
+)
+
+// fakeServiceError 模拟 oci-go-sdk 的 common.ServiceError，驱动 respondOCIError 分支。
+type fakeServiceError struct {
+	status  int
+	code    string
+	message string
+}
+
+func (e fakeServiceError) Error() string {
+	return fmt.Sprintf("Error returned by service. Http Status Code: %d. Error Code: %s. Message: %s",
+		e.status, e.code, e.message)
+}
+
+func (e fakeServiceError) GetHTTPStatusCode() int  { return e.status }
+func (e fakeServiceError) GetMessage() string      { return e.message }
+func (e fakeServiceError) GetCode() string         { return e.code }
+func (e fakeServiceError) GetOpcRequestID() string { return "req-1" }
+
+// TestRespondErrorOCIStatus 覆盖上游状态码改写规则：
+// 401/429 在前端有本地专属语义（登出 / 跳 /blocked），上游同码须改写 502。
+func TestRespondErrorOCIStatus(t *testing.T) {
+	gin.SetMode(gin.TestMode)
+	cases := []struct {
+		name       string
+		upstream   int
+		wantStatus int
+	}{
+		{"404 原样透传", 404, 404},
+		{"409 原样透传", 409, 409},
+		{"上游 401 改写 502", 401, 502},
+		{"上游 429 改写 502", 429, 502},
+		{"非错误码兜底 502", 200, 502},
+	}
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			rec := httptest.NewRecorder()
+			c, _ := gin.CreateTestContext(rec)
+			err := fmt.Errorf("获取流量: %w", fakeServiceError{tc.upstream, "TooManyRequests", "slow down"})
+			respondError(c, err)
+			if rec.Code != tc.wantStatus {
+				t.Fatalf("status = %d, want %d", rec.Code, tc.wantStatus)
+			}
+		})
+	}
+}
@@ -4,13 +4,15 @@ import (
 	"net/http"

 	"github.com/gin-gonic/gin"
+
+	_ "oci-portal/internal/service" // swagger 注解引用
 )

 // overview 返回总览页聚合数据（本地快照，不发云端请求）。
 //
 // @Summary  返回总览页聚合数据
 // @Tags     租户配置
-// @Success  200 {object} map[string]any
+// @Success  200 {object} service.Overview
 // @Security BearerAuth
 // @Router   /api/v1/overview [get]
 func (h *ociConfigHandler) overview(c *gin.Context) {
@@ -18,7 +18,7 @@ type proxyHandler struct {
 //
 // @Summary  代理列表
 // @Tags     设置
-// @Success  200 {object} map[string]any "items"
+// @Success  200 {object} itemsResponse[service.ProxyView] "items"
 // @Security BearerAuth
 // @Router   /api/v1/proxies [get]
 func (h *proxyHandler) list(c *gin.Context) {
@@ -35,7 +35,7 @@ func (h *proxyHandler) list(c *gin.Context) {
 // @Summary  创建代理
 // @Tags     设置
 // @Param    body body service.ProxyInput true "代理配置(密码只写不回)"
-// @Success  201 {object} map[string]any
+// @Success  201 {object} service.ProxyView
 // @Security BearerAuth
 // @Router   /api/v1/proxies [post]
 func (h *proxyHandler) create(c *gin.Context) {
@@ -58,7 +58,7 @@ func (h *proxyHandler) create(c *gin.Context) {
 // @Tags     设置
 // @Param    id path int true "代理 ID"
 // @Param    body body service.ProxyInput true "代理配置(密码缺省沿用)"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} service.ProxyView
 // @Security BearerAuth
 // @Router   /api/v1/proxies/{id} [put]
 func (h *proxyHandler) update(c *gin.Context) {
@@ -104,7 +104,7 @@ func (h *proxyHandler) remove(c *gin.Context) {
 // @Summary  批量导入代理
 // @Tags     设置
 // @Param    body body object true "请求体(见接口说明)"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} service.ImportResult
 // @Security BearerAuth
 // @Router   /api/v1/proxies/import [post]
 func (h *proxyHandler) importBatch(c *gin.Context) {
@@ -128,7 +128,7 @@ func (h *proxyHandler) importBatch(c *gin.Context) {
 // @Summary  手动重测代理出口地区,同步返回最新视图
 // @Tags     设置
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} service.ProxyView
 // @Security BearerAuth
 // @Router   /api/v1/proxies/{id}/probe [post]
 func (h *proxyHandler) probe(c *gin.Context) {
@@ -144,6 +144,35 @@ func (h *proxyHandler) probe(c *gin.Context) {
 	c.JSON(http.StatusOK, view)
 }

+// setTenants 设置代理关联的租户全集(代理侧批量关联 / 解除)。
+//
+// @Summary  设置代理关联的租户全集
+// @Tags     设置
+// @Param    id path int true "代理 ID"
+// @Param    body body object true "请求体 {\"ociConfigIds\": [1,2]}"
+// @Success  200 {object} usedByResponse "usedBy"
+// @Security BearerAuth
+// @Router   /api/v1/proxies/{id}/tenants [put]
+func (h *proxyHandler) setTenants(c *gin.Context) {
+	id, ok := pathID(c)
+	if !ok {
+		return
+	}
+	var req struct {
+		OciConfigIDs []uint `json:"ociConfigIds"`
+	}
+	if err := c.ShouldBindJSON(&req); err != nil {
+		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
+		return
+	}
+	used, err := h.svc.SetTenants(c.Request.Context(), id, req.OciConfigIDs)
+	if err != nil {
+		h.respond(c, err)
+		return
+	}
+	c.JSON(http.StatusOK, gin.H{"usedBy": used})
+}
+
 // respond 把代理业务错误映射到语义状态码。
 func (h *proxyHandler) respond(c *gin.Context, err error) {
 	switch {
@@ -6,6 +6,8 @@ import (

 	"github.com/gin-gonic/gin"

+	_ "oci-portal/internal/service" // swagger 注解引用
+
 	"oci-portal/internal/oci"
 )

@@ -13,13 +15,13 @@ import (
 //
 // @Summary  返回本地维护的完整区域表
 // @Tags     租户配置
-// @Success  200 {object} map[string]any
+// @Success  200 {array} oci.RegionInfo
 // @Security BearerAuth
 // @Router   /api/v1/regions [get]
 func listRegions(c *gin.Context) {
 	regions, err := oci.AllRegions()
 	if err != nil {
-		c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
+		respondError(c, err)
 		return
 	}
 	c.JSON(http.StatusOK, regions)
@@ -28,7 +30,7 @@ func listRegions(c *gin.Context) {
 // @Summary  区域订阅列表
 // @Tags     租户配置
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {array} service.RegionSubscriptionView
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/region-subscriptions [get]
 func (h *ociConfigHandler) regionSubscriptions(c *gin.Context) {
@@ -52,7 +54,7 @@ type subscribeRegionRequest struct {
 // @Tags     租户配置
 // @Param    id path int true "配置 ID"
 // @Param    body body subscribeRegionRequest true "请求体"
-// @Success  200 {object} map[string]any
+// @Success  200 {array} service.RegionSubscriptionView
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/region-subscriptions [post]
 func (h *ociConfigHandler) subscribeRegion(c *gin.Context) {
@@ -76,7 +78,7 @@ func (h *ociConfigHandler) subscribeRegion(c *gin.Context) {
 // @Summary  服务限额查询
 // @Tags     租户配置
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {array} oci.LimitValue
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/limits [get]
 func (h *ociConfigHandler) limits(c *gin.Context) {
@@ -21,7 +21,8 @@ func NewRouter(auth *service.AuthService, oauth *service.OAuthService, ociConfig
 	accessLog := gin.LoggerWithConfig(gin.LoggerConfig{SkipQueryString: true})
 	r.Use(accessLog, RealIPMiddleware(settings), IPRateMiddleware(settings), gin.Recovery())

-	v1 := r.Group("/api/v1")
+	// 面板 API 请求体统一 1MB 上限(S-03);AI 网关按对话体量单独 10MB
+	v1 := r.Group("/api/v1", bodyLimit(1<<20))
 	registerAuthPublic(v1, auth, oauth, systemLogs)
 	// AI 网关对外端点:独立密钥鉴权,挂在全局 Use 之后,仍受 IP 限速与 Recovery 保护
 	registerAiGateway(r, aiGateway)
@@ -2,6 +2,8 @@ package api

 import (
 	"context"
+	"encoding/json"
+	"fmt"
 	"net/http"
 	"net/http/httptest"
 	"strings"
@@ -55,6 +57,13 @@ func (nullClient) GetSubscription(context.Context, oci.Credentials, string) (oci
 }

 func newTestRouter(t *testing.T) (*gin.Engine, *service.AuthService, *service.SystemLogService) {
+	t.Helper()
+	r, auth, systemLogs, _ := newTestRouterDB(t)
+	return r, auth, systemLogs
+}
+
+// newTestRouterDB 额外回传 db 句柄,供需要直插数据(如网关模型缓存)的测试使用。
+func newTestRouterDB(t *testing.T) (*gin.Engine, *service.AuthService, *service.SystemLogService, *gorm.DB) {
 	t.Helper()
 	gin.SetMode(gin.TestMode)
 	db, err := gorm.Open(sqlite.Open(":memory:"), &gorm.Config{
@@ -88,7 +97,7 @@ func newTestRouter(t *testing.T) (*gin.Engine, *service.AuthService, *service.Sy
 	logEvents := service.NewLogEventService(db)
 	oauth := service.NewOAuthService(db, settings, auth)
 	r := NewRouter(auth, oauth, ociConfigs, tasks, service.NewConsoleService(ociConfigs), settings, notifier, systemLogs, logEvents, service.NewProxyService(db, cipher), service.NewAiGatewayService(db, ociConfigs, nullClient{}))
-	return r, auth, systemLogs
+	return r, auth, systemLogs, db
 }

 func doRequest(t *testing.T, r *gin.Engine, method, path, token, body string) *httptest.ResponseRecorder {
@@ -183,3 +192,103 @@ func TestSystemLogsEndpoint(t *testing.T) {
 		})
 	}
 }
+
+// TestLoginFieldLimits 超长登录字段被绑定校验拒绝(S-03 高基数防护第一道)。
+func TestLoginFieldLimits(t *testing.T) {
+	r, _, _ := newTestRouter(t)
+	longName := strings.Repeat("a", 65)
+	longPass := strings.Repeat("b", 129)
+	tests := []struct {
+		name string
+		body string
+	}{
+		{name: "用户名超长", body: `{"username":"` + longName + `","password":"x"}`},
+		{name: "密码超长", body: `{"username":"admin","password":"` + longPass + `"}`},
+		{name: "验证码超长", body: `{"username":"admin","password":"x","totpCode":"123456789"}`},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			w := doRequest(t, r, http.MethodPost, "/api/v1/auth/login", "", tt.body)
+			if w.Code != http.StatusBadRequest {
+				t.Errorf("status = %d, want 400", w.Code)
+			}
+		})
+	}
+}
+
+// TestBodyLimitRejectsHuge 面板 API 请求体超 1MB 被拒(S-03)。
+func TestBodyLimitRejectsHuge(t *testing.T) {
+	r, _, _ := newTestRouter(t)
+	huge := `{"username":"admin","password":"` + strings.Repeat("x", 2<<20) + `"}`
+	w := doRequest(t, r, http.MethodPost, "/api/v1/auth/login", "", huge)
+	if w.Code != http.StatusBadRequest {
+		t.Errorf("status = %d, want 400(body too large)", w.Code)
+	}
+}
+
+// TestRevokeSessionsEndpoint 撤销全部会话:直接调用即生效,
+// 响应带新 token,旧 token 随即失效。
+func TestRevokeSessionsEndpoint(t *testing.T) {
+	r, _, _ := newTestRouter(t)
+	login := doRequest(t, r, http.MethodPost, "/api/v1/auth/login", "", `{"username":"admin","password":"pass123"}`)
+	var sess struct{ Token string }
+	if err := json.Unmarshal(login.Body.Bytes(), &sess); err != nil || sess.Token == "" {
+		t.Fatalf("login: %s", login.Body.String())
+	}
+	w := doRequest(t, r, http.MethodPost, "/api/v1/auth/revoke-sessions", sess.Token, "")
+	if w.Code != http.StatusOK || !strings.Contains(w.Body.String(), "token") {
+		t.Fatalf("revoke = %d %s", w.Code, w.Body.String())
+	}
+	w = doRequest(t, r, http.MethodGet, "/api/v1/auth/credentials", sess.Token, "")
+	if w.Code != http.StatusUnauthorized {
+		t.Errorf("撤销后旧 token 访问 = %d, want 401", w.Code)
+	}
+}
+
+// TestRespondErrorSanitizesInternal 内部错误边界(S-08):wrap 链细节不透传,
+// 响应为固定文案 + requestId;记录缺失映射 404 固定文案。
+func TestRespondErrorSanitizesInternal(t *testing.T) {
+	gin.SetMode(gin.TestMode)
+	cases := []struct {
+		name       string
+		err        error
+		wantStatus int
+		wantBody   string   // 必须出现
+		leaks      []string // 不得出现
+	}{
+		{
+			name:       "内部错误脱敏",
+			err:        fmt.Errorf("query users: dial tcp 10.0.0.5:3306: dsn user:secretpw"),
+			wantStatus: http.StatusInternalServerError,
+			wantBody:   "requestId",
+			leaks:      []string{"secretpw", "10.0.0.5", "dsn", "dial tcp"},
+		},
+		{
+			name:       "记录缺失固定文案",
+			err:        fmt.Errorf("find oci config 5: %w", gorm.ErrRecordNotFound),
+			wantStatus: http.StatusNotFound,
+			wantBody:   "资源不存在",
+			leaks:      []string{"find oci config"},
+		},
+	}
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			rec := httptest.NewRecorder()
+			c, _ := gin.CreateTestContext(rec)
+			c.Request = httptest.NewRequest(http.MethodGet, "/api/v1/x", nil)
+			respondError(c, tc.err)
+			if rec.Code != tc.wantStatus {
+				t.Fatalf("status = %d, want %d", rec.Code, tc.wantStatus)
+			}
+			body := rec.Body.String()
+			if !strings.Contains(body, tc.wantBody) {
+				t.Errorf("body %q 应包含 %q", body, tc.wantBody)
+			}
+			for _, leak := range tc.leaks {
+				if strings.Contains(body, leak) {
+					t.Errorf("body %q 泄露内部细节 %q", body, leak)
+				}
+			}
+		})
+	}
+}
@@ -10,11 +10,15 @@ import (
 // 挂在全局中间件之后,仍受 IP 限速与 Recovery 保护;高频调用自有 AiCallLog。
 func registerAiGateway(r *gin.Engine, aiGateway *service.AiGatewayService) {
 	aih := &aiGatewayHandler{gw: aiGateway}
-	ai := r.Group("/ai/v1", aih.auth)
+	ai := r.Group("/ai/v1", bodyLimit(10<<20), aih.auth)
 	ai.POST("/chat/completions", aih.chatCompletions)
 	ai.POST("/responses", aih.responses)
 	ai.POST("/messages", aih.messages)
 	ai.POST("/embeddings", aih.embeddings)
+	ai.POST("/audio/speech", aih.audioSpeech)
+	ai.POST("/tts", aih.tts)
+	ai.POST("/rerank", aih.rerank)
+	ai.POST("/moderations", aih.moderations)
 	ai.GET("/models", aih.listModels)
 }

@@ -33,6 +37,11 @@ func registerAiAdmin(secured *gin.RouterGroup, aiGateway *service.AiGatewayServi
 	secured.POST("/ai-channels/:id/probe", aiadmin.probeChannel)
 	secured.POST("/ai-channels/:id/sync-models", aiadmin.syncChannelModels)
 	secured.GET("/ai-models", aiadmin.gatewayModels)
+	secured.GET("/ai-settings", aiadmin.aiSettings)
+	secured.PUT("/ai-settings", aiadmin.updateAiSettings)
+	secured.GET("/ai-blacklist", aiadmin.listBlacklist)
+	secured.POST("/ai-blacklist", aiadmin.addBlacklist)
+	secured.DELETE("/ai-blacklist/:id", aiadmin.removeBlacklist)
 	secured.GET("/ai-logs", aiadmin.listLogs)
 	secured.GET("/ai-content-logs", aiadmin.listContentLogs)
 }
@@ -26,13 +26,14 @@ func registerAuthSecured(secured *gin.RouterGroup, auth *service.AuthService, oa
 	ax := &authxHandler{auth: auth, oauth: oauth}
 	secured.GET("/auth/totp", ax.totpStatus)
 	secured.POST("/auth/totp/setup", ax.totpSetup)
-	secured.POST("/auth/totp/activate", ax.totpActivate)
-	secured.POST("/auth/totp/disable", ax.totpDisable)
 	secured.GET("/auth/credentials", ax.getCredentials)
 	secured.PUT("/auth/credentials", ax.updateCredentials)
-	secured.PUT("/auth/password-login", ax.updatePasswordLogin)
 	secured.GET("/auth/identities", ax.identities)
+	secured.POST("/auth/totp/activate", ax.totpActivate)
+	secured.POST("/auth/totp/disable", ax.totpDisable)
+	secured.PUT("/auth/password-login", ax.updatePasswordLogin)
 	secured.DELETE("/auth/identities/:id", ax.unbindIdentity)
+	secured.POST("/auth/revoke-sessions", ax.revokeSessions)
 	secured.GET("/settings/oauth", func(c *gin.Context) { ax.getOAuthSettings(c, settings) })
 	secured.PUT("/settings/oauth", func(c *gin.Context) { ax.updateOAuthSettings(c, settings) })
 }
@@ -109,6 +109,7 @@ func registerOciCost(g *gin.RouterGroup, h *ociConfigHandler) {
 func registerOciTenantIAM(g *gin.RouterGroup, h *ociConfigHandler) {
 	g.GET("/oci-configs/:id/audit-events", h.getAuditEvents)
 	g.GET("/oci-configs/:id/audit-events/detail", h.getAuditEventDetail)
+	g.GET("/oci-configs/:id/domains", h.listIdentityDomains)
 	g.GET("/oci-configs/:id/users", h.listTenantUsers)
 	g.POST("/oci-configs/:id/users", h.createTenantUser)
 	g.GET("/oci-configs/:id/users/:userId", h.getTenantUserDetail)
@@ -15,6 +15,9 @@ func registerSettings(secured *gin.RouterGroup, settings *service.SettingService
 	secured.GET("/settings/telegram", st.getTelegram)
 	secured.PUT("/settings/telegram", st.updateTelegram)
 	secured.POST("/settings/telegram/test", st.testTelegram)
+	secured.GET("/settings/notify-channels", st.listNotifyChannels)
+	secured.PUT("/settings/notify-channels/:type", st.updateNotifyChannel)
+	secured.POST("/settings/notify-channels/:type/test", st.testNotifyChannel)
 	secured.GET("/settings/notify-events", st.getNotifyEvents)
 	secured.PUT("/settings/notify-events", st.updateNotifyEvents)
 	secured.GET("/settings/notify-templates", st.listNotifyTemplates)
@@ -32,4 +35,5 @@ func registerSettings(secured *gin.RouterGroup, settings *service.SettingService
 	secured.PUT("/proxies/:id", px.update)
 	secured.DELETE("/proxies/:id", px.remove)
 	secured.POST("/proxies/:id/probe", px.probe)
+	secured.PUT("/proxies/:id/tenants", px.setTenants)
 }
@@ -19,7 +19,7 @@ type settingsHandler struct {
 //
 // @Summary  返回脱敏后的 Telegram 配置，绝不回 token 明文
 // @Tags     设置
-// @Success  200 {object} map[string]any
+// @Success  200 {object} service.TelegramView
 // @Security BearerAuth
 // @Router   /api/v1/settings/telegram [get]
 func (h *settingsHandler) getTelegram(c *gin.Context) {
@@ -44,7 +44,7 @@ type updateTelegramRequest struct {
 // @Summary  保存配置并返回最新脱敏视图
 // @Tags     设置
 // @Param    body body updateTelegramRequest true "请求体"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} service.TelegramView
 // @Security BearerAuth
 // @Router   /api/v1/settings/telegram [put]
 func (h *settingsHandler) updateTelegram(c *gin.Context) {
@@ -80,11 +80,87 @@ func (h *settingsHandler) testTelegram(c *gin.Context) {
 	c.Status(http.StatusNoContent)
 }

+// listNotifyChannels 返回全部通知渠道的脱敏视图(webhook/ntfy/bark/smtp,不含 telegram)。
+//
+// @Summary  返回全部通知渠道的脱敏视图
+// @Tags     设置
+// @Success  200 {object} itemsResponse[service.NotifyChannelView]
+// @Security BearerAuth
+// @Router   /api/v1/settings/notify-channels [get]
+func (h *settingsHandler) listNotifyChannels(c *gin.Context) {
+	items, err := h.svc.NotifyChannels(c.Request.Context())
+	if err != nil {
+		respondError(c, err)
+		return
+	}
+	c.JSON(http.StatusOK, gin.H{"items": items})
+}
+
+// updateNotifyChannelRequest 是保存单渠道配置的请求体;
+// secret 承载渠道密文字段(ntfy token / bark device key / smtp 密码),
+// 缺省沿用已存值,空串清除。
+type updateNotifyChannelRequest struct {
+	Enabled      bool    `json:"enabled"`
+	URL          string  `json:"url"`
+	BodyTemplate string  `json:"bodyTemplate"`
+	Server       string  `json:"server"`
+	Topic        string  `json:"topic"`
+	Host         string  `json:"host"`
+	Port         int     `json:"port"`
+	Username     string  `json:"username"`
+	From         string  `json:"from"`
+	To           string  `json:"to"`
+	Secret       *string `json:"secret"`
+}
+
+// updateNotifyChannel 保存单渠道配置并返回全渠道最新视图;未知类型或校验失败返回 400。
+//
+// @Summary  保存单渠道配置并返回全渠道最新视图
+// @Tags     设置
+// @Param    type path string true "渠道类型(webhook/ntfy/bark/smtp)"
+// @Param    body body updateNotifyChannelRequest true "请求体"
+// @Success  200 {object} itemsResponse[service.NotifyChannelView]
+// @Security BearerAuth
+// @Router   /api/v1/settings/notify-channels/{type} [put]
+func (h *settingsHandler) updateNotifyChannel(c *gin.Context) {
+	var req updateNotifyChannelRequest
+	if err := c.ShouldBindJSON(&req); err != nil {
+		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
+		return
+	}
+	err := h.svc.UpdateNotifyChannel(c.Request.Context(), c.Param("type"), service.UpdateNotifyChannelInput{
+		Enabled: req.Enabled, URL: req.URL, BodyTemplate: req.BodyTemplate,
+		Server: req.Server, Topic: req.Topic, Host: req.Host, Port: req.Port,
+		Username: req.Username, From: req.From, To: req.To, Secret: req.Secret,
+	})
+	if err != nil {
+		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
+		return
+	}
+	h.listNotifyChannels(c)
+}
+
+// testNotifyChannel 用已保存配置向指定渠道同步发送测试消息。
+//
+// @Summary  向指定渠道发送测试消息
+// @Tags     设置
+// @Param    type path string true "渠道类型(webhook/ntfy/bark/smtp)"
+// @Success  204 "无内容"
+// @Security BearerAuth
+// @Router   /api/v1/settings/notify-channels/{type}/test [post]
+func (h *settingsHandler) testNotifyChannel(c *gin.Context) {
+	if err := h.notifier.SendChannelTest(c.Request.Context(), c.Param("type")); err != nil {
+		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
+		return
+	}
+	c.Status(http.StatusNoContent)
+}
+
 // getNotifyEvents 返回全部通知事件开关；键缺省（存量部署）视为开启。
 //
 // @Summary  返回全部通知事件开关
 // @Tags     设置
-// @Success  200 {object} map[string]any
+// @Success  200 {object} service.NotifyEventsView
 // @Security BearerAuth
 // @Router   /api/v1/settings/notify-events [get]
 func (h *settingsHandler) getNotifyEvents(c *gin.Context) {
@@ -102,7 +178,7 @@ func (h *settingsHandler) getNotifyEvents(c *gin.Context) {
 // @Summary  全量保存五个事件开关并返回最新值
 // @Tags     设置
 // @Param    body body service.NotifyEventsView true "请求体"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} service.NotifyEventsView
 // @Security BearerAuth
 // @Router   /api/v1/settings/notify-events [put]
 func (h *settingsHandler) updateNotifyEvents(c *gin.Context) {
@@ -122,7 +198,7 @@ func (h *settingsHandler) updateNotifyEvents(c *gin.Context) {
 //
 // @Summary  返回全部通知模板
 // @Tags     设置
-// @Success  200 {object} map[string]any
+// @Success  200 {object} itemsResponse[service.NotifyTemplateView]
 // @Security BearerAuth
 // @Router   /api/v1/settings/notify-templates [get]
 func (h *settingsHandler) listNotifyTemplates(c *gin.Context) {
@@ -187,7 +263,7 @@ func (h *settingsHandler) testNotifyTemplate(c *gin.Context) {
 //
 // @Summary  返回任务行为设置
 // @Tags     设置
-// @Success  200 {object} map[string]any
+// @Success  200 {object} service.TaskSettingsView
 // @Security BearerAuth
 // @Router   /api/v1/settings/task [get]
 func (h *settingsHandler) getTaskSettings(c *gin.Context) {
@@ -204,7 +280,7 @@ func (h *settingsHandler) getTaskSettings(c *gin.Context) {
 // @Summary  保存任务行为设置并返回最新值
 // @Tags     设置
 // @Param    body body service.TaskSettingsView true "请求体"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} service.TaskSettingsView
 // @Security BearerAuth
 // @Router   /api/v1/settings/task [put]
 func (h *settingsHandler) updateTaskSettings(c *gin.Context) {
@@ -228,7 +304,7 @@ func (h *settingsHandler) updateTaskSettings(c *gin.Context) {
 //
 // @Summary  返回安全设置
 // @Tags     设置
-// @Success  200 {object} map[string]any
+// @Success  200 {object} service.SecuritySettings
 // @Security BearerAuth
 // @Router   /api/v1/settings/security [get]
 func (h *settingsHandler) getSecurity(c *gin.Context) {
@@ -245,7 +321,7 @@ func (h *settingsHandler) getSecurity(c *gin.Context) {
 // @Summary  保存安全设置并返回最新值
 // @Tags     设置
 // @Param    body body service.SecuritySettings true "请求体"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} service.SecuritySettings
 // @Security BearerAuth
 // @Router   /api/v1/settings/security [put]
 func (h *settingsHandler) updateSecurity(c *gin.Context) {
@@ -4,12 +4,14 @@ import (
 	"net/http"

 	"github.com/gin-gonic/gin"
+
+	_ "oci-portal/internal/oci" // swagger 注解引用
 )

 // @Summary  订阅列表
 // @Tags     租户配置
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {array} oci.SubscriptionInfo
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/subscriptions [get]
 func (h *ociConfigHandler) subscriptions(c *gin.Context) {
@@ -29,7 +31,7 @@ func (h *ociConfigHandler) subscriptions(c *gin.Context) {
 // @Tags     租户配置
 // @Param    id path int true "配置 ID"
 // @Param    subscriptionId path string true "subscriptionId"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} oci.SubscriptionDetail
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/subscriptions/{subscriptionId} [get]
 func (h *ociConfigHandler) subscriptionDetail(c *gin.Context) {
@@ -48,7 +50,7 @@ func (h *ociConfigHandler) subscriptionDetail(c *gin.Context) {
 // @Summary  限额服务列表
 // @Tags     租户配置
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {array} oci.LimitService
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/limits/services [get]
 func (h *ociConfigHandler) limitServices(c *gin.Context) {
@@ -0,0 +1,176 @@
+package api
+
+import (
+	"encoding/json"
+	"time"
+
+	"oci-portal/internal/model"
+	"oci-portal/internal/oci"
+	"oci-portal/internal/service"
+)
+
+// 本文件的类型仅供 swagger 文档渲染(@Success/@Failure 注解引用),
+// 运行时代码不使用;字段与各 handler 返回的 gin.H 外壳保持一致。
+
+// errorResponse 是统一错误响应外壳。
+type errorResponse struct {
+	Error string `json:"error"`
+}
+
+// itemsResponse 是 {"items": [...]} 列表外壳。
+type itemsResponse[T any] struct {
+	Items []T `json:"items"`
+}
+
+// pagedResponse 是 {"items": [...], "total": n} 分页外壳。
+type pagedResponse[T any] struct {
+	Items []T   `json:"items"`
+	Total int64 `json:"total"`
+}
+
+// itemResponse 是 {"item": {...}} 单项外壳。
+type itemResponse[T any] struct {
+	Item T `json:"item"`
+}
+
+// triggerResponse 是任务触发受理响应。
+type triggerResponse struct {
+	Triggered bool `json:"triggered"`
+}
+
+// tokenResponse 是登录 / 换发令牌响应。
+type tokenResponse struct {
+	Token     string    `json:"token"`
+	ExpiresAt time.Time `json:"expiresAt"`
+}
+
+// totpRequiredResponse 是密码通过但需两步验证码的 428 响应。
+type totpRequiredResponse struct {
+	Error        string `json:"error"`
+	TotpRequired bool   `json:"totpRequired"`
+}
+
+// enabledResponse 是布尔开关查询响应。
+type enabledResponse struct {
+	Enabled bool `json:"enabled"`
+}
+
+// totpSetupResponse 是 TOTP 初始化响应。
+type totpSetupResponse struct {
+	Secret     string `json:"secret"`
+	OtpauthUri string `json:"otpauthUri"`
+}
+
+// credentialsResponse 是当前登录账号信息。
+type credentialsResponse struct {
+	Username              string `json:"username"`
+	PasswordLoginDisabled bool   `json:"passwordLoginDisabled"`
+}
+
+// oauthProvidersResponse 是外部登录 provider 列表。
+type oauthProvidersResponse struct {
+	Providers             []service.ProviderInfo `json:"providers"`
+	PasswordLoginDisabled bool                   `json:"passwordLoginDisabled"`
+}
+
+// urlResponse 是跳转地址响应。
+type urlResponse struct {
+	URL string `json:"url"`
+}
+
+// aboutResponse 是「设置 · 关于」页构建与运行信息。
+type aboutResponse struct {
+	Version       string         `json:"version"`
+	BuildTime     string         `json:"buildTime"`
+	GoVersion     string         `json:"goVersion"`
+	Platform      string         `json:"platform"`
+	StartedAt     string         `json:"startedAt"`
+	UptimeSeconds int64          `json:"uptimeSeconds"`
+	Resources     aboutResources `json:"resources"`
+}
+
+// aboutResources 是进程资源占用快照。
+type aboutResources struct {
+	CpuAvgPercent float64 `json:"cpuAvgPercent"`
+	NumCpu        int     `json:"numCpu"`
+	Goroutines    int     `json:"goroutines"`
+	MemHeapBytes  uint64  `json:"memHeapBytes"`
+	MemSysBytes   uint64  `json:"memSysBytes"`
+	DbEngine      string  `json:"dbEngine"`
+	DbBytes       int64   `json:"dbBytes"`
+}
+
+// publicIpResponse 是换绑公网 IP 结果。
+type publicIpResponse struct {
+	PublicIp string `json:"publicIp"`
+}
+
+// ipv6AddressResponse 是实例新增 IPv6 结果。
+type ipv6AddressResponse struct {
+	Ipv6Address string `json:"ipv6Address"`
+}
+
+// addressResponse 是 VNIC 新增 IPv6 结果。
+type addressResponse struct {
+	Address string `json:"address"`
+}
+
+// ipv6StepsResponse 是 VCN 开启 IPv6 的分步结果。
+type ipv6StepsResponse struct {
+	Steps []oci.IPv6Step `json:"steps"`
+}
+
+// configChangesResponse 是租户配置校验 / 更新结果(config 与字段变更对照)。
+type configChangesResponse struct {
+	Config  *model.OciConfig `json:"config"`
+	Changes service.Changes  `json:"changes"`
+}
+
+// aiKeyCreateResponse 是 AI 密钥创建结果;key 为明文,仅本次返回。
+type aiKeyCreateResponse struct {
+	Key  string      `json:"key"`
+	Item model.AiKey `json:"item"`
+}
+
+// usedByResponse 是代理关联租户结果(关联数)。
+type usedByResponse struct {
+	UsedBy int64 `json:"usedBy"`
+}
+
+// rawDetailResponse 是审计事件原文响应。
+type rawDetailResponse struct {
+	Raw json.RawMessage `json:"raw"`
+}
+
+// passwordResponse 是租户用户重置密码结果(一次性明文)。
+type passwordResponse struct {
+	Password string `json:"password"`
+}
+
+// deletedTotpResponse 是清除 MFA 设备结果。
+type deletedTotpResponse struct {
+	DeletedTotpDevices int `json:"deletedTotpDevices"`
+}
+
+// deletedApiKeysResponse 是清理用户 API Key 结果。
+type deletedApiKeysResponse struct {
+	DeletedApiKeys int `json:"deletedApiKeys"`
+}
+
+// createInstancesResponse 是批量创建实例结果;部分成功仍 201,errors 为逐台失败信息。
+type createInstancesResponse struct {
+	Instances []oci.Instance `json:"instances"`
+	Errors    []string       `json:"errors"`
+}
+
+// logWebhookStatusResponse 是日志回传 webhook 状态。
+type logWebhookStatusResponse struct {
+	Exists  bool                    `json:"exists"`
+	Webhook *service.LogWebhookInfo `json:"webhook,omitempty"`
+}
+
+// webConsoleSessionResponse 是网页控制台会话创建结果。
+type webConsoleSessionResponse struct {
+	SessionId string `json:"sessionId"`
+	Type      string `json:"type"`
+}
@@ -6,6 +6,7 @@ import (

 	"github.com/gin-gonic/gin"

+	_ "oci-portal/internal/model" // swagger 注解引用
 	"oci-portal/internal/service"
 )

@@ -18,7 +19,7 @@ type systemLogHandler struct {
 //
 // @Summary  系统操作日志列表
 // @Tags     设置
-// @Success  200 {object} map[string]any "items 与 total"
+// @Success  200 {object} pagedResponse[model.SystemLog] "items 与 total"
 // @Security BearerAuth
 // @Router   /api/v1/system-logs [get]
 func (h *systemLogHandler) list(c *gin.Context) {
@@ -2,11 +2,13 @@ package api

 import (
 	"encoding/json"
+	"errors"
 	"net/http"
 	"strconv"

 	"github.com/gin-gonic/gin"

+	_ "oci-portal/internal/model" // swagger 注解引用
 	"oci-portal/internal/service"
 )

@@ -32,7 +34,7 @@ type updateTaskRequest struct {
 // @Summary  创建任务
 // @Tags     任务与日志回传
 // @Param    body body createTaskRequest true "任务定义(类型/cron/payload)"
-// @Success  201 {object} map[string]any
+// @Success  201 {object} model.Task
 // @Security BearerAuth
 // @Router   /api/v1/tasks [post]
 func (h *taskHandler) create(c *gin.Context) {
@@ -56,7 +58,7 @@ func (h *taskHandler) create(c *gin.Context) {

 // @Summary  任务列表
 // @Tags     任务与日志回传
-// @Success  200 {array} map[string]any
+// @Success  200 {array} model.Task
 // @Security BearerAuth
 // @Router   /api/v1/tasks [get]
 func (h *taskHandler) list(c *gin.Context) {
@@ -71,7 +73,7 @@ func (h *taskHandler) list(c *gin.Context) {
 // @Summary  任务详情
 // @Tags     任务与日志回传
 // @Param    id path int true "任务 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} model.Task
 // @Security BearerAuth
 // @Router   /api/v1/tasks/{id} [get]
 func (h *taskHandler) get(c *gin.Context) {
@@ -91,7 +93,7 @@ func (h *taskHandler) get(c *gin.Context) {
 // @Tags     任务与日志回传
 // @Param    id path int true "任务 ID"
 // @Param    body body updateTaskRequest true "可更新字段"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} model.Task
 // @Security BearerAuth
 // @Router   /api/v1/tasks/{id} [put]
 func (h *taskHandler) update(c *gin.Context) {
@@ -138,7 +140,7 @@ func (h *taskHandler) remove(c *gin.Context) {
 // @Summary  任务执行日志
 // @Tags     任务与日志回传
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {array} model.TaskLog
 // @Security BearerAuth
 // @Router   /api/v1/tasks/{id}/logs [get]
 func (h *taskHandler) logs(c *gin.Context) {
@@ -155,10 +157,11 @@ func (h *taskHandler) logs(c *gin.Context) {
 	c.JSON(http.StatusOK, logs)
 }

-// @Summary  立即执行任务
+// @Summary  立即执行任务(异步触发,结果经任务日志轮询获取)
 // @Tags     任务与日志回传
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  202 {object} triggerResponse
+// @Failure  409 {object} errorResponse "任务正在执行中"
 // @Security BearerAuth
 // @Router   /api/v1/tasks/{id}/run [post]
 func (h *taskHandler) run(c *gin.Context) {
@@ -166,10 +169,13 @@ func (h *taskHandler) run(c *gin.Context) {
 	if !ok {
 		return
 	}
-	entry, err := h.svc.RunTaskNow(c.Request.Context(), id)
-	if err != nil {
+	if err := h.svc.TriggerTask(c.Request.Context(), id); err != nil {
+		if errors.Is(err, service.ErrTaskRunning) {
+			c.JSON(http.StatusConflict, gin.H{"error": err.Error()})
+			return
+		}
 		respondError(c, err)
 		return
 	}
-	c.JSON(http.StatusOK, entry)
+	c.JSON(http.StatusAccepted, gin.H{"triggered": true})
 }
@@ -18,7 +18,7 @@ import (
 // @Tags     计算
 // @Param    id path int true "配置 ID"
 // @Param    instanceId path string true "instanceId"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} oci.InstanceTraffic
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/instances/{instanceId}/traffic [get]
 func (h *ociConfigHandler) instanceTraffic(c *gin.Context) {
@@ -38,7 +38,7 @@ func (h *ociConfigHandler) instanceTraffic(c *gin.Context) {
 // @Summary  配置成本快照
 // @Tags     成本
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {array} oci.CostItem
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/costs [get]
 func (h *ociConfigHandler) costs(c *gin.Context) {
@@ -67,13 +67,15 @@ func (h *ociConfigHandler) costs(c *gin.Context) {

 // ---- 租户审计日志 ----

-// getAuditEvents 实时查询租户 OCI 审计事件:hours 首查(缺省 24);
-// start/end/page 为截断后的同窗续查;窗口越界或非法返回 400。
+// getAuditEvents 批式懒加载查询审计事件:cursor 为空自当前时刻首查,
+// 非空从上次响应游标继续向更早回溯;limit 单批目标条数(缺省 100,上限 200)。
 //
-// @Summary  实时查询租户 OCI 审计事件:hours 首查
+// @Summary  批式懒加载查询租户 OCI 审计事件
 // @Tags     租户 IAM
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Param    cursor query string false "续查游标(上次响应原样带回)"
+// @Param    limit query int false "单批目标条数,缺省 100,上限 200"
+// @Success  200 {object} service.AuditEventsView
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/audit-events [get]
 func (h *ociConfigHandler) getAuditEvents(c *gin.Context) {
@@ -81,13 +83,10 @@ func (h *ociConfigHandler) getAuditEvents(c *gin.Context) {
 	if !ok {
 		return
 	}
-	hours, _ := strconv.Atoi(c.DefaultQuery("hours", "24"))
-	q := service.AuditQuery{
-		Region: c.Query("region"), Hours: hours,
-		Start: c.Query("start"), End: c.Query("end"), Page: c.Query("page"),
-	}
+	limit, _ := strconv.Atoi(c.Query("limit"))
+	q := service.AuditQuery{Region: c.Query("region"), Cursor: c.Query("cursor"), Limit: limit}
 	result, err := h.svc.AuditEvents(c.Request.Context(), id, q)
-	if errors.Is(err, service.ErrInvalidAuditHours) || errors.Is(err, service.ErrInvalidAuditWindow) {
+	if errors.Is(err, service.ErrInvalidAuditCursor) {
 		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
 		return
 	}
@@ -104,7 +103,7 @@ func (h *ociConfigHandler) getAuditEvents(c *gin.Context) {
 // @Summary  按 eventId 取回原始事件 JSON:缓存命中秒开,
 // @Tags     租户 IAM
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} rawDetailResponse "raw 为事件原文 JSON"
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/audit-events/detail [get]
 func (h *ociConfigHandler) getAuditEventDetail(c *gin.Context) {
@@ -143,10 +142,30 @@ type createTenantUserRequest struct {
 	AddToAdminGroup  bool `json:"addToAdminGroup"`
 }

+// @Summary  租户身份域列表
+// @Tags     租户 IAM
+// @Param    id path int true "配置 ID"
+// @Success  200 {array} oci.IdentityDomain
+// @Security BearerAuth
+// @Router   /api/v1/oci-configs/{id}/domains [get]
+func (h *ociConfigHandler) listIdentityDomains(c *gin.Context) {
+	id, ok := pathID(c)
+	if !ok {
+		return
+	}
+	domains, err := h.svc.IdentityDomains(c.Request.Context(), id)
+	if err != nil {
+		respondError(c, err)
+		return
+	}
+	c.JSON(http.StatusOK, domains)
+}
+
 // @Summary  租户 IAM 用户列表
 // @Tags     租户 IAM
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Param    domainId query string false "身份域 OCID(缺省 Default 域)"
+// @Success  200 {array} oci.TenantUser
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/users [get]
 func (h *ociConfigHandler) listTenantUsers(c *gin.Context) {
@@ -154,7 +173,7 @@ func (h *ociConfigHandler) listTenantUsers(c *gin.Context) {
 	if !ok {
 		return
 	}
-	users, err := h.svc.TenantUsers(c.Request.Context(), id)
+	users, err := h.svc.TenantUsers(c.Request.Context(), id, c.Query("domainId"))
 	if err != nil {
 		respondError(c, err)
 		return
@@ -165,8 +184,9 @@ func (h *ociConfigHandler) listTenantUsers(c *gin.Context) {
 // @Summary  租户用户详情
 // @Tags     租户 IAM
 // @Param    id path int true "配置 ID"
+// @Param    domainId query string false "身份域 OCID(缺省 Default 域)"
 // @Param    userId path string true "userId"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} oci.TenantUserDetail
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/users/{userId} [get]
 func (h *ociConfigHandler) getTenantUserDetail(c *gin.Context) {
@@ -174,7 +194,7 @@ func (h *ociConfigHandler) getTenantUserDetail(c *gin.Context) {
 	if !ok {
 		return
 	}
-	detail, err := h.svc.TenantUserDetail(c.Request.Context(), id, c.Param("userId"))
+	detail, err := h.svc.TenantUserDetail(c.Request.Context(), id, c.Query("domainId"), c.Param("userId"))
 	if err != nil {
 		respondError(c, err)
 		return
@@ -185,8 +205,9 @@ func (h *ociConfigHandler) getTenantUserDetail(c *gin.Context) {
 // @Summary  创建租户 IAM 用户
 // @Tags     租户 IAM
 // @Param    id path int true "配置 ID"
+// @Param    domainId query string false "身份域 OCID(缺省 Default 域)"
 // @Param    body body createTenantUserRequest true "请求体"
-// @Success  201 {object} map[string]any
+// @Success  201 {object} oci.TenantUser
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/users [post]
 func (h *ociConfigHandler) createTenantUser(c *gin.Context) {
@@ -199,7 +220,7 @@ func (h *ociConfigHandler) createTenantUser(c *gin.Context) {
 		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
 		return
 	}
-	user, err := h.svc.CreateTenantUser(c.Request.Context(), id, oci.CreateTenantUserInput{
+	user, err := h.svc.CreateTenantUser(c.Request.Context(), id, c.Query("domainId"), oci.CreateTenantUserInput{
 		Name:             req.Name,
 		Description:      req.Description,
 		Email:            req.Email,
@@ -229,9 +250,10 @@ type updateTenantUserRequest struct {
 // @Summary  更新租户 IAM 用户
 // @Tags     租户 IAM
 // @Param    id path int true "配置 ID"
+// @Param    domainId query string false "身份域 OCID(缺省 Default 域)"
 // @Param    userId path string true "userId"
 // @Param    body body updateTenantUserRequest true "请求体"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} oci.TenantUser
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/users/{userId} [put]
 func (h *ociConfigHandler) updateTenantUser(c *gin.Context) {
@@ -244,7 +266,7 @@ func (h *ociConfigHandler) updateTenantUser(c *gin.Context) {
 		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
 		return
 	}
-	user, err := h.svc.UpdateTenantUser(c.Request.Context(), id, c.Param("userId"), oci.UpdateTenantUserInput{
+	user, err := h.svc.UpdateTenantUser(c.Request.Context(), id, c.Query("domainId"), c.Param("userId"), oci.UpdateTenantUserInput{
 		Description:      req.Description,
 		Email:            req.Email,
 		GivenName:        req.GivenName,
@@ -262,6 +284,7 @@ func (h *ociConfigHandler) updateTenantUser(c *gin.Context) {
 // @Summary  删除租户 IAM 用户
 // @Tags     租户 IAM
 // @Param    id path int true "配置 ID"
+// @Param    domainId query string false "身份域 OCID(缺省 Default 域)"
 // @Param    userId path string true "userId"
 // @Success  204 "无内容"
 // @Security BearerAuth
@@ -271,7 +294,7 @@ func (h *ociConfigHandler) deleteTenantUser(c *gin.Context) {
 	if !ok {
 		return
 	}
-	if err := h.svc.DeleteTenantUser(c.Request.Context(), id, c.Param("userId")); err != nil {
+	if err := h.svc.DeleteTenantUser(c.Request.Context(), id, c.Query("domainId"), c.Param("userId")); err != nil {
 		respondError(c, err)
 		return
 	}
@@ -281,8 +304,9 @@ func (h *ociConfigHandler) deleteTenantUser(c *gin.Context) {
 // @Summary  重置租户用户密码
 // @Tags     租户 IAM
 // @Param    id path int true "配置 ID"
+// @Param    domainId query string false "身份域 OCID(缺省 Default 域)"
 // @Param    userId path string true "userId"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} passwordResponse "一次性明文密码"
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/users/{userId}/reset-password [post]
 func (h *ociConfigHandler) resetTenantUserPassword(c *gin.Context) {
@@ -290,7 +314,7 @@ func (h *ociConfigHandler) resetTenantUserPassword(c *gin.Context) {
 	if !ok {
 		return
 	}
-	password, err := h.svc.ResetTenantUserPassword(c.Request.Context(), id, c.Param("userId"))
+	password, err := h.svc.ResetTenantUserPassword(c.Request.Context(), id, c.Query("domainId"), c.Param("userId"))
 	if err != nil {
 		respondError(c, err)
 		return
@@ -301,8 +325,9 @@ func (h *ociConfigHandler) resetTenantUserPassword(c *gin.Context) {
 // @Summary  清除用户 MFA 设备
 // @Tags     租户 IAM
 // @Param    id path int true "配置 ID"
+// @Param    domainId query string false "身份域 OCID(缺省 Default 域)"
 // @Param    userId path string true "userId"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} deletedTotpResponse
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/users/{userId}/mfa-devices [delete]
 func (h *ociConfigHandler) deleteTenantUserMfa(c *gin.Context) {
@@ -310,7 +335,7 @@ func (h *ociConfigHandler) deleteTenantUserMfa(c *gin.Context) {
 	if !ok {
 		return
 	}
-	deleted, err := h.svc.DeleteTenantUserMfa(c.Request.Context(), id, c.Param("userId"))
+	deleted, err := h.svc.DeleteTenantUserMfa(c.Request.Context(), id, c.Query("domainId"), c.Param("userId"))
 	if err != nil {
 		respondError(c, err)
 		return
@@ -322,7 +347,7 @@ func (h *ociConfigHandler) deleteTenantUserMfa(c *gin.Context) {
 // @Tags     租户 IAM
 // @Param    id path int true "配置 ID"
 // @Param    userId path string true "userId"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} deletedApiKeysResponse
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/users/{userId}/api-keys [delete]
 func (h *ociConfigHandler) deleteTenantUserApiKeys(c *gin.Context) {
@@ -344,7 +369,8 @@ func (h *ociConfigHandler) deleteTenantUserApiKeys(c *gin.Context) {
 // @Summary  ---- 域通知收件人与密码策略 ----
 // @Tags     租户 IAM
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Param    domainId query string false "身份域 OCID(缺省 Default 域)"
+// @Success  200 {object} oci.NotificationRecipients
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/notification-recipients [get]
 func (h *ociConfigHandler) getNotificationRecipients(c *gin.Context) {
@@ -352,7 +378,7 @@ func (h *ociConfigHandler) getNotificationRecipients(c *gin.Context) {
 	if !ok {
 		return
 	}
-	recipients, err := h.svc.NotificationRecipients(c.Request.Context(), id)
+	recipients, err := h.svc.NotificationRecipients(c.Request.Context(), id, c.Query("domainId"))
 	if err != nil {
 		respondError(c, err)
 		return
@@ -363,8 +389,9 @@ func (h *ociConfigHandler) getNotificationRecipients(c *gin.Context) {
 // @Summary  更新通知收件人
 // @Tags     租户 IAM
 // @Param    id path int true "配置 ID"
+// @Param    domainId query string false "身份域 OCID(缺省 Default 域)"
 // @Param    body body object true "请求体(见接口说明)"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} oci.NotificationRecipients
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/notification-recipients [put]
 func (h *ociConfigHandler) updateNotificationRecipients(c *gin.Context) {
@@ -379,7 +406,7 @@ func (h *ociConfigHandler) updateNotificationRecipients(c *gin.Context) {
 		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
 		return
 	}
-	recipients, err := h.svc.UpdateNotificationRecipients(c.Request.Context(), id, req.Recipients)
+	recipients, err := h.svc.UpdateNotificationRecipients(c.Request.Context(), id, c.Query("domainId"), req.Recipients)
 	if err != nil {
 		respondError(c, err)
 		return
@@ -390,7 +417,8 @@ func (h *ociConfigHandler) updateNotificationRecipients(c *gin.Context) {
 // @Summary  密码策略列表
 // @Tags     租户 IAM
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Param    domainId query string false "身份域 OCID(缺省 Default 域)"
+// @Success  200 {array} oci.PasswordPolicyInfo
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/password-policies [get]
 func (h *ociConfigHandler) listPasswordPolicies(c *gin.Context) {
@@ -398,7 +426,7 @@ func (h *ociConfigHandler) listPasswordPolicies(c *gin.Context) {
 	if !ok {
 		return
 	}
-	policies, err := h.svc.PasswordPolicies(c.Request.Context(), id)
+	policies, err := h.svc.PasswordPolicies(c.Request.Context(), id, c.Query("domainId"))
 	if err != nil {
 		respondError(c, err)
 		return
@@ -409,9 +437,10 @@ func (h *ociConfigHandler) listPasswordPolicies(c *gin.Context) {
 // @Summary  更新密码策略
 // @Tags     租户 IAM
 // @Param    id path int true "配置 ID"
+// @Param    domainId query string false "身份域 OCID(缺省 Default 域)"
 // @Param    policyId path string true "policyId"
 // @Param    body body object true "请求体(见接口说明)"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} oci.PasswordPolicyInfo
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/password-policies/{policyId} [put]
 func (h *ociConfigHandler) updatePasswordPolicy(c *gin.Context) {
@@ -428,7 +457,7 @@ func (h *ociConfigHandler) updatePasswordPolicy(c *gin.Context) {
 		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
 		return
 	}
-	policy, err := h.svc.UpdatePasswordPolicy(c.Request.Context(), id, c.Param("policyId"), oci.UpdatePasswordPolicyInput{
+	policy, err := h.svc.UpdatePasswordPolicy(c.Request.Context(), id, c.Query("domainId"), c.Param("policyId"), oci.UpdatePasswordPolicyInput{
 		PasswordExpiresAfter:  req.PasswordExpiresAfter,
 		MinLength:             req.MinLength,
 		NumPasswordsInHistory: req.NumPasswordsInHistory,
@@ -443,7 +472,8 @@ func (h *ociConfigHandler) updatePasswordPolicy(c *gin.Context) {
 // @Summary  身份域设置
 // @Tags     租户 IAM
 // @Param    id path int true "配置 ID"
-// @Success  200 {object} map[string]any
+// @Param    domainId query string false "身份域 OCID(缺省 Default 域)"
+// @Success  200 {object} oci.IdentitySettingInfo
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/identity-settings [get]
 func (h *ociConfigHandler) getIdentitySetting(c *gin.Context) {
@@ -451,7 +481,7 @@ func (h *ociConfigHandler) getIdentitySetting(c *gin.Context) {
 	if !ok {
 		return
 	}
-	setting, err := h.svc.IdentitySetting(c.Request.Context(), id)
+	setting, err := h.svc.IdentitySetting(c.Request.Context(), id, c.Query("domainId"))
 	if err != nil {
 		respondError(c, err)
 		return
@@ -462,8 +492,9 @@ func (h *ociConfigHandler) getIdentitySetting(c *gin.Context) {
 // @Summary  更新身份域设置
 // @Tags     租户 IAM
 // @Param    id path int true "配置 ID"
+// @Param    domainId query string false "身份域 OCID(缺省 Default 域)"
 // @Param    body body object true "请求体(见接口说明)"
-// @Success  200 {object} map[string]any
+// @Success  200 {object} oci.IdentitySettingInfo
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/identity-settings [put]
 func (h *ociConfigHandler) updateIdentitySetting(c *gin.Context) {
@@ -478,7 +509,7 @@ func (h *ociConfigHandler) updateIdentitySetting(c *gin.Context) {
 		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
 		return
 	}
-	setting, err := h.svc.UpdateIdentitySetting(c.Request.Context(), id, *req.PrimaryEmailRequired)
+	setting, err := h.svc.UpdateIdentitySetting(c.Request.Context(), id, c.Query("domainId"), *req.PrimaryEmailRequired)
 	if err != nil {
 		respondError(c, err)
 		return
@@ -29,7 +29,7 @@ type createConsoleSessionRequest struct {
 // @Param    id path int true "配置 ID"
 // @Param    instanceId path string true "实例 OCID"
 // @Param    body body object true "{type: serial|vnc}"
-// @Success  200 {object} map[string]any "sessionId 与 wsPath"
+// @Success  201 {object} webConsoleSessionResponse "sessionId 与 type"
 // @Security BearerAuth
 // @Router   /api/v1/oci-configs/{id}/instances/{instanceId}/console-sessions [post]
 func (h *consoleHandler) create(c *gin.Context) {
@@ -79,7 +79,7 @@ var wsUpgrader = websocket.Upgrader{
 // @Success  101 "升级为 WebSocket"
 // @Router   /api/v1/console-sessions/{sessionId}/ws [get]
 func (h *consoleHandler) ws(c *gin.Context) {
-	if _, err := h.auth.ParseToken(c.Query("token")); err != nil {
+	if _, err := h.auth.ParseToken(c.Request.Context(), c.Query("token")); err != nil {
 		c.JSON(http.StatusUnauthorized, gin.H{"error": "invalid or expired token"})
 		return
 	}
@@ -53,7 +53,7 @@ type webhookHandler struct {
 // @Param    secret path string true "面板生成的回传密钥(鉴权凭据)"
 // @Param    body body object true "ONS 消息原文(SubscriptionConfirmation / Notification)"
 // @Success  200 "已受理"
-// @Failure  403 {object} map[string]string "时间戳超窗或证书源非 Oracle 域"
+// @Failure  403 {object} errorResponse "时间戳超窗或证书源非 Oracle 域"
 // @Failure  404 "secret 无效(不暴露端点存在性)"
 // @Router   /api/v1/webhooks/oci-logs/{secret} [post]
 func (h *webhookHandler) handle(c *gin.Context) {
@@ -2,6 +2,9 @@ package database

 import (
 	"fmt"
+	"log"
+	"os"
+	"time"

 	"github.com/glebarez/sqlite"
 	"gorm.io/driver/mysql"
@@ -12,6 +15,18 @@ import (
 	"oci-portal/internal/model"
 )

+// newDBLogger 构造脱敏的 GORM 日志器(S-08):SQL 一律参数化输出,
+// 实参(密码哈希 / token / 任务 payload)不落日志;业务正常路径的
+// record-not-found 不打 SQL。
+func newDBLogger() logger.Interface {
+	return logger.New(log.New(os.Stdout, "\r\n", log.LstdFlags), logger.Config{
+		SlowThreshold:             200 * time.Millisecond,
+		LogLevel:                  logger.Warn,
+		IgnoreRecordNotFoundError: true,
+		ParameterizedQueries:      true,
+	})
+}
+
 // Open 按驱动打开数据库并自动迁移全部模型。
 // driver 取值 sqlite(默认)/mysql/postgres;sqlite 用 path,其余用 dsn。
 func Open(driver, dsn, path string) (*gorm.DB, error) {
@@ -20,7 +35,7 @@ func Open(driver, dsn, path string) (*gorm.DB, error) {
 		return nil, err
 	}
 	db, err := gorm.Open(dialector, &gorm.Config{
-		Logger: logger.Default.LogMode(logger.Warn),
+		Logger: newDBLogger(),
 	})
 	if err != nil {
 		return nil, fmt.Errorf("open %s database: %w", dialector.Name(), err)
@@ -53,7 +68,7 @@ func autoMigrate(db *gorm.DB) error {
 		&model.CheckSnapshot{}, &model.CostSnapshot{},
 		&model.RegionCache{}, &model.CompartmentCache{}, &model.Setting{},
 		&model.SystemLog{}, &model.LogEvent{}, &model.Proxy{},
-		&model.AiKey{}, &model.AiChannel{}, &model.AiModelCache{}, &model.AiCallLog{},
+		&model.AiKey{}, &model.AiChannel{}, &model.AiModelCache{}, &model.AiModelBlacklist{}, &model.AiCallLog{},
 		&model.AiContentLog{},
 	)
 }
@@ -10,10 +10,11 @@ const (
 	AccountTypeUnknown = "unknown"
 )

-// 测活状态取值。
+// 测活状态取值;suspended 来自账户能力接口的云端暂停标记(区别于失联)。
 const (
 	AliveStatusAlive     = "alive"
 	AliveStatusDead      = "dead"
+	AliveStatusSuspended = "suspended"
 	AliveStatusUnknown   = "unknown"
 )

@@ -65,6 +66,9 @@ type User struct {
 	PasswordHash string `json:"-"`
 	// TOTP 共享密钥 AES-GCM 密文;空串表示未启用两步验证
 	TotpSecretEnc string `json:"-"`
+	// TokenVersion 是 JWT 版本号:凭据/TOTP/外部身份/登录策略变更或撤销会话时
+	// 原子递增,旧版本令牌随即失效(存量令牌无 ver 视为 0,与零值兼容)
+	TokenVersion uint      `json:"-"`
 	CreatedAt    time.Time `json:"createdAt"`
 	UpdatedAt    time.Time `json:"updatedAt"`
 }
@@ -237,6 +241,8 @@ type AiKey struct {
 	KeyHash string `gorm:"uniqueIndex;size:64" json:"-"`
 	Tail    string `gorm:"size:4" json:"tail"`
 	Group   string `gorm:"column:key_group;size:32" json:"group"`
+	// Models 是模型白名单(精确匹配模型名);空 = 不限,非空时网关仅放行列表内模型
+	Models     []string   `gorm:"serializer:json;type:text" json:"models"`
 	Enabled    bool       `json:"enabled"`
 	LastUsedAt *time.Time `json:"lastUsedAt"`
 	// ContentLogUntil 是内容日志开启截止时间;nil = 永关(默认,红线),
@@ -267,7 +273,7 @@ type AiChannel struct {
 }

 // AiModelCache 是渠道区域的可用模型缓存(整渠道覆盖式同步)。
-// Capability 为 CHAT / EMBEDDING;存量空串视为 CHAT(加列前只同步对话模型)。
+// Capability 为 CHAT / EMBEDDING / RERANK / TTS;存量空串视为 CHAT(加列前只同步对话模型)。
 type AiModelCache struct {
 	ID         uint      `gorm:"primaryKey" json:"-"`
 	ChannelID  uint      `gorm:"index" json:"channelId"`
@@ -282,6 +288,14 @@ type AiModelCache struct {
 	RetiredAt    *time.Time `json:"retiredAt"`
 }

+// AiModelBlacklist 是全局模型黑名单(按模型名):加入即从全部渠道缓存删除该模型,
+// 同步与探测过滤不再入库;适用于微调基座等实际不可按需调用的模型,由用户手动维护。
+type AiModelBlacklist struct {
+	ID        uint      `gorm:"primaryKey" json:"id"`
+	Name      string    `gorm:"size:96;uniqueIndex" json:"name"`
+	CreatedAt time.Time `json:"createdAt"`
+}
+
 // AiCallLog 是网关调用日志:仅元数据与 token 用量,绝不记录 prompt / 响应正文。
 type AiCallLog struct {
 	ID          uint   `gorm:"primaryKey" json:"id"`
@@ -1,9 +1,16 @@
 package oci

 import (
+	"context"
+	"encoding/json"
+	"fmt"
+	"io"
+	"net/http"
 	"strings"
 	"time"

+	"github.com/oracle/oci-go-sdk/v65/common"
+
 	"oci-portal/internal/model"
 )

@@ -19,6 +26,89 @@ type AccountProfile struct {
 	PromotionExpires *time.Time
 }

+// AccountCapabilities 是账户能力接口的关键字段。来源为 Limits 服务
+// 20230301 版 `/compartments/{tenancy}/capabilities?accountInfoOnly=true`
+// (与控制台同源;oci-go-sdk v65 未收录,自签名裸调)。
+// 该接口对暂停/已终止租户仍可访问,借此把「暂停」从「失联」中区分出来。
+type AccountCapabilities struct {
+	Suspended           bool   `json:"suspended"`
+	AccountStatus       string `json:"accountStatus"`
+	FreeTierEnabled     bool   `json:"freeTierEnabled"`
+	FreeTierOnly        bool   `json:"freeTierOnly"`
+	PromotionStatus     string `json:"promotionStatus"`
+	ProgramType         string `json:"programType"`
+	HasSaasSubscription bool   `json:"hasSaasSubscription"`
+	Deletable           bool   `json:"deletable"`
+}
+
+// capabilitiesEnvelope 对应接口原始响应:顶层与 accountInfo 各有一个
+// suspended,任一为 true 即视为暂停。
+type capabilitiesEnvelope struct {
+	Capabilities struct {
+		Suspended   bool `json:"suspended"`
+		AccountInfo struct {
+			FreeTierEnabled     bool   `json:"freeTierEnabled"`
+			FreeTierOnly        bool   `json:"freeTierOnly"`
+			PromotionStatus     string `json:"promotionStatus"`
+			Suspended           bool   `json:"suspended"`
+			ProgramType         string `json:"programType"`
+			HasSaasSubscription bool   `json:"hasSaasSubscription"`
+			AccountStatus       string `json:"accountStatus"`
+			Deletable           bool   `json:"deletable"`
+		} `json:"accountInfo"`
+	} `json:"capabilities"`
+}
+
+// parseAccountCapabilities 把原始响应体压平为关键字段视图。
+func parseAccountCapabilities(body []byte) (AccountCapabilities, error) {
+	var env capabilitiesEnvelope
+	if err := json.Unmarshal(body, &env); err != nil {
+		return AccountCapabilities{}, fmt.Errorf("decode account capabilities: %w", err)
+	}
+	info := env.Capabilities.AccountInfo
+	return AccountCapabilities{
+		Suspended:           env.Capabilities.Suspended || info.Suspended,
+		AccountStatus:       info.AccountStatus,
+		FreeTierEnabled:     info.FreeTierEnabled,
+		FreeTierOnly:        info.FreeTierOnly,
+		PromotionStatus:     info.PromotionStatus,
+		ProgramType:         info.ProgramType,
+		HasSaasSubscription: info.HasSaasSubscription,
+		Deletable:           info.Deletable,
+	}, nil
+}
+
+// GetAccountCapabilities 实现 Client:查询账户能力(暂停/账户状态/免费层)。
+// region 应传 home region;请求带 API Key 签名,并沿用租户出站代理。
+func (c *RealClient) GetAccountCapabilities(ctx context.Context, cred Credentials, region string) (AccountCapabilities, error) {
+	url := fmt.Sprintf("https://limits.%s.oci.oraclecloud.com/20230301/compartments/%s/capabilities?accountInfoOnly=true",
+		normalizeRegion(region), cred.TenancyOCID)
+	req, err := http.NewRequestWithContext(ctx, http.MethodGet, url, nil)
+	if err != nil {
+		return AccountCapabilities{}, fmt.Errorf("new capabilities request: %w", err)
+	}
+	if err := common.DefaultRequestSigner(provider(cred)).Sign(req); err != nil {
+		return AccountCapabilities{}, fmt.Errorf("sign capabilities request: %w", err)
+	}
+	hc := proxyHTTPClient(cred.Proxy)
+	if hc == nil {
+		hc = &http.Client{Timeout: proxyClientTimeout}
+	}
+	resp, err := hc.Do(req)
+	if err != nil {
+		return AccountCapabilities{}, fmt.Errorf("get account capabilities: %w", err)
+	}
+	defer resp.Body.Close()
+	body, err := io.ReadAll(io.LimitReader(resp.Body, 1<<20))
+	if err != nil {
+		return AccountCapabilities{}, fmt.Errorf("read account capabilities: %w", err)
+	}
+	if resp.StatusCode != http.StatusOK {
+		return AccountCapabilities{}, fmt.Errorf("get account capabilities: status %d", resp.StatusCode)
+	}
+	return parseAccountCapabilities(body)
+}
+
 // ClassifyAccount 按订阅字段判定账户类别：
 //   - paymentModel 非空且不是 FREE_TRIAL：付费；
 //   - paymentModel 为 FREE_TRIAL 时，promotion ACTIVE 或 tier FREE_AND_TRIAL
@@ -36,3 +36,53 @@ func TestClassifyAccount(t *testing.T) {
 		})
 	}
 }
+
+// TestParseAccountCapabilities 用控制台同源接口的真实响应样例驱动。
+func TestParseAccountCapabilities(t *testing.T) {
+	cases := []struct {
+		name       string
+		body       string
+		wantSusp   bool
+		wantStatus string
+		wantErr    bool
+	}{
+		{
+			name: "已终止租户(顶层与 accountInfo 双 suspended)",
+			body: `{"compartmentId":"ocid1.tenancy.oc1..x","capabilities":{"suspended":true,` +
+				`"accountInfo":{"freeTierEnabled":false,"freeTierOnly":false,"promotionStatus":"none",` +
+				`"suspended":true,"programType":"default","limitsProvisioned":true,"intentToPay":false,` +
+				`"hasSaasSubscription":true,"accountStatus":"terminated","accountFlags":0,"deletable":true,` +
+				`"limitIncrease":false,"orgProperties":"0"}}}`,
+			wantSusp:   true,
+			wantStatus: "terminated",
+		},
+		{
+			name:     "正常租户",
+			body:     `{"capabilities":{"suspended":false,"accountInfo":{"accountStatus":"active","freeTierEnabled":true}}}`,
+			wantSusp: false, wantStatus: "active",
+		},
+		{
+			name:     "仅 accountInfo 标记 suspended 也算暂停",
+			body:     `{"capabilities":{"suspended":false,"accountInfo":{"suspended":true}}}`,
+			wantSusp: true,
+		},
+		{name: "非 JSON 报错", body: `<html>`, wantErr: true},
+	}
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			got, err := parseAccountCapabilities([]byte(tc.body))
+			if tc.wantErr {
+				if err == nil {
+					t.Fatal("expected error")
+				}
+				return
+			}
+			if err != nil {
+				t.Fatalf("parse: %v", err)
+			}
+			if got.Suspended != tc.wantSusp || got.AccountStatus != tc.wantStatus {
+				t.Fatalf("got %+v, want suspended=%v status=%q", got, tc.wantSusp, tc.wantStatus)
+			}
+		})
+	}
+}
@@ -93,6 +93,11 @@ func (c *RealClient) ListAuditEvents(ctx context.Context, cred Credentials, regi

 // appendAuditEvents 过滤噪声后追加一页事件;原始事件只对保留条目序列化。
 func appendAuditEvents(result *AuditEventsResult, items []audit.AuditEvent) {
+	result.Items = appendKeptAuditEvents(result.Items, items)
+}
+
+// appendKeptAuditEvents 是过滤追加的通用形态,窗口式与批式查询共用。
+func appendKeptAuditEvents(dst []AuditEvent, items []audit.AuditEvent) []AuditEvent {
 	for _, ev := range items {
 		out := toAuditEvent(ev)
 		if !keepAuditEvent(out) {
@@ -101,8 +106,115 @@ func appendAuditEvents(result *AuditEventsResult, items []audit.AuditEvent) {
 		if raw, mErr := json.Marshal(ev); mErr == nil {
 			out.Raw = raw
 		}
-		result.Items = append(result.Items, out)
+		dst = append(dst, out)
 	}
+	return dst
+}
+
+// ---- 批式懒加载查询:分窗回溯 + 游标续查 ----
+
+// 批式查询参数:单批 OCI 翻页预算沿用 maxAuditPages;首窗 24h,
+// 连续空窗倍增(上限 30 天)加速跨越闲置期;回溯下限为事件保留期 365 天。
+const (
+	auditWindowHours    = 24
+	auditWindowMaxHours = 720
+	auditRetentionDays  = 365
+)
+
+// AuditCursor 是批式查询的续查位置:当前时间窗、窗内 OCI 翻页游标
+// 与当前窗宽(小时,空窗倍增的记忆)。序列化为不透明 cursor 由 service 层负责。
+type AuditCursor struct {
+	Start       time.Time `json:"s"`
+	End         time.Time `json:"e"`
+	Page        string    `json:"p,omitempty"`
+	WindowHours int       `json:"w"`
+}
+
+// NewAuditCursor 构造首查游标:自 now 起回溯第一个 24h 窗。
+func NewAuditCursor(now time.Time) AuditCursor {
+	end := now.UTC().Truncate(time.Minute)
+	return AuditCursor{Start: end.Add(-auditWindowHours * time.Hour), End: end, WindowHours: auditWindowHours}
+}
+
+// advance 推进到紧邻更早的窗;empty 表示刚结束的窗无保留事件,窗宽倍增,
+// 否则重置 24h。done 为 true 表示已越过保留期尽头。
+func (cur AuditCursor) advance(now time.Time, empty bool) (AuditCursor, bool) {
+	w := cur.WindowHours
+	if w <= 0 {
+		w = auditWindowHours
+	}
+	if empty {
+		if w *= 2; w > auditWindowMaxHours {
+			w = auditWindowMaxHours
+		}
+	} else {
+		w = auditWindowHours
+	}
+	end := cur.Start
+	if end.Before(now.UTC().AddDate(0, 0, -auditRetentionDays)) {
+		return cur, true
+	}
+	return AuditCursor{Start: end.Add(-time.Duration(w) * time.Hour), End: end, WindowHours: w}, false
+}
+
+// AuditBatchResult 是一批懒加载结果;Cursor 为 nil 且 Exhausted 为 true
+// 表示已回溯到保留期尽头,无更早数据。
+type AuditBatchResult struct {
+	Items     []AuditEvent
+	Cursor    *AuditCursor
+	Exhausted bool
+}
+
+// ListAuditEventsBatch 实现 Client:从 cur 位置向更早方向收集约 limit 条
+// 保留事件;单批最多消费 maxAuditPages 页 OCI 调用,不足额也按预算返回,
+// 由前端按需续查。窗口不重叠 + 窗内游标续翻保证跨批不重不漏。
+func (c *RealClient) ListAuditEventsBatch(ctx context.Context, cred Credentials, region string, cur AuditCursor, limit int) (AuditBatchResult, error) {
+	ac, err := c.auditClient(cred, region)
+	if err != nil {
+		return AuditBatchResult{}, err
+	}
+	res := AuditBatchResult{Items: []AuditEvent{}}
+	windowHasKept := false
+	for budget := maxAuditPages; budget > 0 && len(res.Items) < limit; budget-- {
+		items, next, err := listAuditPage(ctx, ac, cred.TenancyOCID, cur)
+		if err != nil {
+			return AuditBatchResult{}, err
+		}
+		before := len(res.Items)
+		res.Items = appendKeptAuditEvents(res.Items, items)
+		windowHasKept = windowHasKept || len(res.Items) > before
+		if next != "" {
+			cur.Page = next
+			continue
+		}
+		nextCur, done := cur.advance(time.Now(), !windowHasKept)
+		if done {
+			res.Exhausted = true
+			sortAuditEvents(res.Items)
+			return res, nil
+		}
+		cur, windowHasKept = nextCur, false
+	}
+	sortAuditEvents(res.Items)
+	res.Cursor = &cur
+	return res, nil
+}
+
+// listAuditPage 拉取当前游标位置的一页原始事件。
+func listAuditPage(ctx context.Context, ac audit.AuditClient, tenancyOCID string, cur AuditCursor) ([]audit.AuditEvent, string, error) {
+	req := audit.ListEventsRequest{
+		CompartmentId: &tenancyOCID,
+		StartTime:     &common.SDKTime{Time: cur.Start.UTC().Truncate(time.Minute)},
+		EndTime:       &common.SDKTime{Time: cur.End.UTC().Truncate(time.Minute)},
+	}
+	if cur.Page != "" {
+		req.Page = &cur.Page
+	}
+	resp, err := ac.ListEvents(ctx, req)
+	if err != nil {
+		return nil, "", fmt.Errorf("list audit events: %w", err)
+	}
+	return resp.Items, deref(resp.OpcNextPage), nil
 }

 // auditInternalCIDRs 是 OCI 服务内部互调的发起方网段(RFC1918 + CGNAT)。
@@ -142,3 +142,48 @@ func TestKeepAuditEvent(t *testing.T) {
 		})
 	}
 }
+
+func TestAuditCursorAdvance(t *testing.T) {
+	now := time.Date(2026, 7, 10, 12, 0, 0, 0, time.UTC)
+	base := AuditCursor{
+		Start:       now.Add(-24 * time.Hour),
+		End:         now,
+		WindowHours: 24,
+	}
+	cases := []struct {
+		name      string
+		cur       AuditCursor
+		empty     bool
+		wantHours int
+		wantDone  bool
+	}{
+		{"有事件重置 24h 窗", AuditCursor{Start: base.Start, End: base.End, WindowHours: 96}, false, 24, false},
+		{"空窗倍增", base, true, 48, false},
+		{"倍增封顶 720h", AuditCursor{Start: base.Start, End: base.End, WindowHours: 512}, true, 720, false},
+		{"窗宽缺省按 24h 起算", AuditCursor{Start: base.Start, End: base.End}, true, 48, false},
+		{"越过保留期即尽头", AuditCursor{Start: now.AddDate(0, 0, -366), End: now.AddDate(0, 0, -365), WindowHours: 24}, false, 0, true},
+	}
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			next, done := tc.cur.advance(now, tc.empty)
+			if done != tc.wantDone {
+				t.Fatalf("done = %v, want %v", done, tc.wantDone)
+			}
+			if done {
+				return
+			}
+			if next.WindowHours != tc.wantHours {
+				t.Fatalf("WindowHours = %d, want %d", next.WindowHours, tc.wantHours)
+			}
+			if !next.End.Equal(tc.cur.Start) {
+				t.Fatalf("新窗 End = %v, 应紧邻上窗 Start %v", next.End, tc.cur.Start)
+			}
+			if got := next.End.Sub(next.Start); got != time.Duration(tc.wantHours)*time.Hour {
+				t.Fatalf("窗宽 = %v, want %dh", got, tc.wantHours)
+			}
+			if next.Page != "" {
+				t.Fatalf("新窗应清空窗内游标, got %q", next.Page)
+			}
+		})
+	}
+}
@@ -39,10 +39,18 @@ func ckey(cred Credentials, parts ...string) string {
 // bust 写操作成功后失效该租户全部读缓存。
 func (c *CachedClient) bust(err error, cred Credentials) {
 	if err == nil {
-		c.store.DeletePrefix(cred.TenancyOCID + "|")
+		c.InvalidateTenancy(cred.TenancyOCID)
 	}
 }

+// InvalidateTenancy 主动失效指定 tenancy 的全部 OCI 读缓存。
+func (c *CachedClient) InvalidateTenancy(tenancyOCID string) {
+	if tenancyOCID == "" {
+		return
+	}
+	c.store.DeletePrefix(tenancyOCID + "|")
+}
+
 // cachedList 缓存切片结果;命中返回浅拷贝,防调用方排序共享底层数组。
 func cachedList[T any](c *CachedClient, key string, ttl time.Duration, fn func() ([]T, error)) ([]T, error) {
 	v, err := cache.Do(c.store, key, ttl, fn)
@@ -120,6 +128,13 @@ func (c *CachedClient) ListVolumeAttachments(ctx context.Context, cred Credentia
 	})
 }

+// ListIdentityDomains 域列表准静态且是所有按域操作的 URL 解析源,长 TTL 缓存。
+func (c *CachedClient) ListIdentityDomains(ctx context.Context, cred Credentials, region string) ([]IdentityDomain, error) {
+	return cachedList(c, ckey(cred, "iddomains", region), cacheTTLStatic, func() ([]IdentityDomain, error) {
+		return c.Client.ListIdentityDomains(ctx, cred, region)
+	})
+}
+
 // ---- 写:直通,成功后按租户失效 ----

 func (c *CachedClient) LaunchInstance(ctx context.Context, cred Credentials, in CreateInstanceInput) (Instance, error) {
@@ -86,3 +86,18 @@ func TestCachedClientReturnsClone(t *testing.T) {
 		t.Errorf("缓存底层数组被调用方污染: second[0]=%s, want i-1", second[0].ID)
 	}
 }
+
+func TestCachedClientInvalidateTenancy(t *testing.T) {
+	inner := &countingClient{}
+	c := NewCachedClient(inner)
+	ctx := context.Background()
+
+	_, _ = c.ListInstances(ctx, testCred("t1"), "r1")
+	_, _ = c.ListInstances(ctx, testCred("t2"), "r1")
+	c.InvalidateTenancy("t1")
+	_, _ = c.ListInstances(ctx, testCred("t1"), "r1")
+	_, _ = c.ListInstances(ctx, testCred("t2"), "r1")
+	if inner.instCalls != 3 {
+		t.Fatalf("invalidate tenancy calls = %d, want 3", inner.instCalls)
+	}
+}
@@ -6,6 +6,7 @@ import (
 	"context"
 	"errors"
 	"fmt"
+	"io"
 	"strings"
 	"sync"
 	"time"
@@ -99,10 +100,18 @@ type Client interface {
 	AddVnicIpv6(ctx context.Context, cred Credentials, region, vnicID, address string) (string, error)
 	// GenAI 网关:区域模型列表、聊天(非流式/流式)、配额探测、文本向量化。
 	ListGenAiModels(ctx context.Context, cred Credentials, region string) ([]GenAiModel, error)
-	GenAiChat(ctx context.Context, cred Credentials, region, modelOcid string, ir aiwire.ChatRequest) (*aiwire.ChatResponse, error)
-	GenAiChatStream(ctx context.Context, cred Credentials, region, modelOcid string, ir aiwire.ChatRequest) (GenAiStream, error)
 	GenAiProbeChat(ctx context.Context, cred Credentials, region, modelOcid, modelName string) (int, error)
 	GenAiEmbed(ctx context.Context, cred Credentials, region, modelOcid string, inputs []string, dimensions *int) ([][]float32, *aiwire.Usage, error)
+	// GenAiCompatResponses 直通 OpenAI Responses 请求体到 /actions/v1/responses(xAI 服务端工具通路)。
+	GenAiCompatResponses(ctx context.Context, cred Credentials, region string, body []byte) ([]byte, error)
+	// GenAiCompatResponsesStream 流式直通 /actions/v1/responses,建立成功返回 SSE body。
+	GenAiCompatResponsesStream(ctx context.Context, cred Credentials, region string, body []byte) (io.ReadCloser, error)
+	// GenAiCompatSpeech 直通 OpenAI Audio Speech 请求体到 /openai/v1/audio/speech,返回音频与 Content-Type。
+	GenAiCompatSpeech(ctx context.Context, cred Credentials, region string, body []byte) ([]byte, string, error)
+	// GenAiRerank 文档重排,返回按相关度排序的下标与得分。
+	GenAiRerank(ctx context.Context, cred Credentials, region, modelOcid, query string, documents []string, topN *int) ([]RerankRank, error)
+	// GenAiApplyGuardrails 对单条文本执行内容审核 / PII / 提示注入检测。
+	GenAiApplyGuardrails(ctx context.Context, cred Credentials, region, text string) (*GuardrailsOutcome, error)
 	// 控制台连接：创建（VNC/串口连接串）、列出、删除。
 	CreateConsoleConnection(ctx context.Context, cred Credentials, region, instanceID, sshPublicKey string) (ConsoleConnection, error)
 	ListConsoleConnections(ctx context.Context, cred Credentials, region, instanceID string) ([]ConsoleConnection, error)
@@ -122,32 +131,41 @@ type Client interface {
 	SummarizeCosts(ctx context.Context, cred Credentials, q CostQuery) ([]CostItem, error)
 	// ListAuditEvents 实时查询租户根 compartment 的审计事件(最多翻 maxAuditPages 页,
 	// 到限截断并回传续查游标);page 传上次结果的 NextPage 可从断点续翻。
+	// 批式懒加载走 ListAuditEventsBatch,本方法保留给详情小窗反查。
 	ListAuditEvents(ctx context.Context, cred Credentials, region string, start, end time.Time, page string) (AuditEventsResult, error)
-	// 租户用户管理：经典 IAM 为主，新式租户自动回退 Identity Domains。
-	ListTenantUsers(ctx context.Context, cred Credentials) ([]TenantUser, error)
-	GetTenantUserDetail(ctx context.Context, cred Credentials, homeRegion, userID string) (TenantUserDetail, error)
-	CreateTenantUser(ctx context.Context, cred Credentials, homeRegion string, in CreateTenantUserInput) (TenantUser, error)
-	UpdateTenantUser(ctx context.Context, cred Credentials, homeRegion, userID string, in UpdateTenantUserInput) (TenantUser, error)
-	DeleteTenantUser(ctx context.Context, cred Credentials, homeRegion, userID string) error
-	ResetTenantUserPassword(ctx context.Context, cred Credentials, homeRegion, userID string) (string, error)
-	DeleteTenantUserMfaDevices(ctx context.Context, cred Credentials, homeRegion, userID string) (int, error)
+	// ListAuditEventsBatch 从游标位置向更早方向收集约 limit 条保留事件,
+	// 分窗回溯(空窗倍增),跨批不重不漏;到 365 天保留期尽头置 Exhausted。
+	ListAuditEventsBatch(ctx context.Context, cred Credentials, region string, cur AuditCursor, limit int) (AuditBatchResult, error)
+	// 身份域:租户 ACTIVE 域枚举(多域租户选择器数据源)。
+	ListIdentityDomains(ctx context.Context, cred Credentials, region string) ([]IdentityDomain, error)
+	// 账户能力:暂停标记/账户状态/免费层信息(对暂停租户仍可访问,测活据此判「暂停」)。
+	GetAccountCapabilities(ctx context.Context, cred Credentials, region string) (AccountCapabilities, error)
+	// 租户用户管理:domainID 非空按域走 Identity Domains SCIM,
+	// 为空保持经典 IAM 为主、新式租户自动回退的旧行为。
+	ListTenantUsers(ctx context.Context, cred Credentials, homeRegion, domainID string) ([]TenantUser, error)
+	GetTenantUserDetail(ctx context.Context, cred Credentials, homeRegion, domainID, userID string) (TenantUserDetail, error)
+	CreateTenantUser(ctx context.Context, cred Credentials, homeRegion, domainID string, in CreateTenantUserInput) (TenantUser, error)
+	UpdateTenantUser(ctx context.Context, cred Credentials, homeRegion, domainID, userID string, in UpdateTenantUserInput) (TenantUser, error)
+	DeleteTenantUser(ctx context.Context, cred Credentials, homeRegion, domainID, userID string) error
+	ResetTenantUserPassword(ctx context.Context, cred Credentials, homeRegion, domainID, userID string) (string, error)
+	DeleteTenantUserMfaDevices(ctx context.Context, cred Credentials, homeRegion, domainID, userID string) (int, error)
 	DeleteTenantUserApiKeys(ctx context.Context, cred Credentials, homeRegion, userID string, includeCurrent bool) (int, error)
 	// 域设置：通知收件人、密码策略、身份设置。
-	GetNotificationRecipients(ctx context.Context, cred Credentials, region string) (NotificationRecipients, error)
-	UpdateNotificationRecipients(ctx context.Context, cred Credentials, region string, emails []string) (NotificationRecipients, error)
-	ListPasswordPolicies(ctx context.Context, cred Credentials, region string) ([]PasswordPolicyInfo, error)
-	UpdatePasswordPolicy(ctx context.Context, cred Credentials, region, policyID string, in UpdatePasswordPolicyInput) (PasswordPolicyInfo, error)
-	GetIdentitySetting(ctx context.Context, cred Credentials, region string) (IdentitySettingInfo, error)
-	UpdateIdentitySetting(ctx context.Context, cred Credentials, region string, primaryEmailRequired bool) (IdentitySettingInfo, error)
+	GetNotificationRecipients(ctx context.Context, cred Credentials, region, domainID string) (NotificationRecipients, error)
+	UpdateNotificationRecipients(ctx context.Context, cred Credentials, region, domainID string, emails []string) (NotificationRecipients, error)
+	ListPasswordPolicies(ctx context.Context, cred Credentials, region, domainID string) ([]PasswordPolicyInfo, error)
+	UpdatePasswordPolicy(ctx context.Context, cred Credentials, region, domainID, policyID string, in UpdatePasswordPolicyInput) (PasswordPolicyInfo, error)
+	GetIdentitySetting(ctx context.Context, cred Credentials, region, domainID string) (IdentitySettingInfo, error)
+	UpdateIdentitySetting(ctx context.Context, cred Credentials, region, domainID string, primaryEmailRequired bool) (IdentitySettingInfo, error)
 	// Federation：SAML IdP 管理、域 SP 元数据、sign-on 免 MFA 规则。
-	ListIdentityProviders(ctx context.Context, cred Credentials, region string) ([]IdentityProviderInfo, error)
-	CreateSamlIdentityProvider(ctx context.Context, cred Credentials, region string, in CreateIdpInput) (IdentityProviderInfo, error)
-	SetIdentityProviderEnabled(ctx context.Context, cred Credentials, region, idpID string, enabled bool) (IdentityProviderInfo, error)
-	DeleteIdentityProvider(ctx context.Context, cred Credentials, region, idpID string) error
-	DownloadDomainSamlMetadata(ctx context.Context, cred Credentials, region string) ([]byte, error)
-	ListConsoleSignOnRules(ctx context.Context, cred Credentials, region string) ([]SignOnRuleInfo, error)
-	CreateMfaExemptionRule(ctx context.Context, cred Credentials, region, idpID, ruleName string) (SignOnRuleInfo, error)
-	DeleteMfaExemptionRule(ctx context.Context, cred Credentials, region, ruleID string) error
+	ListIdentityProviders(ctx context.Context, cred Credentials, region, domainID string) ([]IdentityProviderInfo, error)
+	CreateSamlIdentityProvider(ctx context.Context, cred Credentials, region, domainID string, in CreateIdpInput) (IdentityProviderInfo, error)
+	SetIdentityProviderEnabled(ctx context.Context, cred Credentials, region, domainID, idpID string, enabled bool) (IdentityProviderInfo, error)
+	DeleteIdentityProvider(ctx context.Context, cred Credentials, region, domainID, idpID string) error
+	DownloadDomainSamlMetadata(ctx context.Context, cred Credentials, region, domainID string) ([]byte, error)
+	ListConsoleSignOnRules(ctx context.Context, cred Credentials, region, domainID string) ([]SignOnRuleInfo, error)
+	CreateMfaExemptionRule(ctx context.Context, cred Credentials, region, domainID, idpID, ruleName string) (SignOnRuleInfo, error)
+	DeleteMfaExemptionRule(ctx context.Context, cred Credentials, region, domainID, ruleID string) error
 	// 日志回传链路(方案A):Topic/CUSTOM_HTTPS 订阅/IAM Policy/Service Connector
 	// 的幂等创建、状态聚合与销毁;IAM 写操作发往 homeRegion。
 	EnsureRelayTopic(ctx context.Context, cred Credentials) (RelayResource, error)
@@ -47,26 +47,82 @@ type IdentitySettingInfo struct {
 	PrimaryEmailRequired bool `json:"primaryEmailRequired"`
 }

-// domainURL 解析租户 Default Identity Domain 的 endpoint URL。
-func (c *RealClient) domainURL(ctx context.Context, cred Credentials, region string) (string, error) {
+// IdentityDomain 是租户下一个 ACTIVE 身份域的列表视图;
+// URL 仅供后端解析 SCIM 端点,不下发给前端(防面板被引导签名任意地址)。
+type IdentityDomain struct {
+	ID          string `json:"id"`
+	DisplayName string `json:"displayName"`
+	URL         string `json:"-"`
+	HomeRegion  string `json:"homeRegion"`
+	Type        string `json:"type"`
+	LicenseType string `json:"licenseType"`
+}
+
+// ListIdentityDomains 实现 Client:列出租户全部 ACTIVE 身份域。
+// 无域老租户返回空列表(非错误),调用方据此回退经典 IAM 路径。
+func (c *RealClient) ListIdentityDomains(ctx context.Context, cred Credentials, region string) ([]IdentityDomain, error) {
+	items, err := c.listDomainSummaries(ctx, cred, region)
+	if err != nil {
+		return nil, err
+	}
+	out := make([]IdentityDomain, 0, len(items))
+	for _, d := range items {
+		if d.LifecycleState != identity.DomainLifecycleStateActive {
+			continue
+		}
+		out = append(out, IdentityDomain{
+			ID:          deref(d.Id),
+			DisplayName: deref(d.DisplayName),
+			URL:         deref(d.Url),
+			HomeRegion:  deref(d.HomeRegion),
+			Type:        string(d.Type),
+			LicenseType: deref(d.LicenseType),
+		})
+	}
+	return out, nil
+}
+
+// listDomainSummaries 拉取租户全部域(数量极少,仍按游标翻全以防万一)。
+func (c *RealClient) listDomainSummaries(ctx context.Context, cred Credentials, region string) ([]identity.DomainSummary, error) {
 	ic, err := c.identityClientAt(cred, region)
+	if err != nil {
+		return nil, err
+	}
+	var items []identity.DomainSummary
+	var page *string
+	for {
+		resp, err := ic.ListDomains(ctx, identity.ListDomainsRequest{CompartmentId: &cred.TenancyOCID, Page: page})
+		if err != nil {
+			return nil, fmt.Errorf("list identity domains: %w", err)
+		}
+		items = append(items, resp.Items...)
+		if resp.OpcNextPage == nil {
+			return items, nil
+		}
+		page = resp.OpcNextPage
+	}
+}
+
+// domainURL 解析身份域的 endpoint URL:domainID 非空按 OCID 精确匹配
+// (URL 只能来自租户自己的域列表,不信任外部输入),为空回退 Default 优先。
+func (c *RealClient) domainURL(ctx context.Context, cred Credentials, region, domainID string) (string, error) {
+	items, err := c.listDomainSummaries(ctx, cred, region)
 	if err != nil {
 		return "", err
 	}
-	resp, err := ic.ListDomains(ctx, identity.ListDomainsRequest{CompartmentId: &cred.TenancyOCID})
-	if err != nil {
-		return "", fmt.Errorf("list identity domains: %w", err)
-	}
-	url := pickDomainURL(resp.Items)
+	url := pickDomainURL(items, domainID)
 	if url == "" {
+		if domainID != "" {
+			return "", fmt.Errorf("identity domain %s not found or inactive", domainID)
+		}
 		return "", fmt.Errorf("no active identity domain found")
 	}
 	return url, nil
 }

-// domainsClient 解析租户的 Default Identity Domain 并构造 SCIM 客户端。
-func (c *RealClient) domainsClient(ctx context.Context, cred Credentials, region string) (identitydomains.IdentityDomainsClient, error) {
-	url, err := c.domainURL(ctx, cred, region)
+// domainsClient 解析身份域(domainID 为空取 Default)并构造 SCIM 客户端。
+func (c *RealClient) domainsClient(ctx context.Context, cred Credentials, region, domainID string) (identitydomains.IdentityDomainsClient, error) {
+	url, err := c.domainURL(ctx, cred, region, domainID)
 	if err != nil {
 		return identitydomains.IdentityDomainsClient{}, err
 	}
@@ -80,13 +136,20 @@ func (c *RealClient) domainsClient(ctx context.Context, cred Credentials, region
 	return dc, nil
 }

-// pickDomainURL 优先返回名为 Default 的域，否则取第一个 ACTIVE 的域。
-func pickDomainURL(domains []identity.DomainSummary) string {
+// pickDomainURL 在 ACTIVE 域中选取:domainID 非空按 OCID 匹配;
+// 为空优先名为 Default 的域,否则取第一个 ACTIVE 的域。
+func pickDomainURL(domains []identity.DomainSummary, domainID string) string {
 	fallback := ""
 	for _, d := range domains {
 		if d.LifecycleState != identity.DomainLifecycleStateActive {
 			continue
 		}
+		if domainID != "" {
+			if deref(d.Id) == domainID {
+				return deref(d.Url)
+			}
+			continue
+		}
 		if deref(d.DisplayName) == "Default" {
 			return deref(d.Url)
 		}
@@ -98,8 +161,8 @@ func pickDomainURL(domains []identity.DomainSummary) string {
 }

 // GetNotificationRecipients 实现 Client：查询域通知收件人设置。
-func (c *RealClient) GetNotificationRecipients(ctx context.Context, cred Credentials, region string) (NotificationRecipients, error) {
-	dc, err := c.domainsClient(ctx, cred, region)
+func (c *RealClient) GetNotificationRecipients(ctx context.Context, cred Credentials, region, domainID string) (NotificationRecipients, error) {
+	dc, err := c.domainsClient(ctx, cred, region, domainID)
 	if err != nil {
 		return NotificationRecipients{}, err
 	}
@@ -115,8 +178,8 @@ func (c *RealClient) GetNotificationRecipients(ctx context.Context, cred Credent

 // UpdateNotificationRecipients 实现 Client：把域通知改为只发给指定收件人；
 // emails 为空时关闭 test mode 恢复默认发送。其余字段原样保留。
-func (c *RealClient) UpdateNotificationRecipients(ctx context.Context, cred Credentials, region string, emails []string) (NotificationRecipients, error) {
-	dc, err := c.domainsClient(ctx, cred, region)
+func (c *RealClient) UpdateNotificationRecipients(ctx context.Context, cred Credentials, region, domainID string, emails []string) (NotificationRecipients, error) {
+	dc, err := c.domainsClient(ctx, cred, region, domainID)
 	if err != nil {
 		return NotificationRecipients{}, err
 	}
@@ -153,8 +216,8 @@ func getNotificationSetting(ctx context.Context, dc identitydomains.IdentityDoma
 }

 // ListPasswordPolicies 实现 Client：列出域内全部密码策略。
-func (c *RealClient) ListPasswordPolicies(ctx context.Context, cred Credentials, region string) ([]PasswordPolicyInfo, error) {
-	dc, err := c.domainsClient(ctx, cred, region)
+func (c *RealClient) ListPasswordPolicies(ctx context.Context, cred Credentials, region, domainID string) ([]PasswordPolicyInfo, error) {
+	dc, err := c.domainsClient(ctx, cred, region, domainID)
 	if err != nil {
 		return nil, err
 	}
@@ -171,12 +234,12 @@ func (c *RealClient) ListPasswordPolicies(ctx context.Context, cred Credentials,

 // UpdatePasswordPolicy 实现 Client：以 SCIM Patch 修改密码策略的指定字段。
 // 仅 Custom 强度的策略可修改，Simple/Standard 内置策略 OCI 返回 403。
-func (c *RealClient) UpdatePasswordPolicy(ctx context.Context, cred Credentials, region, policyID string, in UpdatePasswordPolicyInput) (PasswordPolicyInfo, error) {
+func (c *RealClient) UpdatePasswordPolicy(ctx context.Context, cred Credentials, region, domainID, policyID string, in UpdatePasswordPolicyInput) (PasswordPolicyInfo, error) {
 	ops := buildPolicyPatchOps(in)
 	if len(ops) == 0 {
 		return PasswordPolicyInfo{}, fmt.Errorf("update password policy: nothing to update")
 	}
-	dc, err := c.domainsClient(ctx, cred, region)
+	dc, err := c.domainsClient(ctx, cred, region, domainID)
 	if err != nil {
 		return PasswordPolicyInfo{}, err
 	}
@@ -233,8 +296,8 @@ func removeOp(path string) identitydomains.Operations {
 }

 // GetIdentitySetting 实现 Client：读取域身份设置（IdentitySettings 单例资源）。
-func (c *RealClient) GetIdentitySetting(ctx context.Context, cred Credentials, region string) (IdentitySettingInfo, error) {
-	dc, err := c.domainsClient(ctx, cred, region)
+func (c *RealClient) GetIdentitySetting(ctx context.Context, cred Credentials, region, domainID string) (IdentitySettingInfo, error) {
+	dc, err := c.domainsClient(ctx, cred, region, domainID)
 	if err != nil {
 		return IdentitySettingInfo{}, err
 	}
@@ -254,12 +317,12 @@ func (c *RealClient) GetIdentitySetting(ctx context.Context, cred Credentials, r
 }

 // UpdateIdentitySetting 实现 Client：修改「用户需要提供主电子邮件地址」开关。
-func (c *RealClient) UpdateIdentitySetting(ctx context.Context, cred Credentials, region string, primaryEmailRequired bool) (IdentitySettingInfo, error) {
-	current, err := c.GetIdentitySetting(ctx, cred, region)
+func (c *RealClient) UpdateIdentitySetting(ctx context.Context, cred Credentials, region, domainID string, primaryEmailRequired bool) (IdentitySettingInfo, error) {
+	current, err := c.GetIdentitySetting(ctx, cred, region, domainID)
 	if err != nil {
 		return IdentitySettingInfo{}, err
 	}
-	dc, err := c.domainsClient(ctx, cred, region)
+	dc, err := c.domainsClient(ctx, cred, region, domainID)
 	if err != nil {
 		return IdentitySettingInfo{}, err
 	}
@@ -76,3 +76,34 @@ func ServiceStatus(err error) (int, bool) {
 	}
 	return 0, false
 }
+
+// IsOnDemandUnsupported 识别「模型在该区域仅为微调基座、不支持按需调用」的 400:
+// OCI 消息形如 "Not allowed to call finetune base model …, use Endpoint: false"。
+// ListModels 无字段可事先区分,只能在调用报错时识别并换渠道。
+func IsOnDemandUnsupported(err error) bool {
+	var svcErr common.ServiceError
+	if !errors.As(err, &svcErr) {
+		return false
+	}
+	return svcErr.GetHTTPStatusCode() == 400 &&
+		strings.Contains(strings.ToLower(svcErr.GetMessage()), "finetune base model")
+}
+
+// IsEntityNotFound 识别「实体不存在」404(消息 "Entity with key … not found"):
+// GenAI 对区域内无按需供给的模型 OCID 返回此类 404,属模型级错误;
+// 鉴权失败的 404 是 NotAuthorizedOrNotFound(消息为 Authorization failed…),不在此列。
+func IsEntityNotFound(err error) bool {
+	var svcErr common.ServiceError
+	if !errors.As(err, &svcErr) {
+		return false
+	}
+	msg := strings.ToLower(svcErr.GetMessage())
+	return svcErr.GetHTTPStatusCode() == 404 &&
+		strings.Contains(msg, "entity with key") && strings.Contains(msg, "not found")
+}
+
+// IsModelUnavailable 归并「模型×区域不可按需调用」两类报错:微调基座 400 与实体不存在 404;
+// 命中即应把该 (渠道, 模型) 从池中剔除并换候选/换渠道,而非定论租户配额问题。
+func IsModelUnavailable(err error) bool {
+	return IsOnDemandUnsupported(err) || IsEntityNotFound(err)
+}
@@ -109,3 +109,52 @@ func TestErrorHint(t *testing.T) {
 		})
 	}
 }
+
+func TestIsOnDemandUnsupported(t *testing.T) {
+	ft := fakeServiceError{status: 400, code: "InvalidParameter",
+		message: "Not allowed to call finetune base model ocid1.generativeaimodel.oc1.eu-frankfurt-1.tpel5q, use Endpoint: false"}
+	tests := []struct {
+		name string
+		err  error
+		want bool
+	}{
+		{"微调基座 400 命中(含包装链)", fmt.Errorf("genai chat: %w", ft), true},
+		{"同消息但非 400 不命中", fakeServiceError{status: 500, code: "InternalError", message: ft.message}, false},
+		{"普通 400 不命中", fakeServiceError{status: 400, code: "InvalidParameter", message: "bad request"}, false},
+		{"非服务端错误不命中", errors.New("finetune base model"), false},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			if got := IsOnDemandUnsupported(tt.err); got != tt.want {
+				t.Errorf("IsOnDemandUnsupported() = %v, want %v", got, tt.want)
+			}
+		})
+	}
+}
+
+func TestIsModelUnavailable(t *testing.T) {
+	entity404 := fakeServiceError{status: 404, code: "NotFound",
+		message: "Entity with key ocid1.generativeaimodel.oc1.eu-frankfurt-1.2flsfq not found"}
+	auth404 := fakeServiceError{status: 404, code: "NotAuthorizedOrNotFound",
+		message: "Authorization failed or requested resource not found."}
+	ft400 := fakeServiceError{status: 400, code: "InvalidParameter",
+		message: "Not allowed to call finetune base model ocid1.generativeaimodel.oc1..x, use Endpoint: false"}
+	tests := []struct {
+		name string
+		err  error
+		want bool
+	}{
+		{"实体不存在 404 命中(含包装链)", fmt.Errorf("genai chat: %w", entity404), true},
+		{"鉴权类 404 不命中(仍属租户级)", auth404, false},
+		{"微调基座 400 命中", ft400, true},
+		{"其他 404 消息不命中", fakeServiceError{status: 404, code: "NotFound", message: "route not found"}, false},
+		{"非服务端错误不命中", errors.New("entity with key x not found"), false},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			if got := IsModelUnavailable(tt.err); got != tt.want {
+				t.Errorf("IsModelUnavailable() = %v, want %v", got, tt.want)
+			}
+		})
+	}
+}
@@ -56,8 +56,8 @@ type CreateIdpInput struct {
 }

 // ListIdentityProviders 实现 Client：列出域内 SAML 身份提供者。
-func (c *RealClient) ListIdentityProviders(ctx context.Context, cred Credentials, region string) ([]IdentityProviderInfo, error) {
-	dc, err := c.domainsClient(ctx, cred, region)
+func (c *RealClient) ListIdentityProviders(ctx context.Context, cred Credentials, region, domainID string) ([]IdentityProviderInfo, error) {
+	dc, err := c.domainsClient(ctx, cred, region, domainID)
 	if err != nil {
 		return nil, err
 	}
@@ -76,8 +76,8 @@ func (c *RealClient) ListIdentityProviders(ctx context.Context, cred Credentials
 }

 // CreateSamlIdentityProvider 实现 Client：按输入创建禁用态 SAML IdP 并配置 JIT。
-func (c *RealClient) CreateSamlIdentityProvider(ctx context.Context, cred Credentials, region string, in CreateIdpInput) (IdentityProviderInfo, error) {
-	dc, err := c.domainsClient(ctx, cred, region)
+func (c *RealClient) CreateSamlIdentityProvider(ctx context.Context, cred Credentials, region, domainID string, in CreateIdpInput) (IdentityProviderInfo, error) {
+	dc, err := c.domainsClient(ctx, cred, region, domainID)
 	if err != nil {
 		return IdentityProviderInfo{}, err
 	}
@@ -139,23 +139,28 @@ func buildSamlIdp(in CreateIdpInput, adminGroup *identitydomains.IdentityProvide
 	return idp
 }

-// adminGroupRef 查询域内 Administrators 组作为 JIT 静态分配目标；不存在时返回 nil。
+// adminGroupRef 查询域内管理员组作为 JIT 静态分配与用户授权目标;
+// 按 adminGroupNames 优先序取第一个存在的组,都不存在返回 nil。
 func adminGroupRef(ctx context.Context, dc identitydomains.IdentityDomainsClient) (*identitydomains.IdentityProviderJitUserProvAssignedGroups, error) {
-	filter := fmt.Sprintf("displayName eq %q", administratorsGroupName)
-	count := 1
+	filter := fmt.Sprintf("displayName eq %q or displayName eq %q", adminGroupNames[0], adminGroupNames[1])
+	count := len(adminGroupNames)
 	resp, err := dc.ListGroups(ctx, identitydomains.ListGroupsRequest{
 		Filter:     &filter,
 		Attributes: common.String("id,displayName"),
 		Count:      &count,
 	})
 	if err != nil {
-		return nil, fmt.Errorf("find Administrators group: %w", err)
+		return nil, fmt.Errorf("find administrators group: %w", err)
+	}
+	for _, name := range adminGroupNames {
+		for _, g := range resp.Resources {
+			if deref(g.DisplayName) == name && g.Id != nil {
+				return &identitydomains.IdentityProviderJitUserProvAssignedGroups{Value: g.Id}, nil
+			}
+		}
 	}
-	if len(resp.Resources) == 0 || resp.Resources[0].Id == nil {
 	return nil, nil
 }
-	return &identitydomains.IdentityProviderJitUserProvAssignedGroups{Value: resp.Resources[0].Id}, nil
-}

 // jitMappings 生成 JIT 属性映射：来源为 NameID（或指定断言属性），
 // 固定映射 用户名 + 姓，按需追加主邮箱（emails[work].value）。
@@ -207,8 +212,8 @@ func ensureJitAttributeMappings(ctx context.Context, dc identitydomains.Identity

 // SetIdentityProviderEnabled 实现 Client：启用/停用 IdP 并同步登录页显示。
 // 顺序有讲究：启用后才能上登录页，停用前必须先从登录页移除。
-func (c *RealClient) SetIdentityProviderEnabled(ctx context.Context, cred Credentials, region, idpID string, enabled bool) (IdentityProviderInfo, error) {
-	dc, err := c.domainsClient(ctx, cred, region)
+func (c *RealClient) SetIdentityProviderEnabled(ctx context.Context, cred Credentials, region, domainID, idpID string, enabled bool) (IdentityProviderInfo, error) {
+	dc, err := c.domainsClient(ctx, cred, region, domainID)
 	if err != nil {
 		return IdentityProviderInfo{}, err
 	}
@@ -249,8 +254,8 @@ func patchIdpEnabled(ctx context.Context, dc identitydomains.IdentityDomainsClie
 }

 // DeleteIdentityProvider 实现 Client：删除 IdP；先移出登录页并停用。
-func (c *RealClient) DeleteIdentityProvider(ctx context.Context, cred Credentials, region, idpID string) error {
-	dc, err := c.domainsClient(ctx, cred, region)
+func (c *RealClient) DeleteIdentityProvider(ctx context.Context, cred Credentials, region, domainID, idpID string) error {
+	dc, err := c.domainsClient(ctx, cred, region, domainID)
 	if err != nil {
 		return err
 	}
@@ -346,8 +351,8 @@ func toggleJSONList(list, item string, add bool) (string, bool, error) {
 // DownloadDomainSamlMetadata 实现 Client：下载域的 SP SAML 元数据 XML。
 // 域设置「访问签名证书」未开启时端点不可用，自动开启后重试
 // （联邦配置要求 IdP 侧能读取该元数据，开启公开访问是官方标准做法）。
-func (c *RealClient) DownloadDomainSamlMetadata(ctx context.Context, cred Credentials, region string) ([]byte, error) {
-	url, err := c.domainURL(ctx, cred, region)
+func (c *RealClient) DownloadDomainSamlMetadata(ctx context.Context, cred Credentials, region, domainID string) ([]byte, error) {
+	url, err := c.domainURL(ctx, cred, region, domainID)
 	if err != nil {
 		return nil, err
 	}
@@ -358,7 +363,7 @@ func (c *RealClient) DownloadDomainSamlMetadata(ctx context.Context, cred Creden
 	if status == http.StatusOK {
 		return body, nil
 	}
-	if err := c.enableSigningCertPublicAccess(ctx, cred, region); err != nil {
+	if err := c.enableSigningCertPublicAccess(ctx, cred, region, domainID); err != nil {
 		return nil, err
 	}
 	body, status, err = fetchSamlMetadata(ctx, url)
@@ -390,8 +395,8 @@ func fetchSamlMetadata(ctx context.Context, url string) ([]byte, int, error) {
 }

 // enableSigningCertPublicAccess 开启域设置「访问签名证书」公开访问。
-func (c *RealClient) enableSigningCertPublicAccess(ctx context.Context, cred Credentials, region string) error {
-	dc, err := c.domainsClient(ctx, cred, region)
+func (c *RealClient) enableSigningCertPublicAccess(ctx context.Context, cred Credentials, region, domainID string) error {
+	dc, err := c.domainsClient(ctx, cred, region, domainID)
 	if err != nil {
 		return err
 	}
@@ -2,13 +2,11 @@ package oci

 import (
 	"context"
+	"encoding/json"
 	"fmt"
-	"io"
 	"net/http"
-	"strings"
 	"time"

-	"github.com/oracle/oci-go-sdk/v65/common"
 	"github.com/oracle/oci-go-sdk/v65/generativeai"
 	"github.com/oracle/oci-go-sdk/v65/generativeaiinference"

@@ -20,7 +18,6 @@ type GenAiModel struct {
 	Ocid   string   `json:"ocid"`
 	Name   string   `json:"name"`
 	Vendor string   `json:"vendor"`
-	ChatOnly bool     `json:"-"`
 	Caps   []string `json:"capabilities"`
 	// Capability 是网关侧归一能力:CHAT / EMBEDDING(兼具时算 CHAT)
 	Capability string `json:"capability"`
@@ -31,12 +28,6 @@ type GenAiModel struct {
 	Retired *time.Time `json:"retired"`
 }

-// GenAiStream 逐事件读取流式聊天;Next 在流结束时返回 io.EOF。
-type GenAiStream interface {
-	Next() (aiwire.ChatChunk, error)
-	Close() error
-}
-
 func (c *RealClient) genAiClient(cred Credentials, region string) (generativeai.GenerativeAiClient, error) {
 	gc, err := generativeai.NewGenerativeAiClientWithConfigurationProvider(provider(cred))
 	if err != nil {
@@ -71,18 +62,39 @@ func (c *RealClient) ListGenAiModels(ctx context.Context, cred Credentials, regi
 	if err != nil {
 		return nil, fmt.Errorf("list genai models: %w", err)
 	}
-	seen := map[string]bool{}
-	now := time.Now()
+	return dedupGenAiModels(resp.Items, time.Now()), nil
+}
+
+// dedupGenAiModels 压平并按名称去重;同名多条目时优先保留不含 FINE_TUNE 能力的条目
+// (微调基座条目在部分区域不支持按需调用,缓存其 OCID 会导致调用 400)。
+func dedupGenAiModels(items []generativeai.ModelSummary, now time.Time) []GenAiModel {
+	seen := map[string]int{}
 	var out []GenAiModel
-	for _, m := range resp.Items {
+	for _, m := range items {
 		gm, ok := toGenAiModel(m, now)
-		if !ok || seen[gm.Name] {
+		if !ok {
 			continue
 		}
-		seen[gm.Name] = true
+		if i, dup := seen[gm.Name]; dup {
+			if hasFineTune(out[i].Caps) && !hasFineTune(gm.Caps) {
+				out[i] = gm
+			}
+			continue
+		}
+		seen[gm.Name] = len(out)
 		out = append(out, gm)
 	}
-	return out, nil
+	return out
+}
+
+// hasFineTune 判断能力列表是否含 FINE_TUNE(微调基座条目)。
+func hasFineTune(caps []string) bool {
+	for _, c := range caps {
+		if c == string(generativeai.ModelCapabilityFineTune) {
+			return true
+		}
+	}
+	return false
 }

 // toGenAiModel 压平模型摘要;无归一能力、无名称或按需推理已退役
@@ -119,6 +131,11 @@ func modelCapability(m generativeai.ModelSummary) string {
 			return "CHAT"
 		case generativeai.ModelCapabilityTextEmbeddings:
 			capability = "EMBEDDING"
+		case generativeai.ModelCapabilityTextRerank:
+			capability = "RERANK"
+		case generativeai.ModelCapabilityEnum("TEXT_TO_AUDIO"):
+			// SDK v65.120 尚无该枚举常量,按原始字符串匹配(xai.grok-tts)
+			capability = "TTS"
 		}
 	}
 	return capability
@@ -132,121 +149,15 @@ func capStrings(caps []generativeai.ModelCapabilityEnum) []string {
 	return out
 }

-// GenAiChat 实现 Client:非流式聊天;modelOcid 走 on-demand serving,
-// cohere.* 模型走 COHERE 请求格式,其余走 GENERIC。
-func (c *RealClient) GenAiChat(ctx context.Context, cred Credentials, region, modelOcid string, ir aiwire.ChatRequest) (*aiwire.ChatResponse, error) {
-	ic, err := c.genAiInferenceClient(cred, region)
-	if err != nil {
-		return nil, err
-	}
-	req, err := buildChatRequest(ir, false)
-	if err != nil {
-		return nil, err
-	}
-	resp, err := ic.Chat(ctx, chatRequest(cred, modelOcid, req))
-	if err != nil {
-		return nil, fmt.Errorf("genai chat: %w", err)
-	}
-	return chatResponseToIR(resp.ChatResult.ChatResponse, ir.Model)
-}
-
-// buildChatRequest 按模型 vendor 组装底层请求体。
-func buildChatRequest(ir aiwire.ChatRequest, stream bool) (generativeaiinference.BaseChatRequest, error) {
-	if isCohereModel(ir.Model) {
-		return irToCohereSDK(ir, stream)
-	}
-	return irToSDK(ir, stream), nil
-}
-
-// chatResponseToIR 按响应实际形态(GENERIC / COHERE)转回 IR。
-func chatResponseToIR(resp generativeaiinference.BaseChatResponse, model string) (*aiwire.ChatResponse, error) {
-	switch r := resp.(type) {
-	case generativeaiinference.GenericChatResponse:
-		return sdkToIR(r, model), nil
-	case generativeaiinference.CohereChatResponse:
-		return cohereSDKToIR(r, model), nil
-	default:
-		return nil, fmt.Errorf("genai chat: unexpected response format %T", resp)
-	}
-}
-
-func chatRequest(cred Credentials, modelOcid string, req generativeaiinference.BaseChatRequest) generativeaiinference.ChatRequest {
-	return generativeaiinference.ChatRequest{
-		ChatDetails: generativeaiinference.ChatDetails{
-			CompartmentId: &cred.TenancyOCID,
-			ServingMode:   generativeaiinference.OnDemandServingMode{ModelId: &modelOcid},
-			ChatRequest:   req,
-		},
-	}
-}
-
-// GenAiChatStream 实现 Client:流式聊天,返回逐事件读取器。
-// SDK 对 text/event-stream 跳过 unmarshal,原始流经 RawResponse 交由 SSEReader 消费。
-func (c *RealClient) GenAiChatStream(ctx context.Context, cred Credentials, region, modelOcid string, ir aiwire.ChatRequest) (GenAiStream, error) {
-	ic, err := c.genAiInferenceClient(cred, region)
-	if err != nil {
-		return nil, err
-	}
-	req, err := buildChatRequest(ir, true)
-	if err != nil {
-		return nil, err
-	}
-	resp, err := ic.Chat(ctx, chatRequest(cred, modelOcid, req))
-	if err != nil {
-		return nil, fmt.Errorf("genai chat stream: %w", err)
-	}
-	reader, err := common.NewSSEReader(resp.RawResponse)
-	if err != nil {
-		return nil, fmt.Errorf("genai chat stream: sse reader: %w", err)
-	}
-	return &genAiSSEStream{reader: reader, body: resp.RawResponse.Body, model: ir.Model}, nil
-}
-
-// genAiSSEStream 把 OCI SSE 事件流适配为 IR chunk 流。
-type genAiSSEStream struct {
-	reader *common.SseReader
-	body   io.ReadCloser
-	model  string
-}
-
-// Next 读取下一条有效事件;空事件与 [DONE] 哨兵跳过,流尽返回 io.EOF。
-func (s *genAiSSEStream) Next() (aiwire.ChatChunk, error) {
-	for {
-		data, err := s.reader.ReadNextEvent()
-		if err != nil {
-			return aiwire.ChatChunk{}, err
-		}
-		text := strings.TrimSpace(string(data))
-		if text == "" {
-			continue
-		}
-		if text == "[DONE]" {
-			return aiwire.ChatChunk{}, io.EOF
-		}
-		if chunk, ok := parseGenAiEvent([]byte(text), s.model); ok {
-			return chunk, nil
-		}
-	}
-}
-
-func (s *genAiSSEStream) Close() error {
-	if s.body != nil {
-		return s.body.Close()
-	}
-	return nil
-}
-
-// GenAiProbeChat 实现 Client:配额探测专用的最小聊天(maxTokens=1),返回 HTTP 状态码;
+// GenAiProbeChat 实现 Client:经 OpenAI 兼容面(直通同链路)发一次极小请求探测渠道;配额探测专用的最小聊天(maxTokens=1),返回 HTTP 状态码;
 // modelName 决定请求格式(cohere.* 走 COHERE)。
 func (c *RealClient) GenAiProbeChat(ctx context.Context, cred Credentials, region, modelOcid, modelName string) (int, error) {
-	one := 1
-	ir := aiwire.ChatRequest{
-		Model:     modelName,
-		Messages:  []aiwire.ChatMessage{{Role: "user", Content: aiwire.NewTextContent("hi")}},
-		MaxTokens: &one,
+	body, err := json.Marshal(map[string]any{"model": modelName, "input": "hi",
+		"max_output_tokens": 1, "store": false})
+	if err != nil {
+		return 0, err
 	}
-	_, err := c.GenAiChat(ctx, cred, region, modelOcid, ir)
-	if err == nil {
+	if _, err = c.GenAiCompatResponses(ctx, cred, region, body); err == nil {
 		return http.StatusOK, nil
 	}
 	if status, ok := ServiceStatus(err); ok {
@@ -255,6 +166,27 @@ func (c *RealClient) GenAiProbeChat(ctx context.Context, cred Credentials, regio
 	return 0, err
 }

+// sdkUsageToIR 把 SDK 用量转为内部记账结构(缓存命中挂 details,仅命中时出现)。
+func sdkUsageToIR(u *generativeaiinference.Usage) *aiwire.Usage {
+	if u == nil {
+		return nil
+	}
+	out := &aiwire.Usage{}
+	if u.PromptTokens != nil {
+		out.PromptTokens = *u.PromptTokens
+	}
+	if u.CompletionTokens != nil {
+		out.CompletionTokens = *u.CompletionTokens
+	}
+	if u.TotalTokens != nil {
+		out.TotalTokens = *u.TotalTokens
+	}
+	if u.PromptTokensDetails != nil && u.PromptTokensDetails.CachedTokens != nil && *u.PromptTokensDetails.CachedTokens > 0 {
+		out.PromptTokensDetails = &aiwire.PromptTokensDetails{CachedTokens: *u.PromptTokensDetails.CachedTokens}
+	}
+	return out
+}
+
 // GenAiEmbed 实现 Client:文本向量化(on-demand serving);dimensions 透传 OutputDimensions。
 func (c *RealClient) GenAiEmbed(ctx context.Context, cred Credentials, region, modelOcid string, inputs []string, dimensions *int) ([][]float32, *aiwire.Usage, error) {
 	ic, err := c.genAiInferenceClient(cred, region)
@@ -1,266 +0,0 @@
-package oci
-
-import (
-	"encoding/json"
-	"fmt"
-	"strings"
-
-	"github.com/oracle/oci-go-sdk/v65/generativeaiinference"
-
-	"oci-portal/internal/aiwire"
-)
-
-// isCohereModel 依据模型名前缀判定 COHERE 系(走专属请求格式)。
-func isCohereModel(name string) bool {
-	return strings.HasPrefix(strings.ToLower(name), "cohere.")
-}
-
-// irToCohereSDK 把 IR 转为 COHERE 聊天请求;工具与结构化输出做有损降级
-// (参数仅取 JSON Schema 顶层 properties,tool_choice / json_schema 的 name、strict 无对应)。
-func irToCohereSDK(ir aiwire.ChatRequest, stream bool) (generativeaiinference.CohereChatRequest, error) {
-	var req generativeaiinference.CohereChatRequest
-	if err := cohereRejectUnsupported(ir); err != nil {
-		return req, err
-	}
-	p, err := cohereSplitMessages(ir.Messages)
-	if err != nil {
-		return req, err
-	}
-	req = generativeaiinference.CohereChatRequest{
-		Message:          &p.message,
-		ChatHistory:      p.history,
-		ToolResults:      p.toolResults,
-		Tools:            cohereTools(ir.Tools),
-		ResponseFormat:   cohereResponseFormat(ir.ResponseFormat),
-		MaxTokens:        ir.MaxTokens,
-		Temperature:      ir.Temperature,
-		TopP:             ir.TopP,
-		TopK:             ir.TopK,
-		FrequencyPenalty: ir.FrequencyPenalty,
-		PresencePenalty:  ir.PresencePenalty,
-		Seed:             ir.Seed,
-	}
-	if p.preamble != "" {
-		req.PreambleOverride = &p.preamble
-	}
-	if len(ir.Stop) > 0 {
-		req.StopSequences = ir.Stop
-	}
-	if stream {
-		req.IsStream = &stream
-		includeUsage := ir.StreamOptions == nil || ir.StreamOptions.IncludeUsage
-		req.StreamOptions = &generativeaiinference.StreamOptions{IsIncludeUsage: &includeUsage}
-	}
-	return req, nil
-}
-
-// cohereRejectUnsupported 拒绝 COHERE 无法承接的能力(多模态图片)。
-func cohereRejectUnsupported(ir aiwire.ChatRequest) error {
-	for _, m := range ir.Messages {
-		for _, p := range m.Content.Parts {
-			if p.Type == "image_url" {
-				return fmt.Errorf("cohere 系模型暂不支持图片输入,请改用 meta/google 等多模态模型")
-			}
-		}
-	}
-	return nil
-}
-
-// cohereParts 是 IR 消息拆装为 COHERE 请求的中间结果。
-type cohereParts struct {
-	message     string
-	history     []generativeaiinference.CohereMessage
-	preamble    string
-	toolResults []generativeaiinference.CohereToolResult
-}
-
-// cohereSplitMessages 拆装消息:system 拼 preamble,末尾 user 抽为 message,
-// assistant(含 toolCalls)与其余 user 进 history,tool 结果转顶层 toolResults
-// (工具结果在末尾时 message 留空,COHERE 以 toolResults 续跑)。
-func cohereSplitMessages(msgs []aiwire.ChatMessage) (cohereParts, error) {
-	lastUser, lastTool := -1, -1
-	for i, m := range msgs {
-		switch m.Role {
-		case "user":
-			lastUser = i
-		case "tool":
-			lastTool = i
-		}
-	}
-	if lastUser == -1 && lastTool == -1 {
-		return cohereParts{}, fmt.Errorf("cohere 系模型至少需要一条 user 消息")
-	}
-	var p cohereParts
-	var preamble []string
-	calls := map[string]generativeaiinference.CohereToolCall{}
-	for i, m := range msgs {
-		text := m.Content.JoinText()
-		switch m.Role {
-		case "system", "developer":
-			preamble = append(preamble, text)
-		case "assistant":
-			p.history = append(p.history, cohereBotMessage(text, m.ToolCalls, calls))
-		case "tool":
-			p.toolResults = append(p.toolResults, cohereToolResult(m, calls))
-		default: // user
-			if i == lastUser && lastUser > lastTool {
-				p.message = text
-				continue
-			}
-			p.history = append(p.history, generativeaiinference.CohereUserMessage{Message: &text})
-		}
-	}
-	p.preamble = strings.Join(preamble, "\n")
-	return p, nil
-}
-
-// cohereBotMessage 转 assistant 消息;toolCalls 同步登记进 id→call 映射供 tool 结果回查。
-func cohereBotMessage(text string, tcs []aiwire.ToolCall, calls map[string]generativeaiinference.CohereToolCall) generativeaiinference.CohereChatBotMessage {
-	msg := generativeaiinference.CohereChatBotMessage{}
-	if text != "" {
-		msg.Message = &text
-	}
-	for _, tc := range tcs {
-		name := tc.Function.Name
-		var params interface{}
-		if json.Unmarshal([]byte(tc.Function.Arguments), &params) != nil || params == nil {
-			params = map[string]any{}
-		}
-		call := generativeaiinference.CohereToolCall{Name: &name, Parameters: &params}
-		calls[tc.ID] = call
-		msg.ToolCalls = append(msg.ToolCalls, call)
-	}
-	return msg
-}
-
-// cohereToolResult 把 IR tool 消息转顶层工具结果;COHERE 工具调用无 id,
-// 靠 assistant 历史登记的映射回查,缺失时以 tool_call_id 名义调用兜底。
-func cohereToolResult(m aiwire.ChatMessage, calls map[string]generativeaiinference.CohereToolCall) generativeaiinference.CohereToolResult {
-	call, ok := calls[m.ToolCallID]
-	if !ok {
-		name := m.ToolCallID
-		var params interface{} = map[string]any{}
-		call = generativeaiinference.CohereToolCall{Name: &name, Parameters: &params}
-	}
-	text := m.Content.JoinText()
-	var output interface{}
-	if json.Unmarshal([]byte(text), &output) != nil || output == nil {
-		output = map[string]any{"output": text}
-	}
-	if arr, isArr := output.([]interface{}); isArr {
-		return generativeaiinference.CohereToolResult{Call: &call, Outputs: arr}
-	}
-	if _, isMap := output.(map[string]interface{}); !isMap {
-		output = map[string]any{"output": output}
-	}
-	return generativeaiinference.CohereToolResult{Call: &call, Outputs: []interface{}{output}}
-}
-
-// cohereTools 把 JSON Schema 工具定义降级为 COHERE 扁平参数表(嵌套结构有损:仅取顶层)。
-func cohereTools(tools []aiwire.Tool) []generativeaiinference.CohereTool {
-	if len(tools) == 0 {
-		return nil
-	}
-	out := make([]generativeaiinference.CohereTool, 0, len(tools))
-	for _, t := range tools {
-		name, desc := t.Function.Name, t.Function.Description
-		if desc == "" {
-			desc = name // Description 为 COHERE 必填
-		}
-		out = append(out, generativeaiinference.CohereTool{
-			Name: &name, Description: &desc,
-			ParameterDefinitions: cohereParams(t.Function.Parameters),
-		})
-	}
-	return out
-}
-
-// cohereParams 取 JSON Schema 顶层 properties 转扁平参数定义;嵌套 schema 只保留类型名。
-func cohereParams(schema json.RawMessage) map[string]generativeaiinference.CohereParameterDefinition {
-	var s struct {
-		Properties map[string]struct {
-			Type        string `json:"type"`
-			Description string `json:"description"`
-		} `json:"properties"`
-		Required []string `json:"required"`
-	}
-	if json.Unmarshal(schema, &s) != nil || len(s.Properties) == 0 {
-		return nil
-	}
-	req := map[string]bool{}
-	for _, r := range s.Required {
-		req[r] = true
-	}
-	out := make(map[string]generativeaiinference.CohereParameterDefinition, len(s.Properties))
-	for k, v := range s.Properties {
-		typ, desc := v.Type, v.Description
-		if typ == "" {
-			typ = "string"
-		}
-		pd := generativeaiinference.CohereParameterDefinition{Type: &typ}
-		if desc != "" {
-			pd.Description = &desc
-		}
-		if req[k] {
-			t := true
-			pd.IsRequired = &t
-		}
-		out[k] = pd
-	}
-	return out
-}
-
-// cohereResponseFormat 映射 json_object / json_schema(name、strict 无对应,有损降级)。
-func cohereResponseFormat(rf *aiwire.ResponseFormat) generativeaiinference.CohereResponseFormat {
-	if rf == nil {
-		return nil
-	}
-	switch rf.Type {
-	case "json_object", "json_schema":
-		out := generativeaiinference.CohereResponseJsonFormat{}
-		var in struct {
-			Schema json.RawMessage `json:"schema"`
-		}
-		if json.Unmarshal(rf.JSONSchema, &in) == nil && len(in.Schema) > 0 {
-			var schema interface{}
-			if json.Unmarshal(in.Schema, &schema) == nil {
-				out.Schema = &schema
-			}
-		}
-		return out
-	}
-	return nil
-}
-
-// cohereSDKToIR 把 COHERE 非流式响应转回 IR;工具调用补派生 ID,存在时结束原因归 tool_calls。
-func cohereSDKToIR(resp generativeaiinference.CohereChatResponse, model string) *aiwire.ChatResponse {
-	msg := aiwire.ChatMessage{Role: "assistant", Content: aiwire.NewTextContent(deref(resp.Text))}
-	for i, tc := range resp.ToolCalls {
-		msg.ToolCalls = append(msg.ToolCalls, cohereCallToIR(deref(tc.Name), tc.Parameters, i))
-	}
-	finish := mapFinishReason(string(resp.FinishReason))
-	if len(msg.ToolCalls) > 0 {
-		finish = "tool_calls"
-	}
-	return &aiwire.ChatResponse{
-		Object:  "chat.completion",
-		Model:   model,
-		Choices: []aiwire.Choice{{Message: msg, FinishReason: finish}},
-		Usage:   sdkUsageToIR(resp.Usage),
-	}
-}
-
-// cohereCallToIR 生成 IR 工具调用;COHERE 无调用 id,按名称+序号派生(回传时按此回查)。
-func cohereCallToIR(name string, params *interface{}, idx int) aiwire.ToolCall {
-	args := "{}"
-	if params != nil && *params != nil {
-		if b, err := json.Marshal(*params); err == nil {
-			args = string(b)
-		}
-	}
-	return aiwire.ToolCall{
-		ID:       fmt.Sprintf("call_%s_%d", name, idx),
-		Type:     "function",
-		Function: aiwire.FunctionCall{Name: name, Arguments: args},
-	}
-}
@@ -1,367 +0,0 @@
-package oci
-
-import (
-	"encoding/json"
-	"strings"
-
-	"github.com/oracle/oci-go-sdk/v65/generativeaiinference"
-
-	"oci-portal/internal/aiwire"
-)
-
-// irToSDK 把 IR(OpenAI 线格式)转为 OCI GENERIC 聊天请求;字段近一一镜像。
-func irToSDK(ir aiwire.ChatRequest, stream bool) generativeaiinference.GenericChatRequest {
-	req := generativeaiinference.GenericChatRequest{
-		Messages:            irMessages(ir.Messages),
-		MaxTokens:           ir.MaxTokens,
-		MaxCompletionTokens: ir.MaxCompletionTokens,
-		Temperature:         ir.Temperature,
-		TopP:                ir.TopP,
-		TopK:                ir.TopK,
-		FrequencyPenalty:    ir.FrequencyPenalty,
-		PresencePenalty:     ir.PresencePenalty,
-		Seed:                ir.Seed,
-		NumGenerations:      ir.N,
-		Tools:               irTools(ir.Tools),
-		ToolChoice:          irToolChoice(ir.ToolChoice),
-		ResponseFormat:      irResponseFormat(ir.ResponseFormat),
-	}
-	if len(ir.Stop) > 0 {
-		req.Stop = ir.Stop
-	}
-	if ir.ReasoningEffort != "" {
-		req.ReasoningEffort = generativeaiinference.GenericChatRequestReasoningEffortEnum(strings.ToUpper(ir.ReasoningEffort))
-	}
-	if stream {
-		req.IsStream = &stream
-		includeUsage := ir.StreamOptions == nil || ir.StreamOptions.IncludeUsage
-		req.StreamOptions = &generativeaiinference.StreamOptions{IsIncludeUsage: &includeUsage}
-	}
-	return req
-}
-
-// irMessages 按 role 拆装消息;user 消息保留图文块,其余角色只取文本。
-func irMessages(msgs []aiwire.ChatMessage) []generativeaiinference.Message {
-	out := make([]generativeaiinference.Message, 0, len(msgs))
-	for _, m := range msgs {
-		switch m.Role {
-		case "system", "developer":
-			out = append(out, generativeaiinference.SystemMessage{Content: textContents(m.Content.JoinText())})
-		case "assistant":
-			out = append(out, generativeaiinference.AssistantMessage{
-				Content:   textContents(m.Content.JoinText()),
-				ToolCalls: irToolCalls(m.ToolCalls),
-			})
-		case "tool":
-			tm := generativeaiinference.ToolMessage{Content: textContents(m.Content.JoinText())}
-			if m.ToolCallID != "" {
-				id := m.ToolCallID
-				tm.ToolCallId = &id
-			}
-			out = append(out, tm)
-		default: // user
-			out = append(out, generativeaiinference.UserMessage{Content: chatContents(m.Content)})
-		}
-	}
-	return out
-}
-
-// chatContents 把 IR 内容转 SDK 内容块;文本直通,image_url 转 IMAGE 块(data URI / 公网地址)。
-func chatContents(c aiwire.Content) []generativeaiinference.ChatContent {
-	if len(c.Parts) == 0 {
-		return textContents(c.Text)
-	}
-	var out []generativeaiinference.ChatContent
-	for _, p := range c.Parts {
-		switch {
-		case p.Type == "image_url" && p.ImageURL != nil:
-			img := generativeaiinference.ImageUrl{Url: &p.ImageURL.URL}
-			if d, ok := generativeaiinference.GetMappingImageUrlDetailEnum(p.ImageURL.Detail); ok {
-				img.Detail = d
-			}
-			out = append(out, generativeaiinference.ImageContent{ImageUrl: &img})
-		case p.Text != "":
-			t := p.Text
-			out = append(out, generativeaiinference.TextContent{Text: &t})
-		}
-	}
-	return out
-}
-
-// textContents 把纯文本包装为单元素 TextContent 数组;空文本返回 nil。
-func textContents(text string) []generativeaiinference.ChatContent {
-	if text == "" {
-		return nil
-	}
-	return []generativeaiinference.ChatContent{generativeaiinference.TextContent{Text: &text}}
-}
-
-func irToolCalls(calls []aiwire.ToolCall) []generativeaiinference.ToolCall {
-	if len(calls) == 0 {
-		return nil
-	}
-	out := make([]generativeaiinference.ToolCall, 0, len(calls))
-	for _, c := range calls {
-		id, name, args := c.ID, c.Function.Name, c.Function.Arguments
-		out = append(out, generativeaiinference.FunctionCall{Id: &id, Name: &name, Arguments: &args})
-	}
-	return out
-}
-
-func irTools(tools []aiwire.Tool) []generativeaiinference.ToolDefinition {
-	if len(tools) == 0 {
-		return nil
-	}
-	out := make([]generativeaiinference.ToolDefinition, 0, len(tools))
-	for _, t := range tools {
-		name, desc := t.Function.Name, t.Function.Description
-		fd := generativeaiinference.FunctionDefinition{Name: &name}
-		if desc != "" {
-			fd.Description = &desc
-		}
-		if len(t.Function.Parameters) > 0 {
-			var params interface{}
-			if json.Unmarshal(t.Function.Parameters, &params) == nil {
-				fd.Parameters = &params
-			}
-		}
-		out = append(out, fd)
-	}
-	return out
-}
-
-// irToolChoice 解析 "auto"/"none"/"required" 或 {type:function,function:{name}}。
-func irToolChoice(raw json.RawMessage) generativeaiinference.ToolChoice {
-	if len(raw) == 0 {
-		return nil
-	}
-	var s string
-	if json.Unmarshal(raw, &s) == nil {
-		switch s {
-		case "none":
-			return generativeaiinference.ToolChoiceNone{}
-		case "required":
-			return generativeaiinference.ToolChoiceRequired{}
-		case "auto":
-			return generativeaiinference.ToolChoiceAuto{}
-		}
-		return nil
-	}
-	var obj struct {
-		Function struct {
-			Name string `json:"name"`
-		} `json:"function"`
-	}
-	if json.Unmarshal(raw, &obj) == nil && obj.Function.Name != "" {
-		return generativeaiinference.ToolChoiceFunction{Name: &obj.Function.Name}
-	}
-	return nil
-}
-
-func irResponseFormat(rf *aiwire.ResponseFormat) generativeaiinference.ResponseFormat {
-	if rf == nil {
-		return nil
-	}
-	switch rf.Type {
-	case "json_object":
-		return generativeaiinference.JsonObjectResponseFormat{}
-	case "json_schema":
-		return irJSONSchemaFormat(rf.JSONSchema)
-	default:
-		return nil
-	}
-}
-
-// irJSONSchemaFormat 映射 OpenAI json_schema{name,description,schema,strict}。
-func irJSONSchemaFormat(raw json.RawMessage) generativeaiinference.ResponseFormat {
-	var in struct {
-		Name        string          `json:"name"`
-		Description string          `json:"description"`
-		Schema      json.RawMessage `json:"schema"`
-		Strict      *bool           `json:"strict"`
-	}
-	if json.Unmarshal(raw, &in) != nil || in.Name == "" {
-		return generativeaiinference.JsonObjectResponseFormat{}
-	}
-	js := generativeaiinference.ResponseJsonSchema{Name: &in.Name, IsStrict: in.Strict}
-	if in.Description != "" {
-		js.Description = &in.Description
-	}
-	if len(in.Schema) > 0 {
-		var schema interface{}
-		if json.Unmarshal(in.Schema, &schema) == nil {
-			js.Schema = &schema
-		}
-	}
-	return generativeaiinference.JsonSchemaResponseFormat{JsonSchema: &js}
-}
-
-// sdkToIR 把非流式 GenericChatResponse 转回 IR 响应;id/created 由调用方补。
-func sdkToIR(resp generativeaiinference.GenericChatResponse, model string) *aiwire.ChatResponse {
-	out := &aiwire.ChatResponse{Object: "chat.completion", Model: model}
-	if resp.TimeCreated != nil {
-		out.Created = resp.TimeCreated.Unix()
-	}
-	for i, ch := range resp.Choices {
-		idx := i
-		if ch.Index != nil {
-			idx = *ch.Index
-		}
-		out.Choices = append(out.Choices, aiwire.Choice{
-			Index:        idx,
-			Message:      sdkMessageToIR(ch.Message),
-			FinishReason: mapFinishReason(deref(ch.FinishReason)),
-		})
-	}
-	out.Usage = sdkUsageToIR(resp.Usage)
-	return out
-}
-
-func sdkUsageToIR(u *generativeaiinference.Usage) *aiwire.Usage {
-	if u == nil {
-		return nil
-	}
-	out := &aiwire.Usage{}
-	if u.PromptTokens != nil {
-		out.PromptTokens = *u.PromptTokens
-	}
-	if u.CompletionTokens != nil {
-		out.CompletionTokens = *u.CompletionTokens
-	}
-	if u.TotalTokens != nil {
-		out.TotalTokens = *u.TotalTokens
-	}
-	if d := u.PromptTokensDetails; d != nil && d.CachedTokens != nil && *d.CachedTokens > 0 {
-		out.PromptTokensDetails = &aiwire.PromptTokensDetails{CachedTokens: *d.CachedTokens}
-	}
-	return out
-}
-
-// sdkMessageToIR 提取助手消息文本与工具调用。
-func sdkMessageToIR(m generativeaiinference.Message) aiwire.ChatMessage {
-	out := aiwire.ChatMessage{Role: "assistant"}
-	am, ok := m.(generativeaiinference.AssistantMessage)
-	if !ok {
-		return out
-	}
-	var sb strings.Builder
-	for _, c := range am.Content {
-		if tc, ok := c.(generativeaiinference.TextContent); ok && tc.Text != nil {
-			sb.WriteString(*tc.Text)
-		}
-	}
-	out.Content = aiwire.NewTextContent(sb.String())
-	for _, call := range am.ToolCalls {
-		if fc, ok := call.(generativeaiinference.FunctionCall); ok {
-			out.ToolCalls = append(out.ToolCalls, aiwire.ToolCall{
-				ID:   deref(fc.Id),
-				Type: "function",
-				Function: aiwire.FunctionCall{
-					Name:      deref(fc.Name),
-					Arguments: deref(fc.Arguments),
-				},
-			})
-		}
-	}
-	return out
-}
-
-// mapFinishReason 归一化结束原因;OCI 与 OpenAI 取值同名,做小写与别名兜底。
-func mapFinishReason(reason string) string {
-	switch strings.ToLower(reason) {
-	case "", "null":
-		return ""
-	case "stop", "completed", "end_turn":
-		return "stop"
-	case "length", "max_tokens":
-		return "length"
-	case "tool_calls", "tool_call", "tool_use":
-		return "tool_calls"
-	case "complete":
-		return "stop"
-	case "error_toxic":
-		return "content_filter"
-	default:
-		return strings.ToLower(reason)
-	}
-}
-
-// genAiStreamEvent 是 GENERIC 流式事件的宽容解析结构(字段按增量 choice 形状)。
-type genAiStreamEvent struct {
-	Index   *int `json:"index"`
-	Message struct {
-		Role    string `json:"role"`
-		Content []struct {
-			Type string `json:"type"`
-			Text string `json:"text"`
-		} `json:"content"`
-		ToolCalls []struct {
-			ID        string `json:"id"`
-			Name      string `json:"name"`
-			Arguments string `json:"arguments"`
-		} `json:"toolCalls"`
-	} `json:"message"`
-	FinishReason *string                      `json:"finishReason"`
-	Usage        *generativeaiinference.Usage `json:"usage"`
-	Choices      []json.RawMessage            `json:"choices"` // 兼容整包 choices 形态
-	Text         string                       `json:"text"`    // COHERE 流事件的顶层增量文本
-	// CohereCalls 是 COHERE 流事件顶层的工具调用(与 GENERIC 的 message.toolCalls 不同层级)
-	CohereCalls []struct {
-		Name       string      `json:"name"`
-		Parameters interface{} `json:"parameters"`
-	} `json:"toolCalls"`
-}
-
-// parseGenAiEvent 把一条 SSE 事件 JSON 解析为 IR chunk;无有效负载时 ok=false。
-func parseGenAiEvent(data []byte, model string) (aiwire.ChatChunk, bool) {
-	var ev genAiStreamEvent
-	if err := json.Unmarshal(data, &ev); err != nil {
-		return aiwire.ChatChunk{}, false
-	}
-	if len(ev.Choices) > 0 { // choices 包裹形态:取首个展开重解析
-		var inner genAiStreamEvent
-		if json.Unmarshal(ev.Choices[0], &inner) == nil {
-			inner.Usage = ev.Usage
-			ev = inner
-		}
-	}
-	chunk := aiwire.ChatChunk{Object: "chat.completion.chunk", Model: model}
-	choice := aiwire.ChunkChoice{}
-	if ev.Index != nil {
-		choice.Index = *ev.Index
-	}
-	choice.Delta.Role = strings.ToLower(ev.Message.Role)
-	var sb strings.Builder
-	for _, c := range ev.Message.Content {
-		sb.WriteString(c.Text)
-	}
-	if sb.Len() == 0 && ev.Text != "" { // COHERE 事件形态
-		sb.WriteString(ev.Text)
-	}
-	choice.Delta.Content = sb.String()
-	for i, tc := range ev.Message.ToolCalls {
-		d := aiwire.ToolCallDelta{Index: i, ID: tc.ID}
-		if tc.ID != "" {
-			d.Type = "function"
-		}
-		d.Function.Name = tc.Name
-		d.Function.Arguments = tc.Arguments
-		choice.Delta.ToolCalls = append(choice.Delta.ToolCalls, d)
-	}
-	for i, tc := range ev.CohereCalls { // COHERE 顶层工具调用:整包出现,补派生 ID
-		call := cohereCallToIR(tc.Name, &tc.Parameters, i)
-		d := aiwire.ToolCallDelta{Index: len(choice.Delta.ToolCalls) + i, ID: call.ID, Type: "function"}
-		d.Function.Name = call.Function.Name
-		d.Function.Arguments = call.Function.Arguments
-		choice.Delta.ToolCalls = append(choice.Delta.ToolCalls, d)
-	}
-	if r := mapFinishReason(deref(ev.FinishReason)); r != "" {
-		choice.FinishReason = &r
-	}
-	hasPayload := choice.Delta.Content != "" || len(choice.Delta.ToolCalls) > 0 || choice.FinishReason != nil
-	if hasPayload {
-		chunk.Choices = []aiwire.ChunkChoice{choice}
-	}
-	chunk.Usage = sdkUsageToIR(ev.Usage)
-	return chunk, hasPayload || chunk.Usage != nil
-}
@@ -0,0 +1,496 @@
+package oci
+
+// 真实 OCI GenAI effort 探针(重建版)。
+//
+// 前身 genai_cache_integration_test.go 承载缓存/亲和/服务端工具等历轮调研模式,
+// 调研结论均已归档(.trellis/tasks/archive/2026-07/*/research/),该文件已删除;
+// 本文件重建共享基础设施,保留 effort 支持矩阵探测,并新增 responses 端点探测
+// (multi-agent 模型仅允许走 Responses 面)。
+//
+// 运行(默认跳过,不触网):
+//
+//	OCI_GENAI_CACHE_PROBE=effort-matrix OCI_CACHE_DB=<db 绝对路径> DATA_KEY=<主密钥> \
+//	OCI_CACHE_RUN_ID=<唯一标识> go test ./internal/oci/ -run TestRealGenAiEffortMatrix -v
+//
+// 可选:OCI_CACHE_CONFIG_ID(默认 6)、OCI_CACHE_REGION(默认 us-chicago-1)、
+// OCI_EFFORT_CASES 自定义用例("model=effort" 逗号分隔;effort 后缀 "@responses"
+// 表示走 /actions/v1/responses 端点,如 "xai.grok-4.20-multi-agent=low@responses")。
+//
+// 纪律:数据库只读打开;OCI_GO_SDK_DEBUG 必须为空;日志不落原始请求体与凭据,
+// 错误消息经 OCID 遮掩后截断。
+
+import (
+	"context"
+	"encoding/json"
+	"fmt"
+	"io"
+	"net/http"
+	"os"
+	"regexp"
+	"strings"
+	"testing"
+	"time"
+
+	"github.com/glebarez/sqlite"
+	"github.com/oracle/oci-go-sdk/v65/common"
+	"github.com/oracle/oci-go-sdk/v65/generativeaiinference"
+	"gorm.io/gorm"
+	"gorm.io/gorm/logger"
+
+	projectcrypto "oci-portal/internal/crypto"
+	"oci-portal/internal/model"
+)
+
+const (
+	effortProbeRegion = "us-chicago-1"
+	effortProbeLimit  = int64(4 << 20)
+)
+
+var effortRunIDPattern = regexp.MustCompile(`^[A-Za-z0-9._-]{1,64}$`)
+
+type effortProbeCase struct {
+	model     string
+	effort    string
+	responses bool // true 走 /actions/v1/responses,否则 typed /actions/chat
+	stream    bool // 仅 responses 面:请求 SSE 流,观测事件类型序列
+}
+
+type effortProbeResources struct {
+	cred      Credentials
+	region    string
+	modelOCID map[string]string
+	inference generativeaiinference.GenerativeAiInferenceClient
+}
+
+type effortProbeResult struct {
+	status           int
+	requestRef       string
+	duration         time.Duration
+	errorCode        string
+	errorMessage     string
+	completionTokens int
+	reasoningTokens  int
+	answerLen        int
+	outputTypes      []string
+}
+
+// effortMatrixCases 默认批为 grok-4.3 五档基线;全模型矩阵结论见任务归档,
+// 复测用 OCI_EFFORT_CASES 自定义。
+func effortMatrixCases() []effortProbeCase {
+	if raw := os.Getenv("OCI_EFFORT_CASES"); raw != "" {
+		return parseEffortCases(raw)
+	}
+	return []effortProbeCase{
+		{model: "xai.grok-4.3", effort: "none", responses: true},
+		{model: "xai.grok-4.3", effort: "low", responses: true},
+		{model: "xai.grok-4.3", effort: "high", responses: true},
+	}
+}
+
+// parseEffortCases 解析 "model=effort[@responses|@responses-stream]" 逗号分隔用例。
+func parseEffortCases(raw string) []effortProbeCase {
+	var out []effortProbeCase
+	for _, item := range strings.Split(raw, ",") {
+		parts := strings.SplitN(strings.TrimSpace(item), "=", 2)
+		if len(parts) != 2 || parts[0] == "" || parts[1] == "" {
+			continue
+		}
+		effort, streaming := strings.CutSuffix(parts[1], "@responses-stream")
+		if streaming {
+			out = append(out, effortProbeCase{model: parts[0], effort: effort, responses: true, stream: true})
+			continue
+		}
+		effort, _ = strings.CutSuffix(effort, "@responses") // 后缀兼容保留;所有用例均走 responses 面
+		out = append(out, effortProbeCase{model: parts[0], effort: effort, responses: true})
+	}
+	return out
+}
+
+func TestRealGenAiEffortMatrix(t *testing.T) {
+	requireEffortProbeMode(t, "effort-matrix")
+	ctx, cancel := context.WithTimeout(context.Background(), 20*time.Minute)
+	defer cancel()
+	resources := newEffortProbeResources(t, ctx)
+	for _, tc := range effortMatrixCases() {
+		runEffortCase(t, ctx, resources, tc)
+	}
+}
+
+func requireEffortProbeMode(t *testing.T, wanted string) {
+	t.Helper()
+	if os.Getenv("OCI_GENAI_CACHE_PROBE") != wanted {
+		t.Skipf("set OCI_GENAI_CACHE_PROBE=%s to run", wanted)
+	}
+	if os.Getenv("OCI_GO_SDK_DEBUG") != "" {
+		t.Fatal("OCI_GO_SDK_DEBUG must be unset to protect signed requests")
+	}
+	if !effortRunIDPattern.MatchString(os.Getenv("OCI_CACHE_RUN_ID")) {
+		t.Fatal("set a unique OCI_CACHE_RUN_ID (1-64 safe characters)")
+	}
+}
+
+func runEffortCase(t *testing.T, ctx context.Context, resources effortProbeResources, tc effortProbeCase) {
+	t.Helper()
+	ocid, ok := resources.modelOCID[tc.model]
+	if !ok {
+		t.Logf("model=%s effort=%s resolve-failed: model not in on-demand catalog", tc.model, tc.effort)
+		return
+	}
+	_ = ocid
+	body, path, err := effortProbeBody(tc)
+	if err != nil {
+		t.Fatalf("build effort body: %v", err)
+	}
+	result := executeEffortProbe(ctx, resources, path, body)
+	logEffortResult(t, tc, result)
+}
+
+// effortProbeBody 构造 OpenAI Responses 直通请求体(typed chat 面已随网关剔除,
+// 探针仅保留 responses 端点;历史 typed 结论见任务归档)。
+func effortProbeBody(tc effortProbeCase) ([]byte, string, error) {
+	question := "How many positive divisors does 360 have? Answer with the number only."
+	body := map[string]interface{}{"model": tc.model, "input": question,
+		"max_output_tokens": 2048, "store": false}
+	if tc.effort != "" {
+		body["reasoning"] = map[string]string{"effort": tc.effort}
+	}
+	if tc.stream {
+		body["stream"] = true
+	}
+	payload, err := json.Marshal(body)
+	return payload, "/actions/v1/responses", err
+}
+
+// executeEffortProbe 以 IAM 签名裸 POST 指定路径;responses 面按生产实现补 compartment 头。
+func executeEffortProbe(ctx context.Context, resources effortProbeResources, path string, body []byte) effortProbeResult {
+	client := resources.inference.BaseClient
+	client.Configuration.CircuitBreaker = nil
+	common.UpdateEndpointTemplateForOptions(&client)
+	common.SetMissingTemplateParams(&client)
+	request, err := http.NewRequestWithContext(ctx, http.MethodPost, path, strings.NewReader(string(body)))
+	if err != nil {
+		return effortProbeResult{errorCode: "BuildRequest", errorMessage: err.Error()}
+	}
+	request.Header.Set("Content-Type", "application/json")
+	if path != "/actions/chat" {
+		request.Header.Set("CompartmentId", resources.cred.TenancyOCID)
+		request.Header.Set("opc-compartment-id", resources.cred.TenancyOCID)
+	}
+	return effortCallProbe(ctx, client, request)
+}
+
+func effortCallProbe(ctx context.Context, client common.BaseClient, request *http.Request) effortProbeResult {
+	started := time.Now()
+	response, err := client.Call(ctx, request)
+	result := effortProbeResult{duration: time.Since(started)}
+	if response != nil {
+		result.status = response.StatusCode
+		result.requestRef = shortEffortRef(response.Header.Get("opc-request-id"))
+		defer response.Body.Close()
+	}
+	if err != nil {
+		result.errorCode, result.errorMessage = effortProbeError(err)
+		return result
+	}
+	payload, readErr := io.ReadAll(io.LimitReader(response.Body, effortProbeLimit))
+	if readErr != nil {
+		result.errorCode, result.errorMessage = "ReadResponse", readErr.Error()
+		return result
+	}
+	parseEffortPayload(&result, payload)
+	return result
+}
+
+func effortProbeError(err error) (string, string) {
+	if serviceErr, ok := common.IsServiceError(err); ok {
+		return fmt.Sprintf("%d", serviceErr.GetHTTPStatusCode()), sanitizeEffortText(serviceErr.GetMessage())
+	}
+	return "CallError", sanitizeEffortText(err.Error())
+}
+
+// parseEffortPayload 兼容三种 usage 形态:typed 驼峰(chatResponse.usage.completionTokens)、
+// chat 兼容面下划线(usage.completion_tokens)、responses 面(usage.output_tokens)。
+func parseEffortPayload(result *effortProbeResult, payload []byte) {
+	if text := strings.TrimSpace(string(payload)); strings.HasPrefix(text, "event:") || strings.HasPrefix(text, "data:") {
+		parseEffortSSE(result, text)
+		return
+	}
+	var root map[string]interface{}
+	if json.Unmarshal(payload, &root) != nil {
+		return
+	}
+	result.answerLen = len(parseEffortAnswer(root))
+	result.outputTypes = parseEffortOutputTypes(root)
+	if typed, ok := root["chatResponse"].(map[string]interface{}); ok {
+		root = typed
+	}
+	usage, _ := root["usage"].(map[string]interface{})
+	for _, key := range []string{"completionTokens", "completion_tokens", "output_tokens"} {
+		if v, ok := effortInt(usage, key); ok {
+			result.completionTokens = v
+			break
+		}
+	}
+	for _, detailsKey := range []string{"completionTokensDetails", "completion_tokens_details", "output_tokens_details"} {
+		details, _ := usage[detailsKey].(map[string]interface{})
+		for _, key := range []string{"reasoningTokens", "reasoning_tokens"} {
+			if v, ok := effortInt(details, key); ok {
+				result.reasoningTokens = v
+			}
+		}
+	}
+}
+
+// parseEffortAnswer 提取正文文本:typed 取 chatResponse.choices[].message.content[].text,
+// responses 取 output[] 里 message 项的 content[].text。仅在内存中量长度,不落日志。
+func parseEffortAnswer(root map[string]interface{}) string {
+	var sb strings.Builder
+	if typed, ok := root["chatResponse"].(map[string]interface{}); ok {
+		choices, _ := typed["choices"].([]interface{})
+		for _, c := range choices {
+			cm, _ := c.(map[string]interface{})
+			msg, _ := cm["message"].(map[string]interface{})
+			collectEffortText(&sb, msg["content"])
+		}
+		return sb.String()
+	}
+	output, _ := root["output"].([]interface{})
+	for _, item := range output {
+		im, _ := item.(map[string]interface{})
+		if im["type"] == "message" {
+			collectEffortText(&sb, im["content"])
+		}
+	}
+	return sb.String()
+}
+
+func collectEffortText(sb *strings.Builder, content interface{}) {
+	parts, _ := content.([]interface{})
+	for _, p := range parts {
+		pm, _ := p.(map[string]interface{})
+		if text, ok := pm["text"].(string); ok {
+			sb.WriteString(text)
+		}
+	}
+}
+
+// parseEffortSSE 汇总 SSE 流:事件类型去重序列进 outputTypes,正文增量长度进 answerLen,
+// 末尾 completed 事件里的 usage 进 tokens 字段。
+func parseEffortSSE(result *effortProbeResult, text string) {
+	seen := map[string]bool{}
+	for _, line := range strings.Split(text, "\n") {
+		if name, ok := strings.CutPrefix(line, "event: "); ok {
+			if name = strings.TrimSpace(name); !seen[name] {
+				seen[name] = true
+				result.outputTypes = append(result.outputTypes, name)
+			}
+			continue
+		}
+		data, ok := strings.CutPrefix(line, "data: ")
+		if !ok {
+			continue
+		}
+		var ev map[string]interface{}
+		if json.Unmarshal([]byte(data), &ev) != nil {
+			continue
+		}
+		if name, ok := ev["type"].(string); ok && !seen[name] { // 纯 data: 行流(无 event: 行)从事件体取类型
+			seen[name] = true
+			result.outputTypes = append(result.outputTypes, name)
+		}
+		if delta, ok := ev["delta"].(string); ok && strings.Contains(fmt.Sprint(ev["type"]), "output_text") {
+			result.answerLen += len(delta)
+		}
+		if resp, ok := ev["response"].(map[string]interface{}); ok {
+			usage, _ := resp["usage"].(map[string]interface{})
+			if v, ok := effortInt(usage, "output_tokens"); ok {
+				result.completionTokens = v
+			}
+		}
+	}
+}
+
+// parseEffortOutputTypes 收集 responses 面 output 项类型序列(multi-agent 行为观测)。
+func parseEffortOutputTypes(root map[string]interface{}) []string {
+	output, _ := root["output"].([]interface{})
+	var types []string
+	for _, item := range output {
+		im, _ := item.(map[string]interface{})
+		if s, ok := im["type"].(string); ok {
+			types = append(types, s)
+		}
+	}
+	return types
+}
+
+func effortInt(root map[string]interface{}, key string) (int, bool) {
+	if root == nil {
+		return 0, false
+	}
+	if value, ok := root[key].(float64); ok {
+		return int(value), true
+	}
+	return 0, false
+}
+
+func logEffortResult(t *testing.T, tc effortProbeCase, result effortProbeResult) {
+	t.Helper()
+	endpoint := "typed-chat"
+	if tc.responses {
+		endpoint = "compat-responses"
+	}
+	t.Logf("endpoint=%s model=%s effort=%s status=%d completion=%d reasoning=%d answer_len=%d output_types=%v request=%s duration_ms=%d error=%s message=%q",
+		endpoint, tc.model, tc.effort, result.status, result.completionTokens, result.reasoningTokens,
+		result.answerLen, result.outputTypes, result.requestRef, result.duration.Milliseconds(),
+		result.errorCode, result.errorMessage)
+}
+
+// sanitizeEffortText 遮掩 OCID 并截断,避免错误消息携带租户可定位信息。
+func sanitizeEffortText(text string) string {
+	text = regexp.MustCompile(`ocid1\.[a-z0-9._-]+`).ReplaceAllString(text, "ocid1.***")
+	if len(text) > 300 {
+		text = text[:300] + "..."
+	}
+	return text
+}
+
+func shortEffortRef(ref string) string {
+	if len(ref) > 12 {
+		return ref[:12]
+	}
+	return ref
+}
+
+func effortEnvOr(key, fallback string) string {
+	if value := os.Getenv(key); value != "" {
+		return value
+	}
+	return fallback
+}
+
+// newEffortProbeResources 装配凭据、区域、inference 客户端与按需模型目录。
+func newEffortProbeResources(t *testing.T, ctx context.Context) effortProbeResources {
+	t.Helper()
+	cred := effortProbeCredentials(t)
+	region := effortEnvOr("OCI_CACHE_REGION", effortProbeRegion)
+	cred.Region = region
+	real := &RealClient{}
+	ic, err := real.genAiInferenceClient(cred, region)
+	if err != nil {
+		t.Fatalf("new inference client: %v", err)
+	}
+	models, err := real.ListGenAiModels(ctx, cred, region)
+	if err != nil {
+		t.Fatalf("list models: %s", sanitizeEffortText(err.Error()))
+	}
+	index := make(map[string]string, len(models))
+	for _, m := range models {
+		index[m.Name] = m.Ocid
+	}
+	return effortProbeResources{cred: cred, region: region, modelOCID: index, inference: ic}
+}
+
+// effortProbeCredentials 优先从面板数据库(只读)取渠道凭据,否则退回本地 ini 测试凭据。
+func effortProbeCredentials(t *testing.T) Credentials {
+	t.Helper()
+	if dbPath := os.Getenv("OCI_CACHE_DB"); dbPath != "" {
+		return effortDBCredentials(t, dbPath)
+	}
+	return loadTestCredentials(t, effortEnvOr("OCI_TEST_KEY", "试用期"))
+}
+
+func effortDBCredentials(t *testing.T, dbPath string) Credentials {
+	t.Helper()
+	dsn := fmt.Sprintf("file:%s?mode=ro", dbPath)
+	db, err := gorm.Open(sqlite.Open(dsn), &gorm.Config{Logger: logger.Default.LogMode(logger.Silent)})
+	if err != nil {
+		t.Fatalf("open read-only probe database: %v", err)
+	}
+	cipher, err := projectcrypto.NewCipher(os.Getenv("DATA_KEY"))
+	if err != nil {
+		t.Fatalf("create data cipher: %v", err)
+	}
+	var config model.OciConfig
+	if err := db.First(&config, effortEnvOr("OCI_CACHE_CONFIG_ID", "6")).Error; err != nil {
+		t.Fatalf("load OCI config: %v", err)
+	}
+	return decryptEffortCredentials(t, db, cipher, config)
+}
+
+func decryptEffortCredentials(t *testing.T, db *gorm.DB, cipher *projectcrypto.Cipher, config model.OciConfig) Credentials {
+	t.Helper()
+	privateKey, err := cipher.DecryptString(config.PrivateKeyEnc)
+	if err != nil {
+		t.Fatalf("decrypt OCI config %d private key: %v", config.ID, err)
+	}
+	passphrase := ""
+	if config.PassphraseEnc != "" {
+		if passphrase, err = cipher.DecryptString(config.PassphraseEnc); err != nil {
+			t.Fatalf("decrypt OCI config %d passphrase: %v", config.ID, err)
+		}
+	}
+	return Credentials{TenancyOCID: config.TenancyOCID, UserOCID: config.UserOCID,
+		Fingerprint: config.Fingerprint, Region: config.Region, PrivateKey: privateKey,
+		Passphrase: passphrase, Proxy: loadEffortProxy(t, db, cipher, config.ProxyID)}
+}
+
+func loadEffortProxy(t *testing.T, db *gorm.DB, cipher *projectcrypto.Cipher, proxyID *uint) *ProxySpec {
+	t.Helper()
+	if proxyID == nil {
+		return nil
+	}
+	var proxy model.Proxy
+	if err := db.First(&proxy, *proxyID).Error; err != nil {
+		t.Fatalf("load proxy %d: %v", *proxyID, err)
+	}
+	password := ""
+	if proxy.PasswordEnc != "" {
+		decrypted, err := cipher.DecryptString(proxy.PasswordEnc)
+		if err != nil {
+			t.Fatalf("decrypt proxy %d password: %v", *proxyID, err)
+		}
+		password = decrypted
+	}
+	return &ProxySpec{Type: proxy.Type, Host: proxy.Host, Port: proxy.Port,
+		Username: proxy.Username, Password: password}
+}
+
+// TestParseEffortCases 断言自定义用例解析:端点后缀、空项与畸形项跳过。
+func TestParseEffortCases(t *testing.T) {
+	got := parseEffortCases("a=low, b=high@responses ,bad,=x,c=")
+	want := []effortProbeCase{{model: "a", effort: "low", responses: true}, {model: "b", effort: "high", responses: true}}
+	if len(got) != len(want) {
+		t.Fatalf("parseEffortCases = %+v, want %+v", got, want)
+	}
+	for i := range want {
+		if got[i] != want[i] {
+			t.Fatalf("case %d = %+v, want %+v", i, got[i], want[i])
+		}
+	}
+}
+
+// TestParseEffortPayload 断言三种 usage 形态与 output 类型序列的解析。
+func TestParseEffortPayload(t *testing.T) {
+	cases := []struct {
+		name       string
+		payload    string
+		completion int
+		reasoning  int
+		types      int
+	}{
+		{"typed 驼峰", `{"chatResponse":{"choices":[{"message":{"content":[{"type":"TEXT","text":"24"}]}}],"usage":{"completionTokens":5,"completionTokensDetails":{"reasoningTokens":9}}}}`, 5, 9, 0},
+		{"responses 面", `{"output":[{"type":"reasoning"},{"type":"message","content":[{"type":"output_text","text":"24"}]}],"usage":{"output_tokens":7,"output_tokens_details":{"reasoning_tokens":3}}}`, 7, 3, 2},
+	}
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			var result effortProbeResult
+			parseEffortPayload(&result, []byte(tc.payload))
+			if result.completionTokens != tc.completion || result.reasoningTokens != tc.reasoning || len(result.outputTypes) != tc.types {
+				t.Fatalf("parse = %+v, want completion=%d reasoning=%d types=%d", result, tc.completion, tc.reasoning, tc.types)
+			}
+			if result.answerLen != 2 {
+				t.Fatalf("answerLen = %d, want 2", result.answerLen)
+			}
+		})
+	}
+}
@@ -0,0 +1,127 @@
+package oci
+
+import (
+	"context"
+	"fmt"
+
+	"github.com/oracle/oci-go-sdk/v65/common"
+	"github.com/oracle/oci-go-sdk/v65/generativeaiinference"
+)
+
+// RerankRank 是重排结果的一项:index 指向入参 documents 下标。
+type RerankRank struct {
+	Index int
+	Score float64
+}
+
+// GuardrailCategory 是内容审核分类得分(OVERALL / BLOCKLIST)。
+type GuardrailCategory struct {
+	Name  string
+	Score float64
+}
+
+// GuardrailPiiHit 是一处 PII 命中(片段原文与位置)。
+type GuardrailPiiHit struct {
+	Text   string
+	Label  string
+	Score  float64
+	Offset int
+	Length int
+}
+
+// GuardrailsOutcome 汇总 ApplyGuardrails 三能力结果。
+type GuardrailsOutcome struct {
+	Categories           []GuardrailCategory
+	Pii                  []GuardrailPiiHit
+	PromptInjectionScore *float64
+}
+
+// GenAiRerank 实现 Client:文档重排(on-demand serving),返回按相关度排序的下标与得分。
+func (c *RealClient) GenAiRerank(ctx context.Context, cred Credentials, region, modelOcid, query string, documents []string, topN *int) ([]RerankRank, error) {
+	ic, err := c.genAiInferenceClient(cred, region)
+	if err != nil {
+		return nil, err
+	}
+	resp, err := ic.RerankText(ctx, generativeaiinference.RerankTextRequest{
+		RerankTextDetails: generativeaiinference.RerankTextDetails{
+			CompartmentId: &cred.TenancyOCID,
+			ServingMode:   generativeaiinference.OnDemandServingMode{ModelId: &modelOcid},
+			Input:         &query,
+			Documents:     documents,
+			TopN:          topN,
+		},
+	})
+	if err != nil {
+		return nil, fmt.Errorf("genai rerank: %w", err)
+	}
+	ranks := make([]RerankRank, 0, len(resp.DocumentRanks))
+	for _, r := range resp.DocumentRanks {
+		if r.Index == nil || r.RelevanceScore == nil {
+			continue
+		}
+		ranks = append(ranks, RerankRank{Index: *r.Index, Score: *r.RelevanceScore})
+	}
+	return ranks, nil
+}
+
+// GenAiApplyGuardrails 实现 Client:对单条文本执行内容审核 / PII / 提示注入三检测。
+func (c *RealClient) GenAiApplyGuardrails(ctx context.Context, cred Credentials, region, text string) (*GuardrailsOutcome, error) {
+	ic, err := c.genAiInferenceClient(cred, region)
+	if err != nil {
+		return nil, err
+	}
+	resp, err := ic.ApplyGuardrails(ctx, generativeaiinference.ApplyGuardrailsRequest{
+		ApplyGuardrailsDetails: generativeaiinference.ApplyGuardrailsDetails{
+			CompartmentId: &cred.TenancyOCID,
+			Input:         generativeaiinference.GuardrailsTextInput{Content: common.String(text)},
+			GuardrailConfigs: &generativeaiinference.GuardrailConfigs{
+				ContentModerationConfig:                 &generativeaiinference.ContentModerationConfiguration{Categories: []string{"OVERALL"}},
+				PersonallyIdentifiableInformationConfig: &generativeaiinference.PersonallyIdentifiableInformationConfiguration{Types: []string{}},
+				PromptInjectionConfig:                   &generativeaiinference.PromptInjectionConfiguration{},
+			},
+		},
+	})
+	if err != nil {
+		return nil, fmt.Errorf("genai guardrails: %w", err)
+	}
+	return guardrailsOutcome(resp.Results), nil
+}
+
+// guardrailsOutcome 把 SDK 结果换算为 IR;空段置空切片,得分缺失跳过。
+func guardrailsOutcome(r *generativeaiinference.GuardrailsResults) *GuardrailsOutcome {
+	out := &GuardrailsOutcome{}
+	if r == nil {
+		return out
+	}
+	if r.ContentModeration != nil {
+		for _, cat := range r.ContentModeration.Categories {
+			if cat.Name == nil || cat.Score == nil {
+				continue
+			}
+			out.Categories = append(out.Categories, GuardrailCategory{Name: *cat.Name, Score: *cat.Score})
+		}
+	}
+	for _, hit := range r.PersonallyIdentifiableInformation {
+		h := GuardrailPiiHit{}
+		if hit.Text != nil {
+			h.Text = *hit.Text
+		}
+		if hit.Label != nil {
+			h.Label = *hit.Label
+		}
+		if hit.Score != nil {
+			h.Score = *hit.Score
+		}
+		if hit.Offset != nil {
+			h.Offset = *hit.Offset
+		}
+		if hit.Length != nil {
+			h.Length = *hit.Length
+		}
+		out.Pii = append(out.Pii, h)
+	}
+	if r.PromptInjection != nil && r.PromptInjection.Score != nil {
+		out.PromptInjectionScore = r.PromptInjection.Score
+	}
+	return out
+}
@@ -0,0 +1,73 @@
+package oci
+
+import (
+	"bytes"
+	"context"
+	"fmt"
+	"io"
+	"net/http"
+
+	"github.com/oracle/oci-go-sdk/v65/common"
+)
+
+// compatResponsesLimit 限制直通响应体大小;web_search 输出含多段引用,给足余量。
+const compatResponsesLimit = int64(8 << 20)
+
+// GenAiCompatResponses 实现 Client:把 OpenAI Responses 请求体直通到 OCI
+// `/20231130/actions/v1/responses`(IAM 签名)。xAI 服务端工具(web_search /
+// x_search / code_interpreter)与 mcp 已被 Oracle 文档正式支持,工具参数与限制
+// 遵循 xAI 规格;调用方须自行校验并改写请求体(store/stream)。
+func (c *RealClient) GenAiCompatResponses(ctx context.Context, cred Credentials, region string, body []byte) ([]byte, error) {
+	ic, err := c.genAiInferenceClient(cred, region)
+	if err != nil {
+		return nil, err
+	}
+	client := ic.BaseClient
+	common.UpdateEndpointTemplateForOptions(&client)
+	common.SetMissingTemplateParams(&client)
+	request, err := http.NewRequestWithContext(ctx, http.MethodPost, "/actions/v1/responses", bytes.NewReader(body))
+	if err != nil {
+		return nil, fmt.Errorf("build compat responses request: %w", err)
+	}
+	request.Header.Set("Content-Type", "application/json")
+	request.Header.Set("CompartmentId", cred.TenancyOCID)
+	request.Header.Set("opc-compartment-id", cred.TenancyOCID)
+	response, err := client.Call(ctx, request)
+	if err != nil {
+		return nil, err
+	}
+	defer response.Body.Close()
+	payload, err := io.ReadAll(io.LimitReader(response.Body, compatResponsesLimit))
+	if err != nil {
+		return nil, fmt.Errorf("read compat responses body: %w", err)
+	}
+	return payload, nil
+}
+
+// GenAiCompatResponsesStream 实现 Client:以流式直通 OCI `/actions/v1/responses`,
+// 建立成功(2xx)返回 SSE body(调用方负责 Close);建立失败返回 SDK ServiceError,
+// 与既有渠道切换/熔断错误分类兼容。请求体须由调用方置 stream:true。
+func (c *RealClient) GenAiCompatResponsesStream(ctx context.Context, cred Credentials, region string, body []byte) (io.ReadCloser, error) {
+	ic, err := c.genAiInferenceClient(cred, region)
+	if err != nil {
+		return nil, err
+	}
+	client := ic.BaseClient
+	common.UpdateEndpointTemplateForOptions(&client)
+	common.SetMissingTemplateParams(&client)
+	request, err := http.NewRequestWithContext(ctx, http.MethodPost, "/actions/v1/responses", bytes.NewReader(body))
+	if err != nil {
+		return nil, fmt.Errorf("build compat responses stream request: %w", err)
+	}
+	request.Header.Set("Content-Type", "application/json")
+	request.Header.Set("CompartmentId", cred.TenancyOCID)
+	request.Header.Set("opc-compartment-id", cred.TenancyOCID)
+	response, err := client.Call(ctx, request)
+	if err != nil {
+		if response != nil && response.Body != nil {
+			response.Body.Close()
+		}
+		return nil, err
+	}
+	return response.Body, nil
+}
@@ -38,3 +38,33 @@ func TestToGenAiModelRetiredFilter(t *testing.T) {
 		t.Error("正常模型应入池")
 	}
 }
+
+func TestDedupGenAiModelsFineTunePreference(t *testing.T) {
+	now := time.Date(2026, 7, 10, 0, 0, 0, 0, time.UTC)
+	chat := []generativeai.ModelCapabilityEnum{generativeai.ModelCapabilityChat}
+	chatFT := []generativeai.ModelCapabilityEnum{generativeai.ModelCapabilityChat, generativeai.ModelCapabilityFineTune}
+	mk := func(id, name string, caps []generativeai.ModelCapabilityEnum) generativeai.ModelSummary {
+		return generativeai.ModelSummary{
+			Id: common.String(id), DisplayName: common.String(name), Vendor: common.String("meta"),
+			Capabilities: caps, LifecycleState: generativeai.ModelLifecycleStateActive,
+		}
+	}
+	const name = "meta.llama-3-70b-instruct"
+	tests := []struct {
+		name  string
+		items []generativeai.ModelSummary
+		want  string // 期望保留的 OCID
+	}{
+		{"基座在前、纯对话在后:保留纯对话条目", []generativeai.ModelSummary{mk("o-ft", name, chatFT), mk("o-od", name, chat)}, "o-od"},
+		{"纯对话在前、基座在后:保留纯对话条目", []generativeai.ModelSummary{mk("o-od", name, chat), mk("o-ft", name, chatFT)}, "o-od"},
+		{"仅基座条目:保留不误删", []generativeai.ModelSummary{mk("o-ft", name, chatFT)}, "o-ft"},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			got := dedupGenAiModels(tt.items, now)
+			if len(got) != 1 || got[0].Ocid != tt.want {
+				t.Errorf("dedupGenAiModels() = %+v, want 仅保留 %s", got, tt.want)
+			}
+		})
+	}
+}
@@ -0,0 +1,44 @@
+package oci
+
+import (
+	"bytes"
+	"context"
+	"fmt"
+	"io"
+	"net/http"
+
+	"github.com/oracle/oci-go-sdk/v65/common"
+)
+
+// compatSpeechLimit 限制 TTS 音频响应体大小(长文本 mp3 给足余量)。
+const compatSpeechLimit = int64(64 << 20)
+
+// GenAiCompatSpeech 实现 Client:把 OpenAI Audio Speech 请求体直通到 OCI
+// 兼容面 `/openai/v1/audio/speech`(IAM 签名,BasePath 置空——该面与
+// /20231130/actions 面并存,实测仅前者承载 TTS)。返回音频字节与 Content-Type。
+func (c *RealClient) GenAiCompatSpeech(ctx context.Context, cred Credentials, region string, body []byte) ([]byte, string, error) {
+	ic, err := c.genAiInferenceClient(cred, region)
+	if err != nil {
+		return nil, "", err
+	}
+	client := ic.BaseClient
+	common.UpdateEndpointTemplateForOptions(&client)
+	common.SetMissingTemplateParams(&client)
+	client.BasePath = ""
+	request, err := http.NewRequestWithContext(ctx, http.MethodPost, "/openai/v1/audio/speech", bytes.NewReader(body))
+	if err != nil {
+		return nil, "", fmt.Errorf("build compat speech request: %w", err)
+	}
+	request.Header.Set("Content-Type", "application/json")
+	request.Header.Set("opc-compartment-id", cred.TenancyOCID)
+	response, err := client.Call(ctx, request)
+	if err != nil {
+		return nil, "", err
+	}
+	defer response.Body.Close()
+	payload, err := io.ReadAll(io.LimitReader(response.Body, compatSpeechLimit))
+	if err != nil {
+		return nil, "", fmt.Errorf("read compat speech body: %w", err)
+	}
+	return payload, response.Header.Get("Content-Type"), nil
+}
@@ -1,193 +1,27 @@
 package oci

 import (
-	"encoding/json"
 	"testing"

 	"github.com/oracle/oci-go-sdk/v65/generativeaiinference"
-
-	"oci-portal/internal/aiwire"
 )

-// TestIrToSDK 断言 IR→GENERIC 的角色拆装、采样参数与工具映射。
-func TestIrToSDK(t *testing.T) {
-	temp, mt := 0.7, 100
-	ir := aiwire.ChatRequest{
-		Model: "meta.llama-3.3-70b-instruct",
-		Messages: []aiwire.ChatMessage{
-			{Role: "system", Content: aiwire.NewTextContent("你是助手")},
-			{Role: "user", Content: aiwire.NewTextContent("你好")},
-			{Role: "assistant", ToolCalls: []aiwire.ToolCall{{ID: "c1", Type: "function", Function: aiwire.FunctionCall{Name: "get_weather", Arguments: `{"city":"东京"}`}}}},
-			{Role: "tool", ToolCallID: "c1", Content: aiwire.NewTextContent("晴")},
-		},
-		Temperature: &temp,
-		MaxTokens:   &mt,
-		Stop:        aiwire.StringList{"END"},
-		Tools:       []aiwire.Tool{{Type: "function", Function: aiwire.FunctionDef{Name: "get_weather", Parameters: json.RawMessage(`{"type":"object"}`)}}},
-		ToolChoice:  json.RawMessage(`"auto"`),
-	}
-	req := irToSDK(ir, true)
-	if len(req.Messages) != 4 {
-		t.Fatalf("messages = %d, want 4", len(req.Messages))
-	}
-	if _, ok := req.Messages[0].(generativeaiinference.SystemMessage); !ok {
-		t.Fatalf("messages[0] = %T, want SystemMessage", req.Messages[0])
-	}
-	am, ok := req.Messages[2].(generativeaiinference.AssistantMessage)
-	if !ok || len(am.ToolCalls) != 1 {
-		t.Fatalf("assistant tool calls 未映射: %T %+v", req.Messages[2], am)
-	}
-	tm, ok := req.Messages[3].(generativeaiinference.ToolMessage)
-	if !ok || deref(tm.ToolCallId) != "c1" {
-		t.Fatalf("tool message 未映射 toolCallId: %+v", tm)
-	}
-	if req.Temperature == nil || *req.Temperature != 0.7 {
-		t.Fatalf("temperature 未直通")
-	}
-	if req.MaxTokens == nil || *req.MaxTokens != 100 || len(req.Stop) != 1 {
-		t.Fatalf("maxTokens/stop 未直通")
-	}
-	if _, ok := req.ToolChoice.(generativeaiinference.ToolChoiceAuto); !ok {
-		t.Fatalf("toolChoice = %T, want auto", req.ToolChoice)
-	}
-	if len(req.Tools) != 1 || req.IsStream == nil || !*req.IsStream || req.StreamOptions == nil {
-		t.Fatalf("tools/stream 选项未映射")
-	}
-}
-
-// TestSdkToIR 断言响应文本、工具调用与用量的反向映射。
-func TestSdkToIR(t *testing.T) {
-	text, fr := "东京晴", "tool_calls"
-	idx, pt, ct, tt := 0, 10, 5, 15
-	id, name, args := "c1", "get_weather", `{"city":"东京"}`
-	resp := generativeaiinference.GenericChatResponse{
-		Choices: []generativeaiinference.ChatChoice{{
-			Index: &idx,
-			Message: generativeaiinference.AssistantMessage{
-				Content:   []generativeaiinference.ChatContent{generativeaiinference.TextContent{Text: &text}},
-				ToolCalls: []generativeaiinference.ToolCall{generativeaiinference.FunctionCall{Id: &id, Name: &name, Arguments: &args}},
-			},
-			FinishReason: &fr,
-		}},
-		Usage: &generativeaiinference.Usage{PromptTokens: &pt, CompletionTokens: &ct, TotalTokens: &tt},
-	}
-	out := sdkToIR(resp, "m1")
-	if len(out.Choices) != 1 || out.Choices[0].Message.Content.JoinText() != "东京晴" {
-		t.Fatalf("文本未映射: %+v", out)
-	}
-	if out.Choices[0].FinishReason != "tool_calls" || len(out.Choices[0].Message.ToolCalls) != 1 {
-		t.Fatalf("finishReason/toolCalls 未映射: %+v", out.Choices[0])
-	}
-	if out.Usage == nil || out.Usage.TotalTokens != 15 {
-		t.Fatalf("usage 未映射: %+v", out.Usage)
-	}
-}
-
-// TestParseGenAiEvent 断言流事件宽容解析:文本增量、结束原因与 usage 事件。
-func TestParseGenAiEvent(t *testing.T) {
-	chunk, ok := parseGenAiEvent([]byte(`{"index":0,"message":{"role":"ASSISTANT","content":[{"type":"TEXT","text":"你"}]}}`), "m1")
-	if !ok || len(chunk.Choices) != 1 || chunk.Choices[0].Delta.Content != "你" {
-		t.Fatalf("文本增量解析失败: %+v", chunk)
-	}
-	chunk, ok = parseGenAiEvent([]byte(`{"finishReason":"stop","usage":{"promptTokens":3,"completionTokens":2,"totalTokens":5}}`), "m1")
-	if !ok || chunk.Choices[0].FinishReason == nil || *chunk.Choices[0].FinishReason != "stop" {
-		t.Fatalf("finishReason 解析失败: %+v", chunk)
-	}
-	if chunk.Usage == nil || chunk.Usage.TotalTokens != 5 {
-		t.Fatalf("usage 解析失败: %+v", chunk.Usage)
-	}
-	if _, ok := parseGenAiEvent([]byte(`{}`), "m1"); ok {
-		t.Fatal("空事件应返回 ok=false")
-	}
-}
-
+// TestSdkUsageCachedTokens 断言 embed 用量映射:缓存命中挂 details,仅命中 >0 时出现。
 func TestSdkUsageCachedTokens(t *testing.T) {
 	p, c, tot, cached := 10, 5, 15, 8
 	u := sdkUsageToIR(&generativeaiinference.Usage{PromptTokens: &p, CompletionTokens: &c, TotalTokens: &tot,
 		PromptTokensDetails: &generativeaiinference.PromptTokensDetails{CachedTokens: &cached}})
-	if u.CachedTokens() != 8 {
-		t.Errorf("CachedTokens = %d, want 8", u.CachedTokens())
+	if u.PromptTokens != 10 || u.CompletionTokens != 5 || u.TotalTokens != 15 {
+		t.Fatalf("usage 映射错误: %+v", u)
 	}
-	if sdkUsageToIR(&generativeaiinference.Usage{PromptTokens: &p}).CachedTokens() != 0 {
-		t.Error("无细分时 CachedTokens 应为 0")
+	if u.PromptTokensDetails == nil || u.PromptTokensDetails.CachedTokens != 8 {
+		t.Fatalf("cachedTokens 未映射: %+v", u.PromptTokensDetails)
 	}
+	zero := 0
+	if got := sdkUsageToIR(&generativeaiinference.Usage{PromptTokens: &p, PromptTokensDetails: &generativeaiinference.PromptTokensDetails{CachedTokens: &zero}}); got.PromptTokensDetails != nil {
+		t.Fatalf("零命中不应带 details: %+v", got.PromptTokensDetails)
 	}
-
-func TestIrToCohereSDK(t *testing.T) {
-	ir := aiwire.ChatRequest{
-		Model: "cohere.command-r-plus",
-		Messages: []aiwire.ChatMessage{
-			{Role: "system", Content: aiwire.NewTextContent("你是助手")},
-			{Role: "user", Content: aiwire.NewTextContent("第一问")},
-			{Role: "assistant", Content: aiwire.NewTextContent("第一答")},
-			{Role: "user", Content: aiwire.NewTextContent("第二问")},
-		},
-	}
-	req, err := irToCohereSDK(ir, true)
-	if err != nil {
-		t.Fatalf("irToCohereSDK: %v", err)
-	}
-	if *req.Message != "第二问" || *req.PreambleOverride != "你是助手" || len(req.ChatHistory) != 2 {
-		t.Errorf("拆装错误: msg=%q preamble=%v history=%d", *req.Message, req.PreambleOverride, len(req.ChatHistory))
-	}
-	if _, ok := req.ChatHistory[0].(generativeaiinference.CohereUserMessage); !ok {
-		t.Errorf("history[0] 应为 user: %T", req.ChatHistory[0])
-	}
-	if _, ok := req.ChatHistory[1].(generativeaiinference.CohereChatBotMessage); !ok {
-		t.Errorf("history[1] 应为 chatbot: %T", req.ChatHistory[1])
-	}
-	if req.IsStream == nil || !*req.IsStream {
-		t.Error("IsStream 未设置")
-	}
-	if req.StreamOptions == nil || req.StreamOptions.IsIncludeUsage == nil || !*req.StreamOptions.IsIncludeUsage {
-		t.Error("流式应默认开启 usage 回传(StreamOptions.IsIncludeUsage)")
-	}
-	// 工具定义降级为扁平参数表(仅取 JSON Schema 顶层 properties)
-	ir.Tools = []aiwire.Tool{{Type: "function", Function: aiwire.FunctionDef{
-		Name: "get_weather", Parameters: json.RawMessage(`{"type":"object","properties":{"city":{"type":"string","description":"城市"}},"required":["city"]}`)}}}
-	req2, err := irToCohereSDK(ir, false)
-	if err != nil {
-		t.Fatalf("工具请求应支持: %v", err)
-	}
-	tool := req2.Tools[0]
-	if *tool.Name != "get_weather" || *tool.Description != "get_weather" {
-		t.Errorf("tool = %+v", tool)
-	}
-	if pd, ok := tool.ParameterDefinitions["city"]; !ok || *pd.Type != "string" || pd.IsRequired == nil || !*pd.IsRequired {
-		t.Errorf("param city = %+v", pd)
-	}
-	// 图片输入仍拒绝
-	ir.Tools = nil
-	ir.Messages = append(ir.Messages, aiwire.ChatMessage{Role: "user", Content: aiwire.NewPartsContent([]aiwire.ContentPart{
-		{Type: "image_url", ImageURL: &aiwire.ImageURL{URL: "data:image/png;base64,xx"}}})})
-	if _, err := irToCohereSDK(ir, false); err == nil {
-		t.Error("图片输入应被拒绝")
-	}
-	// 无 user 消息被拒
-	if _, err := irToCohereSDK(aiwire.ChatRequest{Model: "cohere.x", Messages: []aiwire.ChatMessage{{Role: "system", Content: aiwire.NewTextContent("s")}}}, false); err == nil {
-		t.Error("无 user 消息应被拒绝")
-	}
-}
-
-func TestCohereSDKToIR(t *testing.T) {
-	text := "回答"
-	resp := generativeaiinference.CohereChatResponse{
-		Text:         &text,
-		FinishReason: generativeaiinference.CohereChatResponseFinishReasonComplete,
-	}
-	out := cohereSDKToIR(resp, "cohere.command-r-plus")
-	if out.Choices[0].Message.Content.JoinText() != "回答" || out.Choices[0].FinishReason != "stop" {
-		t.Errorf("cohereSDKToIR = %+v", out.Choices[0])
-	}
-}
-
-func TestParseGenAiEventCohere(t *testing.T) {
-	chunk, ok := parseGenAiEvent([]byte(`{"apiFormat":"COHERE","text":"你好"}`), "cohere.command-r")
-	if !ok || chunk.Choices[0].Delta.Content != "你好" {
-		t.Errorf("cohere text 事件解析 = %+v, %v", chunk, ok)
-	}
-	chunk, ok = parseGenAiEvent([]byte(`{"apiFormat":"COHERE","finishReason":"COMPLETE","usage":{"promptTokens":3,"completionTokens":5,"totalTokens":8}}`), "cohere.command-r")
-	if !ok || *chunk.Choices[0].FinishReason != "stop" || chunk.Usage.TotalTokens != 8 {
-		t.Errorf("cohere 终帧解析 = %+v, %v", chunk, ok)
+	if sdkUsageToIR(nil) != nil {
+		t.Fatal("nil usage 应返回 nil")
 	}
 }
@@ -0,0 +1,107 @@
+package oci
+
+import (
+	"testing"
+	"time"
+
+	"github.com/oracle/oci-go-sdk/v65/common"
+	"github.com/oracle/oci-go-sdk/v65/identity"
+	"github.com/oracle/oci-go-sdk/v65/identitydomains"
+)
+
+func domainSummary(id, name, url string, state identity.DomainLifecycleStateEnum) identity.DomainSummary {
+	return identity.DomainSummary{Id: &id, DisplayName: &name, Url: &url, LifecycleState: state}
+}
+
+func TestPickDomainURL(t *testing.T) {
+	domains := []identity.DomainSummary{
+		domainSummary("ocid1.domain.idcs", "OracleIdentityCloudService", "https://idcs-a.example.com", identity.DomainLifecycleStateActive),
+		domainSummary("ocid1.domain.default", "Default", "https://idcs-b.example.com", identity.DomainLifecycleStateActive),
+		domainSummary("ocid1.domain.off", "Off", "https://idcs-c.example.com", identity.DomainLifecycleStateInactive),
+	}
+	cases := []struct {
+		name     string
+		domains  []identity.DomainSummary
+		domainID string
+		want     string
+	}{
+		{"缺省优先 Default 域", domains, "", "https://idcs-b.example.com"},
+		{"按 OCID 精确匹配", domains, "ocid1.domain.idcs", "https://idcs-a.example.com"},
+		{"OCID 不存在返回空", domains, "ocid1.domain.miss", ""},
+		{"非 ACTIVE 域不可选", domains, "ocid1.domain.off", ""},
+		{"无 Default 时取第一个 ACTIVE", domains[:1], "", "https://idcs-a.example.com"},
+		{"空列表返回空", nil, "", ""},
+	}
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			if got := pickDomainURL(tc.domains, tc.domainID); got != tc.want {
+				t.Fatalf("pickDomainURL() = %q, want %q", got, tc.want)
+			}
+		})
+	}
+}
+
+func TestToDomainListUser(t *testing.T) {
+	created := "2024-12-12T14:52:00Z"
+	lastLogin := "2026-07-01T08:00:00.123Z"
+	yes, no := true, false
+	u := identitydomains.User{
+		Id:       common.String("scim-1"),
+		Ocid:     common.String("ocid1.user.oc1..a"),
+		UserName: common.String("oci"),
+		Active:   &yes,
+		Meta:     &identitydomains.Meta{Created: &created},
+		Emails: []identitydomains.UserEmails{{
+			Value: common.String("a@b.c"), Primary: &yes, Verified: &yes,
+			Type: identitydomains.UserEmailsTypeWork,
+		}},
+		UrnIetfParamsScimSchemasOracleIdcsExtensionMfaUser: &identitydomains.ExtensionMfaUser{
+			MfaStatus: identitydomains.ExtensionMfaUserMfaStatusEnrolled,
+		},
+		UrnIetfParamsScimSchemasOracleIdcsExtensionUserStateUser: &identitydomains.ExtensionUserStateUser{
+			LastSuccessfulLoginDate: &lastLogin,
+		},
+	}
+	got := toDomainListUser(u, "ocid1.user.oc1..a")
+	if got.ID != "ocid1.user.oc1..a" || !got.IsCurrentUser {
+		t.Fatalf("ID/IsCurrentUser 映射错误: %+v", got)
+	}
+	if !got.MfaActivated || !got.EmailVerified || got.Email != "a@b.c" {
+		t.Fatalf("MFA/邮箱映射错误: %+v", got)
+	}
+	if got.LifecycleState != "ACTIVE" || got.TimeCreated == nil || got.LastLoginTime == nil {
+		t.Fatalf("状态/时间映射错误: %+v", got)
+	}
+
+	inactive := identitydomains.User{Id: common.String("scim-2"), Ocid: common.String("ocid1.user.oc1..b"), Active: &no}
+	if s := toDomainListUser(inactive, "x").LifecycleState; s != "INACTIVE" {
+		t.Fatalf("inactive 用户 LifecycleState = %q, want INACTIVE", s)
+	}
+}
+
+func TestParseScimTime(t *testing.T) {
+	valid := "2026-07-10T01:02:03Z"
+	bad := "not-a-time"
+	empty := ""
+	cases := []struct {
+		name string
+		in   *string
+		want *time.Time
+	}{
+		{"合法 RFC3339", &valid, func() *time.Time { t0, _ := time.Parse(time.RFC3339, valid); return &t0 }()},
+		{"nil 返回 nil", nil, nil},
+		{"空串返回 nil", &empty, nil},
+		{"非法格式返回 nil", &bad, nil},
+	}
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			got := parseScimTime(tc.in)
+			if (got == nil) != (tc.want == nil) {
+				t.Fatalf("parseScimTime() = %v, want %v", got, tc.want)
+			}
+			if got != nil && !got.Equal(*tc.want) {
+				t.Fatalf("parseScimTime() = %v, want %v", got, tc.want)
+			}
+		})
+	}
+}
@@ -25,7 +25,7 @@ func TestRealOCIFederation(t *testing.T) {
 	client := NewClient()
 	region := cred.Region

-	metadata, err := client.DownloadDomainSamlMetadata(ctx, cred, region)
+	metadata, err := client.DownloadDomainSamlMetadata(ctx, cred, region, "")
 	if err != nil {
 		t.Fatalf("DownloadDomainSamlMetadata: %v", err)
 	}
@@ -34,7 +34,7 @@ func TestRealOCIFederation(t *testing.T) {
 	}
 	t.Logf("domain saml metadata: %d bytes", len(metadata))

-	before, err := client.ListConsoleSignOnRules(ctx, cred, region)
+	before, err := client.ListConsoleSignOnRules(ctx, cred, region, "")
 	if err != nil {
 		t.Fatalf("ListConsoleSignOnRules: %v", err)
 	}
@@ -43,7 +43,7 @@ func TestRealOCIFederation(t *testing.T) {
 	idp := createTestIdp(ctx, t, client, cred, region)
 	verifyIdpDefaults(ctx, t, client, cred, region, idp.ID)

-	activated, err := client.SetIdentityProviderEnabled(ctx, cred, region, idp.ID, true)
+	activated, err := client.SetIdentityProviderEnabled(ctx, cred, region, "", idp.ID, true)
 	if err != nil {
 		t.Fatalf("activate idp: %v", err)
 	}
@@ -54,7 +54,7 @@ func TestRealOCIFederation(t *testing.T) {

 	exemptionRoundTrip(ctx, t, client, cred, region, idp.ID, len(before))

-	deactivated, err := client.SetIdentityProviderEnabled(ctx, cred, region, idp.ID, false)
+	deactivated, err := client.SetIdentityProviderEnabled(ctx, cred, region, "", idp.ID, false)
 	if err != nil {
 		t.Fatalf("deactivate idp: %v", err)
 	}
@@ -71,7 +71,7 @@ func createTestIdp(ctx context.Context, t *testing.T, client *RealClient, cred C
 		t.Fatalf("read test-idp.xml: %v", err)
 	}
 	name := fmt.Sprintf("oci-portal-e2e-idp-%d", time.Now().Unix())
-	idp, err := client.CreateSamlIdentityProvider(ctx, cred, region, CreateIdpInput{
+	idp, err := client.CreateSamlIdentityProvider(ctx, cred, region, "", CreateIdpInput{
 		Name:        name,
 		Metadata:    string(metadata),
 		Description: "oci-portal e2e temporary idp",
@@ -86,7 +86,7 @@ func createTestIdp(ctx context.Context, t *testing.T, client *RealClient, cred C
 	t.Cleanup(func() {
 		cctx, ccancel := context.WithTimeout(context.Background(), 2*time.Minute)
 		defer ccancel()
-		if err := client.DeleteIdentityProvider(cctx, cred, region, idp.ID); err != nil {
+		if err := client.DeleteIdentityProvider(cctx, cred, region, "", idp.ID); err != nil {
 			t.Errorf("cleanup: delete idp: %v", err)
 		} else {
 			t.Log("cleanup: idp deleted")
@@ -98,7 +98,7 @@ func createTestIdp(ctx context.Context, t *testing.T, client *RealClient, cred C
 // verifyIdpDefaults 回读 IdP 断言控制台默认值与 JIT 属性映射两条。
 func verifyIdpDefaults(ctx context.Context, t *testing.T, client *RealClient, cred Credentials, region, idpID string) {
 	t.Helper()
-	dc, err := client.domainsClient(ctx, cred, region)
+	dc, err := client.domainsClient(ctx, cred, region, "")
 	if err != nil {
 		t.Fatalf("domainsClient: %v", err)
 	}
@@ -151,7 +151,7 @@ func verifyJitMappings(ctx context.Context, t *testing.T, dc identitydomains.Ide
 // assertLoginPage 校验 DefaultIDPRule 的 SamlIDPs 是否包含目标 IdP。
 func assertLoginPage(ctx context.Context, t *testing.T, client *RealClient, cred Credentials, region, idpID string, want bool) {
 	t.Helper()
-	dc, err := client.domainsClient(ctx, cred, region)
+	dc, err := client.domainsClient(ctx, cred, region, "")
 	if err != nil {
 		t.Fatalf("domainsClient: %v", err)
 	}
@@ -175,7 +175,7 @@ func assertLoginPage(ctx context.Context, t *testing.T, client *RealClient, cred
 // exemptionRoundTrip 创建免 MFA 规则断言置顶与顺延，删除后断言完全复位。
 func exemptionRoundTrip(ctx context.Context, t *testing.T, client *RealClient, cred Credentials, region, idpID string, beforeCount int) {
 	t.Helper()
-	rule, err := client.CreateMfaExemptionRule(ctx, cred, region, idpID, "oci-portal-e2e-skip-mfa")
+	rule, err := client.CreateMfaExemptionRule(ctx, cred, region, "", idpID, "oci-portal-e2e-skip-mfa")
 	if err != nil {
 		t.Fatalf("CreateMfaExemptionRule: %v", err)
 	}
@@ -185,12 +185,12 @@ func exemptionRoundTrip(ctx context.Context, t *testing.T, client *RealClient, c
 		if deleted {
 			return
 		}
-		if err := client.DeleteMfaExemptionRule(ctx, cred, region, rule.ID); err != nil {
+		if err := client.DeleteMfaExemptionRule(ctx, cred, region, "", rule.ID); err != nil {
 			t.Errorf("cleanup: delete exemption rule: %v", err)
 		}
 	}()

-	after, err := client.ListConsoleSignOnRules(ctx, cred, region)
+	after, err := client.ListConsoleSignOnRules(ctx, cred, region, "")
 	if err != nil {
 		t.Fatalf("ListConsoleSignOnRules after create: %v", err)
 	}
@@ -208,11 +208,11 @@ func exemptionRoundTrip(ctx context.Context, t *testing.T, client *RealClient, c
 		t.Errorf("condition value %q does not reference idp", after[0].ConditionValue)
 	}

-	if err := client.DeleteMfaExemptionRule(ctx, cred, region, rule.ID); err != nil {
+	if err := client.DeleteMfaExemptionRule(ctx, cred, region, "", rule.ID); err != nil {
 		t.Fatalf("DeleteMfaExemptionRule: %v", err)
 	}
 	deleted = true
-	restored, err := client.ListConsoleSignOnRules(ctx, cred, region)
+	restored, err := client.ListConsoleSignOnRules(ctx, cred, region, "")
 	if err != nil {
 		t.Fatalf("ListConsoleSignOnRules after delete: %v", err)
 	}
@@ -21,7 +21,7 @@ func TestRealOCITenantUsers(t *testing.T) {
 	client := NewClient()
 	homeRegion := cred.Region

-	users, err := client.ListTenantUsers(ctx, cred)
+	users, err := client.ListTenantUsers(ctx, cred, homeRegion, "")
 	if err != nil {
 		t.Fatalf("ListTenantUsers: %v", err)
 	}
@@ -40,7 +40,7 @@ func TestRealOCITenantUsers(t *testing.T) {
 	t.Logf("tenant users: %d", len(users))

 	name := fmt.Sprintf("oci-portal-e2e-user-%d", time.Now().Unix())
-	user, err := client.CreateTenantUser(ctx, cred, homeRegion, CreateTenantUserInput{
+	user, err := client.CreateTenantUser(ctx, cred, homeRegion, "", CreateTenantUserInput{
 		Name: name, Description: "e2e temp user", Email: name + "@example.com",
 		GivenName: "E2E", FamilyName: "Temp",
 	})
@@ -49,7 +49,7 @@ func TestRealOCITenantUsers(t *testing.T) {
 	}
 	t.Logf("created user %s (%s)", user.Name, user.ID)
 	defer func() {
-		if err := client.DeleteTenantUser(ctx, cred, homeRegion, user.ID); err != nil {
+		if err := client.DeleteTenantUser(ctx, cred, homeRegion, "", user.ID); err != nil {
 			t.Errorf("cleanup: DeleteTenantUser: %v", err)
 		} else {
 			t.Log("cleanup: user deleted")
@@ -57,7 +57,7 @@ func TestRealOCITenantUsers(t *testing.T) {
 	}()

 	newDesc, newEmail, newGiven := "e2e updated", name+"+upd@example.com", "E2E2"
-	updated, err := client.UpdateTenantUser(ctx, cred, homeRegion, user.ID, UpdateTenantUserInput{
+	updated, err := client.UpdateTenantUser(ctx, cred, homeRegion, "", user.ID, UpdateTenantUserInput{
 		Description: &newDesc, Email: &newEmail, GivenName: &newGiven,
 	})
 	if err != nil {
@@ -68,7 +68,7 @@ func TestRealOCITenantUsers(t *testing.T) {
 	}
 	t.Logf("updated user desc=%q email=%q", updated.Description, updated.Email)

-	password, err := client.ResetTenantUserPassword(ctx, cred, homeRegion, user.ID)
+	password, err := client.ResetTenantUserPassword(ctx, cred, homeRegion, "", user.ID)
 	if err != nil {
 		t.Fatalf("ResetTenantUserPassword: %v", err)
 	}
@@ -77,7 +77,7 @@ func TestRealOCITenantUsers(t *testing.T) {
 	}
 	t.Logf("password reset ok (len=%d)", len(password))

-	mfaDeleted, err := client.DeleteTenantUserMfaDevices(ctx, cred, homeRegion, user.ID)
+	mfaDeleted, err := client.DeleteTenantUserMfaDevices(ctx, cred, homeRegion, "", user.ID)
 	if err != nil {
 		t.Fatalf("DeleteTenantUserMfaDevices: %v", err)
 	}
@@ -101,12 +101,12 @@ func TestRealOCIDomainSettings(t *testing.T) {
 	client := NewClient()
 	region := cred.Region

-	original, err := client.GetNotificationRecipients(ctx, cred, region)
+	original, err := client.GetNotificationRecipients(ctx, cred, region, "")
 	if err != nil {
 		t.Fatalf("GetNotificationRecipients: %v", err)
 	}
 	t.Logf("original recipients=%v testMode=%v", original.Recipients, original.TestModeEnabled)
-	updated, err := client.UpdateNotificationRecipients(ctx, cred, region, []string{"e2e@example.com"})
+	updated, err := client.UpdateNotificationRecipients(ctx, cred, region, "", []string{"e2e@example.com"})
 	if err != nil {
 		t.Fatalf("UpdateNotificationRecipients: %v", err)
 	}
@@ -117,13 +117,13 @@ func TestRealOCIDomainSettings(t *testing.T) {
 	if !original.TestModeEnabled {
 		restoreRecipients = nil
 	}
-	if _, err := client.UpdateNotificationRecipients(ctx, cred, region, restoreRecipients); err != nil {
+	if _, err := client.UpdateNotificationRecipients(ctx, cred, region, "", restoreRecipients); err != nil {
 		t.Errorf("restore recipients: %v", err)
 	} else {
 		t.Log("recipients restored")
 	}

-	policies, err := client.ListPasswordPolicies(ctx, cred, region)
+	policies, err := client.ListPasswordPolicies(ctx, cred, region, "")
 	if err != nil {
 		t.Fatalf("ListPasswordPolicies: %v", err)
 	}
@@ -133,7 +133,7 @@ func TestRealOCIDomainSettings(t *testing.T) {
 	}
 	t.Logf("policies: %d, patch target %s strength=%s expires=%v", len(policies), target.Name, target.PasswordStrength, target.PasswordExpiresAfter)
 	days := 350
-	patched, err := client.UpdatePasswordPolicy(ctx, cred, region, target.ID, UpdatePasswordPolicyInput{PasswordExpiresAfter: &days})
+	patched, err := client.UpdatePasswordPolicy(ctx, cred, region, "", target.ID, UpdatePasswordPolicyInput{PasswordExpiresAfter: &days})
 	if err != nil {
 		t.Fatalf("UpdatePasswordPolicy: %v", err)
 	}
@@ -144,7 +144,7 @@ func TestRealOCIDomainSettings(t *testing.T) {
 	if target.PasswordExpiresAfter != nil {
 		restore = *target.PasswordExpiresAfter
 	}
-	if _, err := client.UpdatePasswordPolicy(ctx, cred, region, target.ID, UpdatePasswordPolicyInput{PasswordExpiresAfter: &restore}); err != nil {
+	if _, err := client.UpdatePasswordPolicy(ctx, cred, region, "", target.ID, UpdatePasswordPolicyInput{PasswordExpiresAfter: &restore}); err != nil {
 		t.Errorf("restore password policy: %v", err)
 	} else {
 		t.Logf("policy restored to %d", restore)
@@ -29,8 +29,8 @@ type SignOnRuleInfo struct {
 }

 // ListConsoleSignOnRules 实现 Client：按优先级列出 OCI Console sign-on 策略的规则。
-func (c *RealClient) ListConsoleSignOnRules(ctx context.Context, cred Credentials, region string) ([]SignOnRuleInfo, error) {
-	dc, err := c.domainsClient(ctx, cred, region)
+func (c *RealClient) ListConsoleSignOnRules(ctx context.Context, cred Credentials, region, domainID string) ([]SignOnRuleInfo, error) {
+	dc, err := c.domainsClient(ctx, cred, region, domainID)
 	if err != nil {
 		return nil, err
 	}
@@ -99,8 +99,8 @@ func fillRuleCondition(ctx context.Context, dc identitydomains.IdentityDomainsCl

 // CreateMfaExemptionRule 实现 Client：为指定 IdP 创建免 MFA sign-on 规则并置顶。
 // 规则语义与控制台一致：subject.authenticatedBy in [idpID] → authenticationFactor=IDP。
-func (c *RealClient) CreateMfaExemptionRule(ctx context.Context, cred Credentials, region, idpID, ruleName string) (SignOnRuleInfo, error) {
-	dc, err := c.domainsClient(ctx, cred, region)
+func (c *RealClient) CreateMfaExemptionRule(ctx context.Context, cred Credentials, region, domainID, idpID, ruleName string) (SignOnRuleInfo, error) {
+	dc, err := c.domainsClient(ctx, cred, region, domainID)
 	if err != nil {
 		return SignOnRuleInfo{}, err
 	}
@@ -239,11 +239,11 @@ func patchPolicyRules(ctx context.Context, dc identitydomains.IdentityDomainsCli

 // DeleteMfaExemptionRule 实现 Client：删除免 MFA 规则并连带清理其条件。
 // 拒绝删除 Oracle 预置规则。
-func (c *RealClient) DeleteMfaExemptionRule(ctx context.Context, cred Credentials, region, ruleID string) error {
+func (c *RealClient) DeleteMfaExemptionRule(ctx context.Context, cred Credentials, region, domainID, ruleID string) error {
 	if strings.HasPrefix(ruleID, "OciConsole") {
 		return fmt.Errorf("delete sign-on rule: %s is a built-in rule and cannot be deleted", ruleID)
 	}
-	dc, err := c.domainsClient(ctx, cred, region)
+	dc, err := c.domainsClient(ctx, cred, region, domainID)
 	if err != nil {
 		return err
 	}
@@ -6,6 +6,7 @@ import (
 	"errors"
 	"fmt"
 	"math/big"
+	"slices"
 	"strings"
 	"time"

@@ -28,11 +29,18 @@ type TenantUser struct {
 	LastLoginTime  *time.Time `json:"lastLoginTime,omitempty"`
 }

-// Oracle 预置的管理员角色与管理员组显示名。
-const (
-	domainAdminRoleName     = "Identity Domain Administrator"
-	administratorsGroupName = "Administrators"
-)
+// Oracle 预置的「身份域管理员」应用角色显示名。
+const domainAdminRoleName = "Identity Domain Administrator"
+
+// adminGroupNames 是租户管理员组候选名(按优先序):原生域为 Administrators,
+// IDCS 迁移域(idcs foundation,如 OracleIdentityCloudService)为 OCI_Administrators;
+// 迁移域另有 IDCS_Administrators 属身份域管理范畴,由上面的应用角色概念覆盖,不算在内。
+var adminGroupNames = []string{"Administrators", "OCI_Administrators"}
+
+// isAdminGroupName 判断组显示名是否为租户管理员组。
+func isAdminGroupName(name string) bool {
+	return slices.Contains(adminGroupNames, name)
+}

 // scimAppRolesAttr 是 SCIM 用户 appRoles 扩展属性的完整路径。
 const scimAppRolesAttr = "urn:ietf:params:scim:schemas:oracle:idcs:extension:user:User:appRoles"
@@ -48,8 +56,8 @@ type TenantUserDetail struct {
 }

 // GetTenantUserDetail 实现 Client：查用户域档案与管理员状态（编辑表单预填充用）。
-func (c *RealClient) GetTenantUserDetail(ctx context.Context, cred Credentials, homeRegion, userID string) (TenantUserDetail, error) {
-	dc, err := c.domainsClient(ctx, cred, homeRegion)
+func (c *RealClient) GetTenantUserDetail(ctx context.Context, cred Credentials, homeRegion, domainID, userID string) (TenantUserDetail, error) {
+	dc, err := c.domainsClient(ctx, cred, homeRegion, domainID)
 	if err != nil {
 		return TenantUserDetail{}, err
 	}
@@ -68,7 +76,7 @@ func toTenantUserDetail(u identitydomains.User) TenantUserDetail {
 		d.FamilyName = deref(u.Name.FamilyName)
 	}
 	for _, g := range u.Groups {
-		if deref(g.Display) == administratorsGroupName {
+		if isAdminGroupName(deref(g.Display)) {
 			d.InAdminGroup = true
 		}
 	}
@@ -94,8 +102,13 @@ func (c *RealClient) identityClientAt(cred Credentials, region string) (identity
 	return ic, nil
 }

-// ListTenantUsers 实现 Client：列出租户全部 IAM 用户。
-func (c *RealClient) ListTenantUsers(ctx context.Context, cred Credentials) ([]TenantUser, error) {
+// ListTenantUsers 实现 Client：列出租户 IAM 用户。
+// domainID 非空时经 Identity Domains SCIM 只列该域用户;
+// 为空时走经典 IAM(跨域拍平视图,兼容无域老租户)。
+func (c *RealClient) ListTenantUsers(ctx context.Context, cred Credentials, homeRegion, domainID string) ([]TenantUser, error) {
+	if domainID != "" {
+		return c.listDomainUsers(ctx, cred, homeRegion, domainID)
+	}
 	ic, err := c.identityClientAt(cred, "")
 	if err != nil {
 		return nil, err
@@ -117,6 +130,78 @@ func (c *RealClient) ListTenantUsers(ctx context.Context, cred Credentials) ([]T
 	}
 }

+// scimListAttrs 是按域列用户时请求的属性集:列表字段 + MFA 与最近登录扩展。
+const scimListAttrs = "id,ocid,userName,description,emails,active,meta.created," +
+	"urn:ietf:params:scim:schemas:oracle:idcs:extension:mfa:User:mfaStatus," +
+	"urn:ietf:params:scim:schemas:oracle:idcs:extension:userState:User:lastSuccessfulLoginDate"
+
+// maxScimUserPages 限制按域列用户的翻页数(每页 200,上限 5000 人)。
+const maxScimUserPages = 25
+
+// listDomainUsers 经 SCIM 按 startIndex 翻页列出域内全部用户。
+func (c *RealClient) listDomainUsers(ctx context.Context, cred Credentials, homeRegion, domainID string) ([]TenantUser, error) {
+	dc, err := c.domainsClient(ctx, cred, homeRegion, domainID)
+	if err != nil {
+		return nil, err
+	}
+	var users []TenantUser
+	attrs := scimListAttrs
+	count := 200
+	start := 1
+	for range maxScimUserPages {
+		req := identitydomains.ListUsersRequest{Attributes: &attrs, Count: &count, StartIndex: &start}
+		resp, err := dc.ListUsers(ctx, req)
+		if err != nil {
+			return nil, fmt.Errorf("list domain users: %w", err)
+		}
+		for _, u := range resp.Resources {
+			users = append(users, toDomainListUser(u, cred.UserOCID))
+		}
+		start += len(resp.Resources)
+		if resp.TotalResults == nil || start > *resp.TotalResults || len(resp.Resources) == 0 {
+			return orEmpty(users), nil
+		}
+	}
+	return orEmpty(users), nil
+}
+
+// toDomainListUser 把 SCIM 用户映射为列表 DTO;ID 用 IAM OCID
+// 保持与经典路径同一 id 语义(IsCurrentUser 比对、后续操作路由)。
+func toDomainListUser(u identitydomains.User, currentUserOCID string) TenantUser {
+	out := toDomainTenantUser(u)
+	out.IsCurrentUser = out.ID == currentUserOCID
+	if u.Active != nil && !*u.Active {
+		out.LifecycleState = "INACTIVE"
+	}
+	if u.Meta != nil {
+		out.TimeCreated = parseScimTime(u.Meta.Created)
+	}
+	if ext := u.UrnIetfParamsScimSchemasOracleIdcsExtensionMfaUser; ext != nil {
+		out.MfaActivated = ext.MfaStatus == identitydomains.ExtensionMfaUserMfaStatusEnrolled
+	}
+	if ext := u.UrnIetfParamsScimSchemasOracleIdcsExtensionUserStateUser; ext != nil {
+		out.LastLoginTime = parseScimTime(ext.LastSuccessfulLoginDate)
+	}
+	for _, e := range u.Emails {
+		if e.Primary != nil && *e.Primary {
+			out.EmailVerified = e.Verified != nil && *e.Verified
+		}
+	}
+	return out
+}
+
+// parseScimTime 解析 SCIM 的 RFC3339 时间串;空值或格式异常返回 nil。
+func parseScimTime(s *string) *time.Time {
+	if s == nil || *s == "" {
+		return nil
+	}
+	t, err := time.Parse(time.RFC3339, *s)
+	if err != nil {
+		return nil
+	}
+	return &t
+}
+
 // CreateTenantUserInput 是创建用户的输入；GivenName/FamilyName 走
 // Identity Domains SCIM（经典 IAM 无对应字段）。两个管理员选项同时勾选时
 // 先授身份域管理员角色，再加入 Administrators 组。
@@ -133,12 +218,13 @@ type CreateTenantUserInput struct {
 // CreateTenantUser 实现 Client：创建用户（须发往 home region）。
 // 优先 Identity Domains SCIM（支持姓名与管理员授权）；SCIM 不可用时回退
 // 经典 IAM API，此时管理员选项无法执行、姓名并入描述。
-func (c *RealClient) CreateTenantUser(ctx context.Context, cred Credentials, homeRegion string, in CreateTenantUserInput) (TenantUser, error) {
-	user, err := c.createUserViaDomain(ctx, cred, homeRegion, in)
+// domainID 非空表示用户显式选择了目标域,SCIM 失败不再回退(避免建错域)。
+func (c *RealClient) CreateTenantUser(ctx context.Context, cred Credentials, homeRegion, domainID string, in CreateTenantUserInput) (TenantUser, error) {
+	user, err := c.createUserViaDomain(ctx, cred, homeRegion, domainID, in)
 	if err == nil {
 		return user, nil
 	}
-	if in.GrantDomainAdmin || in.AddToAdminGroup {
+	if domainID != "" || in.GrantDomainAdmin || in.AddToAdminGroup {
 		return TenantUser{}, fmt.Errorf("create user via identity domain: %w", err)
 	}
 	return c.createUserClassic(ctx, cred, homeRegion, in)
@@ -170,8 +256,8 @@ func (c *RealClient) createUserClassic(ctx context.Context, cred Credentials, ho
 }

 // createUserViaDomain 通过 Identity Domains SCIM 创建用户并按选项授权。
-func (c *RealClient) createUserViaDomain(ctx context.Context, cred Credentials, homeRegion string, in CreateTenantUserInput) (TenantUser, error) {
-	dc, err := c.domainsClient(ctx, cred, homeRegion)
+func (c *RealClient) createUserViaDomain(ctx context.Context, cred Credentials, homeRegion, domainID string, in CreateTenantUserInput) (TenantUser, error) {
+	dc, err := c.domainsClient(ctx, cred, homeRegion, domainID)
 	if err != nil {
 		return TenantUser{}, err
 	}
@@ -373,9 +459,9 @@ type UpdateTenantUserInput struct {

 // UpdateTenantUser 实现 Client：编辑用户资料（须发往 home region）。
 // 优先 Identity Domains SCIM（支持姓名）；用户不在域中时回退经典
-// IAM API（仅备注与邮箱，姓名变更报错）。
-func (c *RealClient) UpdateTenantUser(ctx context.Context, cred Credentials, homeRegion, userID string, in UpdateTenantUserInput) (TenantUser, error) {
-	dc, err := c.domainsClient(ctx, cred, homeRegion)
+// IAM API（仅备注与邮箱，姓名变更报错）;domainID 非空时不回退。
+func (c *RealClient) UpdateTenantUser(ctx context.Context, cred Credentials, homeRegion, domainID, userID string, in UpdateTenantUserInput) (TenantUser, error) {
+	dc, err := c.domainsClient(ctx, cred, homeRegion, domainID)
 	if err != nil {
 		return TenantUser{}, err
 	}
@@ -394,6 +480,9 @@ func (c *RealClient) UpdateTenantUser(ctx context.Context, cred Credentials, hom
 		}
 		return user, nil
 	}
+	if domainID != "" {
+		return TenantUser{}, fmt.Errorf("update user %s: user not found in identity domain", userID)
+	}
 	if in.GivenName != nil || in.FamilyName != nil {
 		return TenantUser{}, fmt.Errorf("update user %s: 姓名仅身份域用户可修改（该用户不在身份域中）", userID)
 	}
@@ -495,7 +584,11 @@ func emailPatchOp(u identitydomains.User, email string) identitydomains.Operatio
 }

 // DeleteTenantUser 实现 Client：删除 IAM 用户（须发往 home region）。
-func (c *RealClient) DeleteTenantUser(ctx context.Context, cred Credentials, homeRegion, userID string) error {
+// domainID 非空时经该域 SCIM 删除(经典 API 对非默认域用户不生效)。
+func (c *RealClient) DeleteTenantUser(ctx context.Context, cred Credentials, homeRegion, domainID, userID string) error {
+	if domainID != "" {
+		return c.deleteDomainUser(ctx, cred, homeRegion, domainID, userID)
+	}
 	ic, err := c.identityClientAt(cred, homeRegion)
 	if err != nil {
 		return err
@@ -506,9 +599,31 @@ func (c *RealClient) DeleteTenantUser(ctx context.Context, cred Credentials, hom
 	return nil
 }

+// deleteDomainUser 经 SCIM 强制删除域用户(forceDelete 连带其授权与组成员关系)。
+func (c *RealClient) deleteDomainUser(ctx context.Context, cred Credentials, homeRegion, domainID, userID string) error {
+	dc, err := c.domainsClient(ctx, cred, homeRegion, domainID)
+	if err != nil {
+		return err
+	}
+	scimID, err := scimUserID(ctx, dc, userID)
+	if err != nil {
+		return err
+	}
+	force := true
+	_, err = dc.DeleteUser(ctx, identitydomains.DeleteUserRequest{UserId: &scimID, ForceDelete: &force})
+	if err != nil {
+		return fmt.Errorf("delete domain user %s: %w", userID, err)
+	}
+	return nil
+}
+
 // ResetTenantUserPassword 实现 Client：重置用户控制台密码并返回一次性新密码。
-// 优先经典 IAM API；新式租户（用户只存在于 Identity Domain）自动回退 SCIM。
-func (c *RealClient) ResetTenantUserPassword(ctx context.Context, cred Credentials, homeRegion, userID string) (string, error) {
+// domainID 非空直走该域 SCIM;为空优先经典 IAM API,
+// 新式租户（用户只存在于 Identity Domain）自动回退 SCIM。
+func (c *RealClient) ResetTenantUserPassword(ctx context.Context, cred Credentials, homeRegion, domainID, userID string) (string, error) {
+	if domainID != "" {
+		return c.resetPasswordViaDomain(ctx, cred, homeRegion, domainID, userID)
+	}
 	ic, err := c.identityClientAt(cred, homeRegion)
 	if err != nil {
 		return "", err
@@ -520,12 +635,12 @@ func (c *RealClient) ResetTenantUserPassword(ctx context.Context, cred Credentia
 	if !isClassicUnsupported(err) {
 		return "", fmt.Errorf("reset ui password of %s: %w", userID, err)
 	}
-	return c.resetPasswordViaDomain(ctx, cred, homeRegion, userID)
+	return c.resetPasswordViaDomain(ctx, cred, homeRegion, "", userID)
 }

 // resetPasswordViaDomain 通过 Identity Domains SCIM API 强设随机新密码。
-func (c *RealClient) resetPasswordViaDomain(ctx context.Context, cred Credentials, homeRegion, userID string) (string, error) {
-	dc, err := c.domainsClient(ctx, cred, homeRegion)
+func (c *RealClient) resetPasswordViaDomain(ctx context.Context, cred Credentials, homeRegion, domainID, userID string) (string, error) {
+	dc, err := c.domainsClient(ctx, cred, homeRegion, domainID)
 	if err != nil {
 		return "", err
 	}
@@ -558,7 +673,11 @@ func (c *RealClient) resetPasswordViaDomain(ctx context.Context, cred Credential

 // DeleteTenantUserMfaDevices 实现 Client：清除用户全部 MFA。
 // 先删经典 TOTP 设备，再通过 Identity Domains 移除全部认证因子（尽力）。
-func (c *RealClient) DeleteTenantUserMfaDevices(ctx context.Context, cred Credentials, homeRegion, userID string) (int, error) {
+func (c *RealClient) DeleteTenantUserMfaDevices(ctx context.Context, cred Credentials, homeRegion, domainID, userID string) (int, error) {
+	// 显式指定域时用户的认证因子只在域内,经典 TOTP 清理不适用
+	if domainID != "" {
+		return 0, c.removeDomainFactors(ctx, cred, homeRegion, domainID, userID)
+	}
 	ic, err := c.identityClientAt(cred, homeRegion)
 	if err != nil {
 		return 0, err
@@ -567,7 +686,7 @@ func (c *RealClient) DeleteTenantUserMfaDevices(ctx context.Context, cred Creden
 	if err != nil {
 		return deleted, err
 	}
-	if err := c.removeDomainFactors(ctx, cred, homeRegion, userID); err != nil {
+	if err := c.removeDomainFactors(ctx, cred, homeRegion, "", userID); err != nil {
 		return deleted, err
 	}
 	return deleted, nil
@@ -594,8 +713,8 @@ func deleteClassicTotpDevices(ctx context.Context, ic identity.IdentityClient, u

 // removeDomainFactors 通过 SCIM AuthenticationFactorsRemover 移除用户
 // 在 Identity Domain 注册的全部认证因子；用户不在域中时视为无事可做。
-func (c *RealClient) removeDomainFactors(ctx context.Context, cred Credentials, homeRegion, userID string) error {
-	dc, err := c.domainsClient(ctx, cred, homeRegion)
+func (c *RealClient) removeDomainFactors(ctx context.Context, cred Credentials, homeRegion, domainID, userID string) error {
+	dc, err := c.domainsClient(ctx, cred, homeRegion, domainID)
 	if err != nil {
 		return err
 	}
@@ -47,7 +47,7 @@ func TestToTenantUserDetail(t *testing.T) {
 				},
 				Groups: []identitydomains.UserGroups{
 					{Value: common.String("g1"), Display: common.String("Readers")},
-					{Value: common.String("g2"), Display: common.String(administratorsGroupName)},
+					{Value: common.String("g2"), Display: common.String(adminGroupNames[0])},
 				},
 				UrnIetfParamsScimSchemasOracleIdcsExtensionUserUser: adminRoles,
 			},
@@ -1,332 +0,0 @@
-package service
-
-import (
-	"encoding/json"
-	"fmt"
-	"strings"
-
-	"oci-portal/internal/aiwire"
-)
-
-// ErrAiUnsupportedBlock 表示请求含网关无法承接的内容块(仅支持文本与图片)。
-var ErrAiUnsupportedBlock = fmt.Errorf("暂不支持文本与图片以外的内容块")
-
-// AnthropicToIR 把 Anthropic Messages 请求转为 IR(OpenAI 线格式)。
-// tool_result 块拆为独立 tool 角色消息(一拆多),顶层 system 变首条 system 消息。
-func AnthropicToIR(req aiwire.MessagesRequest) (aiwire.ChatRequest, error) {
-	ir := aiwire.ChatRequest{
-		Model:       req.Model,
-		Temperature: req.Temperature,
-		TopP:        req.TopP,
-		TopK:        req.TopK,
-		Stop:        req.StopSequences,
-		Stream:      req.Stream,
-		Tools:       anthTools(req.Tools),
-		ToolChoice:  anthToolChoice(req.ToolChoice),
-	}
-	if req.MaxTokens > 0 {
-		mt := req.MaxTokens
-		ir.MaxTokens = &mt
-	}
-	if sys := req.SystemText(); sys != "" {
-		ir.Messages = append(ir.Messages, aiwire.ChatMessage{Role: "system", Content: aiwire.NewTextContent(sys)})
-	}
-	for _, m := range req.Messages {
-		msgs, err := anthMessageToIR(m)
-		if err != nil {
-			return ir, err
-		}
-		ir.Messages = append(ir.Messages, msgs...)
-	}
-	return ir, nil
-}
-
-// anthMessageToIR 拆解单条 Anthropic 消息;user 消息里的 tool_result 前置为独立 tool 消息,
-// image 块转为 IR image_url 部件。
-func anthMessageToIR(m aiwire.AnthMessage) ([]aiwire.ChatMessage, error) {
-	var out []aiwire.ChatMessage
-	var parts []aiwire.ContentPart
-	var toolCalls []aiwire.ToolCall
-	hasImage := false
-	for _, b := range m.Content.AllBlocks() {
-		switch b.Type {
-		case "text":
-			parts = append(parts, aiwire.ContentPart{Type: "text", Text: b.Text})
-		case "image":
-			p, err := anthImagePart(b.Source)
-			if err != nil {
-				return nil, err
-			}
-			parts, hasImage = append(parts, p), true
-		case "tool_use":
-			toolCalls = append(toolCalls, aiwire.ToolCall{ID: b.ID, Type: "function",
-				Function: aiwire.FunctionCall{Name: b.Name, Arguments: string(b.Input)}})
-		case "tool_result":
-			out = append(out, aiwire.ChatMessage{Role: "tool", ToolCallID: b.ToolUseID,
-				Content: aiwire.NewTextContent(b.ResultText())})
-		case "thinking", "redacted_thinking":
-			// 一期忽略 thinking 块
-		default:
-			return nil, ErrAiUnsupportedBlock
-		}
-	}
-	content := partsContent(parts, hasImage)
-	if hasImage || content.JoinText() != "" || len(toolCalls) > 0 {
-		out = append(out, aiwire.ChatMessage{Role: m.Role, Content: content, ToolCalls: toolCalls})
-	}
-	return out, nil
-}
-
-// anthImagePart 把 Anthropic image 块转为 IR image_url 部件(base64 → data URI)。
-func anthImagePart(source json.RawMessage) (aiwire.ContentPart, error) {
-	var src struct {
-		Type      string `json:"type"`
-		MediaType string `json:"media_type"`
-		Data      string `json:"data"`
-		URL       string `json:"url"`
-	}
-	if err := json.Unmarshal(source, &src); err != nil {
-		return aiwire.ContentPart{}, fmt.Errorf("image source 解析失败: %w", err)
-	}
-	switch src.Type {
-	case "base64":
-		if src.MediaType == "" || src.Data == "" {
-			return aiwire.ContentPart{}, fmt.Errorf("image source 缺少 media_type 或 data")
-		}
-		url := "data:" + src.MediaType + ";base64," + src.Data
-		return aiwire.ContentPart{Type: "image_url", ImageURL: &aiwire.ImageURL{URL: url}}, nil
-	case "url":
-		if src.URL == "" {
-			return aiwire.ContentPart{}, fmt.Errorf("image source 缺少 url")
-		}
-		return aiwire.ContentPart{Type: "image_url", ImageURL: &aiwire.ImageURL{URL: src.URL}}, nil
-	}
-	return aiwire.ContentPart{}, fmt.Errorf("不支持的 image source 类型 %q", src.Type)
-}
-
-// partsContent 无图片时退回字符串形态(与上游线格式习惯一致),含图片时保留块数组。
-func partsContent(parts []aiwire.ContentPart, hasImage bool) aiwire.Content {
-	if !hasImage {
-		var sb strings.Builder
-		for _, p := range parts {
-			sb.WriteString(p.Text)
-		}
-		return aiwire.NewTextContent(sb.String())
-	}
-	return aiwire.NewPartsContent(parts)
-}
-
-func anthTools(tools []aiwire.AnthTool) []aiwire.Tool {
-	if len(tools) == 0 {
-		return nil
-	}
-	out := make([]aiwire.Tool, 0, len(tools))
-	for _, t := range tools {
-		out = append(out, aiwire.Tool{Type: "function", Function: aiwire.FunctionDef{
-			Name: t.Name, Description: t.Description, Parameters: t.InputSchema}})
-	}
-	return out
-}
-
-// anthToolChoice 映射 {type:auto|any|tool,name} → OpenAI 形态。
-func anthToolChoice(raw json.RawMessage) json.RawMessage {
-	if len(raw) == 0 {
-		return nil
-	}
-	var tc struct {
-		Type string `json:"type"`
-		Name string `json:"name"`
-	}
-	if json.Unmarshal(raw, &tc) != nil {
-		return nil
-	}
-	switch tc.Type {
-	case "auto":
-		return json.RawMessage(`"auto"`)
-	case "any":
-		return json.RawMessage(`"required"`)
-	case "tool":
-		b, _ := json.Marshal(map[string]any{"type": "function", "function": map[string]string{"name": tc.Name}})
-		return b
-	case "none":
-		return json.RawMessage(`"none"`)
-	}
-	return nil
-}
-
-// IRRespToAnthropic 把 IR 非流式响应转为 Anthropic Messages 响应。
-func IRRespToAnthropic(resp *aiwire.ChatResponse, id string) aiwire.MessagesResponse {
-	out := aiwire.MessagesResponse{ID: id, Type: "message", Role: "assistant", Model: resp.Model, Content: []aiwire.AnthBlock{}}
-	if len(resp.Choices) > 0 {
-		choice := resp.Choices[0]
-		if text := choice.Message.Content.JoinText(); text != "" {
-			out.Content = append(out.Content, aiwire.AnthBlock{Type: "text", Text: text})
-		}
-		for _, tc := range choice.Message.ToolCalls {
-			out.Content = append(out.Content, aiwire.AnthBlock{Type: "tool_use", ID: tc.ID,
-				Name: tc.Function.Name, Input: argsToJSON(tc.Function.Arguments)})
-		}
-		out.StopReason = anthStopReason(choice.FinishReason)
-	}
-	if resp.Usage != nil {
-		out.Usage = aiwire.AnthUsage{InputTokens: resp.Usage.PromptTokens, OutputTokens: resp.Usage.CompletionTokens,
-			CacheReadInputTokens: resp.Usage.CachedTokens()}
-	}
-	return out
-}
-
-// argsToJSON 保证 tool_use.input 是合法 JSON 对象(模型可能产出非法片段)。
-func argsToJSON(args string) json.RawMessage {
-	trimmed := strings.TrimSpace(args)
-	if trimmed == "" {
-		return json.RawMessage(`{}`)
-	}
-	if json.Valid([]byte(trimmed)) {
-		return json.RawMessage(trimmed)
-	}
-	b, _ := json.Marshal(map[string]string{"_raw": args})
-	return b
-}
-
-func anthStopReason(finish string) string {
-	switch finish {
-	case "length":
-		return "max_tokens"
-	case "tool_calls":
-		return "tool_use"
-	default:
-		return "end_turn"
-	}
-}
-
-// ---- Anthropic 流式状态机 ----
-
-// AnthEvent 是一条待写出的 Anthropic SSE 事件。
-type AnthEvent struct {
-	Event string
-	Data  any
-}
-
-// AnthStream 把 IR chunk 流聚合为 Anthropic 事件序列:
-// message_start → content_block_start/delta/stop(text 与 tool_use 分块)→ message_delta → message_stop。
-type AnthStream struct {
-	id, model   string
-	started     bool
-	blockOpen   bool
-	blockIsTool bool
-	toolID      string
-	blockIndex  int
-	stopReason  string
-	usage       aiwire.AnthUsage
-}
-
-// NewAnthStream 构造状态机;id 为响应消息 ID。
-func NewAnthStream(id, model string) *AnthStream {
-	return &AnthStream{id: id, model: model, blockIndex: -1, stopReason: "end_turn"}
-}
-
-// Feed 消费一个 IR chunk,返回应立即写出的事件。
-func (st *AnthStream) Feed(chunk aiwire.ChatChunk) []AnthEvent {
-	var events []AnthEvent
-	if !st.started {
-		st.started = true
-		events = append(events, st.startEvent())
-	}
-	if chunk.Usage != nil {
-		st.usage = aiwire.AnthUsage{InputTokens: chunk.Usage.PromptTokens, OutputTokens: chunk.Usage.CompletionTokens,
-			CacheReadInputTokens: chunk.Usage.CachedTokens()}
-	}
-	for _, choice := range chunk.Choices {
-		events = append(events, st.feedDelta(choice.Delta)...)
-		if choice.FinishReason != nil && *choice.FinishReason != "" {
-			st.stopReason = anthStopReason(*choice.FinishReason)
-		}
-	}
-	return events
-}
-
-func (st *AnthStream) startEvent() AnthEvent {
-	return AnthEvent{Event: "message_start", Data: map[string]any{
-		"type": "message_start",
-		"message": map[string]any{
-			"id": st.id, "type": "message", "role": "assistant", "model": st.model,
-			"content": []any{}, "stop_reason": nil,
-			"usage": map[string]int{"input_tokens": 0, "output_tokens": 0},
-		},
-	}}
-}
-
-// feedDelta 处理文本与工具调用增量,必要时切块。
-func (st *AnthStream) feedDelta(d aiwire.Delta) []AnthEvent {
-	var events []AnthEvent
-	if d.Content != "" {
-		if !st.blockOpen || st.blockIsTool {
-			events = append(events, st.openBlock(false, "", "")...)
-		}
-		events = append(events, AnthEvent{Event: "content_block_delta", Data: map[string]any{
-			"type": "content_block_delta", "index": st.blockIndex,
-			"delta": map[string]string{"type": "text_delta", "text": d.Content},
-		}})
-	}
-	for _, tc := range d.ToolCalls {
-		if tc.ID != "" && (!st.blockOpen || !st.blockIsTool || st.toolID != tc.ID) {
-			events = append(events, st.openBlock(true, tc.ID, tc.Function.Name)...)
-		}
-		if tc.Function.Arguments != "" && st.blockOpen && st.blockIsTool {
-			events = append(events, AnthEvent{Event: "content_block_delta", Data: map[string]any{
-				"type": "content_block_delta", "index": st.blockIndex,
-				"delta": map[string]string{"type": "input_json_delta", "partial_json": tc.Function.Arguments},
-			}})
-		}
-	}
-	return events
-}
-
-// openBlock 关闭当前块并打开新块(text 或 tool_use)。
-func (st *AnthStream) openBlock(isTool bool, toolID, toolName string) []AnthEvent {
-	var events []AnthEvent
-	if st.blockOpen {
-		events = append(events, st.closeBlockEvent())
-	}
-	st.blockOpen, st.blockIsTool, st.toolID = true, isTool, toolID
-	st.blockIndex++
-	block := map[string]any{"type": "text", "text": ""}
-	if isTool {
-		block = map[string]any{"type": "tool_use", "id": toolID, "name": toolName, "input": map[string]any{}}
-	}
-	events = append(events, AnthEvent{Event: "content_block_start", Data: map[string]any{
-		"type": "content_block_start", "index": st.blockIndex, "content_block": block,
-	}})
-	return events
-}
-
-func (st *AnthStream) closeBlockEvent() AnthEvent {
-	return AnthEvent{Event: "content_block_stop", Data: map[string]any{
-		"type": "content_block_stop", "index": st.blockIndex,
-	}}
-}
-
-// Finish 在上游流结束后收尾:关块 → message_delta(stop_reason+usage)→ message_stop。
-func (st *AnthStream) Finish() []AnthEvent {
-	var events []AnthEvent
-	if !st.started {
-		events = append(events, st.startEvent())
-		st.started = true
-	}
-	if st.blockOpen {
-		events = append(events, st.closeBlockEvent())
-		st.blockOpen = false
-	}
-	events = append(events,
-		AnthEvent{Event: "message_delta", Data: map[string]any{
-			"type":  "message_delta",
-			"delta": map[string]any{"stop_reason": st.stopReason, "stop_sequence": nil},
-			"usage": map[string]int{"output_tokens": st.usage.OutputTokens},
-		}},
-		AnthEvent{Event: "message_stop", Data: map[string]any{"type": "message_stop"}},
-	)
-	return events
-}
-
-// Usage 返回聚合到的用量(供调用日志)。
-func (st *AnthStream) Usage() aiwire.AnthUsage { return st.usage }
@@ -1,136 +0,0 @@
-package service
-
-import (
-	"encoding/json"
-	"strings"
-	"testing"
-
-	"oci-portal/internal/aiwire"
-)
-
-func mustAnthReq(t *testing.T, body string) aiwire.MessagesRequest {
-	t.Helper()
-	var req aiwire.MessagesRequest
-	if err := json.Unmarshal([]byte(body), &req); err != nil {
-		t.Fatalf("unmarshal anthropic request: %v", err)
-	}
-	return req
-}
-
-func TestAnthropicToIR(t *testing.T) {
-	req := mustAnthReq(t, `{
-		"model": "meta.llama-3.3-70b-instruct", "max_tokens": 128, "system": "你是助手",
-		"messages": [
-			{"role": "user", "content": "东京天气?"},
-			{"role": "assistant", "content": [
-				{"type": "text", "text": "查询中"},
-				{"type": "tool_use", "id": "t1", "name": "get_weather", "input": {"city": "东京"}}
-			]},
-			{"role": "user", "content": [
-				{"type": "tool_result", "tool_use_id": "t1", "content": "晴 25 度"},
-				{"type": "text", "text": "继续"}
-			]}
-		],
-		"tools": [{"name": "get_weather", "input_schema": {"type": "object"}}],
-		"tool_choice": {"type": "any"}
-	}`)
-	ir, err := AnthropicToIR(req)
-	if err != nil {
-		t.Fatalf("AnthropicToIR: %v", err)
-	}
-	roles := make([]string, 0, len(ir.Messages))
-	for _, m := range ir.Messages {
-		roles = append(roles, m.Role)
-	}
-	want := []string{"system", "user", "assistant", "tool", "user"}
-	if strings.Join(roles, ",") != strings.Join(want, ",") {
-		t.Fatalf("roles = %v, want %v", roles, want)
-	}
-	if ir.Messages[2].ToolCalls[0].Function.Name != "get_weather" {
-		t.Errorf("tool_use 未映射: %+v", ir.Messages[2])
-	}
-	if ir.Messages[3].ToolCallID != "t1" || ir.Messages[3].Content.JoinText() != "晴 25 度" {
-		t.Errorf("tool_result 未拆为 tool 消息: %+v", ir.Messages[3])
-	}
-	if ir.MaxTokens == nil || *ir.MaxTokens != 128 || len(ir.Tools) != 1 {
-		t.Errorf("max_tokens/tools 未直通")
-	}
-	if string(ir.ToolChoice) != `"required"` {
-		t.Errorf("tool_choice any → %s, want required", ir.ToolChoice)
-	}
-	// image 块拒绝
-	bad := mustAnthReq(t, `{"model":"m","max_tokens":1,"messages":[{"role":"user","content":[{"type":"image","source":{}}]}]}`)
-	if _, err := AnthropicToIR(bad); err == nil {
-		t.Error("image 块应被拒绝")
-	}
-}
-
-func TestIRRespToAnthropic(t *testing.T) {
-	resp := &aiwire.ChatResponse{
-		Model: "m1",
-		Choices: []aiwire.Choice{{
-			Message: aiwire.ChatMessage{
-				Role:    "assistant",
-				Content: aiwire.NewTextContent("查到了"),
-				ToolCalls: []aiwire.ToolCall{{ID: "t1", Type: "function",
-					Function: aiwire.FunctionCall{Name: "get_weather", Arguments: `{"city":"东京"}`}}},
-			},
-			FinishReason: "tool_calls",
-		}},
-		Usage: &aiwire.Usage{PromptTokens: 10, CompletionTokens: 5},
-	}
-	out := IRRespToAnthropic(resp, "msg_1")
-	if len(out.Content) != 2 || out.Content[0].Type != "text" || out.Content[1].Type != "tool_use" {
-		t.Fatalf("content 块 = %+v", out.Content)
-	}
-	if out.StopReason != "tool_use" || out.Usage.InputTokens != 10 || out.Usage.OutputTokens != 5 {
-		t.Errorf("stop_reason/usage 未映射: %+v", out)
-	}
-	if string(out.Content[1].Input) != `{"city":"东京"}` {
-		t.Errorf("tool input = %s", out.Content[1].Input)
-	}
-}
-
-// eventTypes 提取事件类型序列便于断言。
-func eventTypes(events []AnthEvent) string {
-	types := make([]string, 0, len(events))
-	for _, e := range events {
-		types = append(types, e.Event)
-	}
-	return strings.Join(types, ",")
-}
-
-func TestAnthStreamTextAndTool(t *testing.T) {
-	st := NewAnthStream("msg_1", "m1")
-	fr := "tool_calls"
-	var events []AnthEvent
-	events = append(events, st.Feed(aiwire.ChatChunk{Choices: []aiwire.ChunkChoice{{Delta: aiwire.Delta{Content: "你"}}}})...)
-	events = append(events, st.Feed(aiwire.ChatChunk{Choices: []aiwire.ChunkChoice{{Delta: aiwire.Delta{Content: "好"}}}})...)
-	events = append(events, st.Feed(aiwire.ChatChunk{Choices: []aiwire.ChunkChoice{{Delta: aiwire.Delta{
-		ToolCalls: []aiwire.ToolCallDelta{{ID: "t1", Function: aiwire.FunctionCallDelta{Name: "get_weather", Arguments: `{"ci`}}},
-	}}}})...)
-	events = append(events, st.Feed(aiwire.ChatChunk{
-		Choices: []aiwire.ChunkChoice{{Delta: aiwire.Delta{ToolCalls: []aiwire.ToolCallDelta{{ID: "t1", Function: aiwire.FunctionCallDelta{Arguments: `ty":"东京"}`}}}}, FinishReason: &fr}},
-		Usage:   &aiwire.Usage{PromptTokens: 8, CompletionTokens: 4},
-	})...)
-	events = append(events, st.Finish()...)
-
-	got := eventTypes(events)
-	want := "message_start,content_block_start,content_block_delta,content_block_delta," +
-		"content_block_stop,content_block_start,content_block_delta,content_block_delta," +
-		"content_block_stop,message_delta,message_stop"
-	if got != want {
-		t.Fatalf("事件序列:\n got %s\nwant %s", got, want)
-	}
-	if st.Usage().OutputTokens != 4 {
-		t.Errorf("usage 未聚合: %+v", st.Usage())
-	}
-}
-
-func TestAnthStreamEmpty(t *testing.T) {
-	st := NewAnthStream("msg_1", "m1")
-	events := st.Finish()
-	if got := eventTypes(events); got != "message_start,message_delta,message_stop" {
-		t.Errorf("空流事件序列 = %s", got)
-	}
-}
@@ -5,15 +5,18 @@ import (
 	"crypto/rand"
 	"crypto/sha256"
 	"encoding/hex"
+	"encoding/json"
 	"errors"
 	"fmt"
 	"log"
 	"sort"
 	"strings"
 	"sync"
+	"sync/atomic"
 	"time"

 	"gorm.io/gorm"
+	"gorm.io/gorm/clause"

 	"oci-portal/internal/aiwire"
 	"oci-portal/internal/model"
@@ -56,11 +59,39 @@ type AiGatewayService struct {
 	lastTouch map[uint]time.Time
 	// onChannelsChanged 在渠道增删后触发,由 main 装配为探测任务同步钩子
 	onChannelsChanged func(context.Context)
+	// filterDeprecated 是「过滤弃用模型」开关(内存镜像,持久化在 settings 表)
+	filterDeprecated atomic.Bool
 }

 // NewAiGatewayService 组装依赖;调用 StartCleanup 后开始调用日志周期清理。
 func NewAiGatewayService(db *gorm.DB, configs *OciConfigService, client oci.Client) *AiGatewayService {
-	return &AiGatewayService{db: db, configs: configs, client: client, lastTouch: map[uint]time.Time{}}
+	s := &AiGatewayService{db: db, configs: configs, client: client, lastTouch: map[uint]time.Time{}}
+	var row model.Setting
+	if err := db.Where("key = ?", settingAiFilterDeprecated).First(&row).Error; err == nil {
+		s.filterDeprecated.Store(row.Value == "1")
+	}
+	return s
+}
+
+// settingAiFilterDeprecated 是「过滤弃用模型」开关的配置键,值 "1"/"0",缺省关闭。
+const settingAiFilterDeprecated = "ai_filter_deprecated"
+
+// FilterDeprecated 返回「过滤弃用模型」开关状态。
+func (s *AiGatewayService) FilterDeprecated() bool { return s.filterDeprecated.Load() }
+
+// SetFilterDeprecated 持久化并即时生效开关:开启后已宣布弃用
+// (deprecated_at 非空,即使未退役)的模型从列表与路由中排除。
+func (s *AiGatewayService) SetFilterDeprecated(ctx context.Context, on bool) error {
+	value := "0"
+	if on {
+		value = "1"
+	}
+	err := s.db.WithContext(ctx).Save(&model.Setting{Key: settingAiFilterDeprecated, Value: value}).Error
+	if err != nil {
+		return fmt.Errorf("保存过滤弃用模型开关: %w", err)
+	}
+	s.filterDeprecated.Store(on)
+	return nil
 }

 // SetOnChannelsChanged 注册渠道数量变化钩子(渠道创建/删除成功后调用)。
@@ -78,7 +109,7 @@ func (s *AiGatewayService) fireChannelsChanged(ctx context.Context) {

 // CreateKey 生成网关密钥并返回明文(仅此一次);customValue 非空时使用给定值,
 // group 非空时该密钥只在同分组渠道内路由。
-func (s *AiGatewayService) CreateKey(ctx context.Context, name, customValue, group string) (string, *model.AiKey, error) {
+func (s *AiGatewayService) CreateKey(ctx context.Context, name, customValue, group string, models []string) (string, *model.AiKey, error) {
 	name = strings.TrimSpace(name)
 	if name == "" {
 		return "", nil, fmt.Errorf("密钥名称不能为空")
@@ -94,13 +125,28 @@ func (s *AiGatewayService) CreateKey(ctx context.Context, name, customValue, gro
 	if len(raw) < 8 {
 		return "", nil, fmt.Errorf("自定义密钥至少 8 个字符")
 	}
-	key := &model.AiKey{Name: name, KeyHash: hashKey(raw), Tail: raw[len(raw)-4:], Group: strings.TrimSpace(group), Enabled: true}
+	key := &model.AiKey{Name: name, KeyHash: hashKey(raw), Tail: raw[len(raw)-4:], Group: strings.TrimSpace(group), Models: normalizeKeyModels(models), Enabled: true}
 	if err := s.db.WithContext(ctx).Create(key).Error; err != nil {
 		return "", nil, fmt.Errorf("密钥名称或取值与现有密钥重复")
 	}
 	return raw, key, nil
 }

+// normalizeKeyModels 规范化模型白名单:trim、剔空串、保序去重;结果为空返回 nil(= 不限)。
+func normalizeKeyModels(models []string) []string {
+	var out []string
+	seen := map[string]bool{}
+	for _, m := range models {
+		m = strings.TrimSpace(m)
+		if m == "" || seen[m] {
+			continue
+		}
+		seen[m] = true
+		out = append(out, m)
+	}
+	return out
+}
+
 func hashKey(raw string) string {
 	sum := sha256.Sum256([]byte(raw))
 	return hex.EncodeToString(sum[:])
@@ -113,8 +159,8 @@ func (s *AiGatewayService) Keys(ctx context.Context) ([]model.AiKey, error) {
 	return keys, err
 }

-// UpdateKey 修改密钥名称 / 启用状态 / 分组(group 指针非空即覆盖,可置空)。
-func (s *AiGatewayService) UpdateKey(ctx context.Context, id uint, name string, enabled *bool, group *string) error {
+// UpdateKey 修改密钥名称 / 启用状态 / 分组 / 模型白名单(指针非空即覆盖,可置空)。
+func (s *AiGatewayService) UpdateKey(ctx context.Context, id uint, name string, enabled *bool, group *string, models *[]string) error {
 	updates := map[string]any{}
 	if name = strings.TrimSpace(name); name != "" {
 		updates["name"] = name
@@ -125,6 +171,14 @@ func (s *AiGatewayService) UpdateKey(ctx context.Context, id uint, name string,
 	if group != nil {
 		updates["key_group"] = strings.TrimSpace(*group)
 	}
+	if models != nil {
+		// 手动序列化走 map 更新,不依赖 GORM map 路径对 serializer 的支持
+		b, err := json.Marshal(normalizeKeyModels(*models))
+		if err != nil {
+			return fmt.Errorf("serialize models: %w", err)
+		}
+		updates["models"] = string(b)
+	}
 	if len(updates) == 0 {
 		return nil
 	}
@@ -289,12 +343,17 @@ func (s *AiGatewayService) ProbeChannel(ctx context.Context, id uint) (*model.Ai
 	return &fresh, nil
 }

-// probe 执行探测并返回 (状态, 错误摘要);同时完成模型缓存同步。
+// probe 执行探测并返回 (状态, 错误摘要);同时完成模型缓存同步(黑名单模型不入库)。
 func (s *AiGatewayService) probe(ctx context.Context, cred oci.Credentials, ch *model.AiChannel) (string, string) {
 	models, err := s.client.ListGenAiModels(ctx, cred, ch.Region)
 	if err != nil {
 		return classifyProbeErr(err), truncateErr(oci.CompactError(err))
 	}
+	models = supportedGatewayModels(models)
+	models, err = s.withoutBlacklisted(ctx, models)
+	if err != nil {
+		return "error", truncateErr(err.Error())
+	}
 	if len(models) == 0 {
 		_ = s.replaceModels(ctx, ch.ID, nil)
 		return "no_service", "区域无可用模型(GenAI 服务不可用或未开放)"
@@ -305,40 +364,88 @@ func (s *AiGatewayService) probe(ctx context.Context, cred oci.Credentials, ch *
 	return s.probeChat(ctx, cred, ch, models)
 }

-// probeChat 按偏好挑选至多 3 个模型依次试调:部分模型元数据标 CHAT 但实际
-// 不可对话(如 voice agent),遇 400/5xx 换下一个;401/403/404 属租户级直接定论。
+// probeChat 按候选顺序试调(上限 8):遇「模型不可按需调用」(微调基座 400 / 实体
+// 不存在 404)换下一个候选,错误信息带模型名供用户加入黑名单;401/403 与鉴权类 404
+// 属租户级直接定论 no_quota;其余错误(元数据标 CHAT 但实际不可对话等)累计 3 次止损。
 func (s *AiGatewayService) probeChat(ctx context.Context, cred oci.Credentials, ch *model.AiChannel, models []oci.GenAiModel) (string, string) {
 	status, detail := "error", "无可试调对话模型"
+	errBudget := 3
 	for _, m := range probeCandidates(models) {
 		code, err := s.client.GenAiProbeChat(ctx, cred, ch.Region, m.Ocid, m.Name)
 		switch {
 		case code == 200 || code == 429:
 			return "ok", ""
+		case oci.IsModelUnavailable(err):
+			status, detail = "error", truncateErr(fmt.Sprintf("%s: 不可按需调用,建议加入模型黑名单", m.Name))
 		case code == 401 || code == 403 || code == 404:
 			return "no_quota", truncateErr(oci.CompactError(err))
 		default:
 			status, detail = "error", truncateErr(fmt.Sprintf("%s: %s", m.Name, oci.CompactError(err)))
+			if errBudget--; errBudget == 0 {
+				return status, detail
+			}
 		}
 	}
 	return status, detail
 }

-// probeCandidates 只取对话模型并按可靠度排序取前 3:主流文本模型优先,
-// voice 等非常规形态殿后(embedding / rerank 已被能力筛选排除)。
+// probeCandidateCap 是单次探测的候选上限:放宽到 8 让一次探测有机会越过整批
+// 不可按需调用的坏模型找到可用者;其他错误另有 3 次止损预算。
+const probeCandidateCap = 8
+
+// probeCandidates 只取对话模型,按可靠度排序后跨厂商取候选:主流文本模型优先;
+// voice 等负分形态(元数据标 CHAT 但实际不可对话)直接排除,不浪费试调预算;
+// 每厂商先取最高分再按分数补位——部分区域某厂商全为微调基座(调用必失败),
+// 不能让单一厂商占满候选名额拖垮整个渠道的探测结论。
 func probeCandidates(models []oci.GenAiModel) []oci.GenAiModel {
 	var sorted []oci.GenAiModel
 	for _, m := range models {
-		if m.Capability == "" || m.Capability == "CHAT" {
+		if (m.Capability == "" || m.Capability == "CHAT") && probeScore(m.Name) >= 0 {
 			sorted = append(sorted, m)
 		}
 	}
 	sort.SliceStable(sorted, func(i, j int) bool {
 		return probeScore(sorted[i].Name) > probeScore(sorted[j].Name)
 	})
-	if len(sorted) > 3 {
-		sorted = sorted[:3]
+	return diversifyByVendor(sorted, probeCandidateCap)
 }
-	return sorted
+
+// diversifyByVendor 从已排序列表先每厂商各取一个,不足 limit 再按原序补位。
+func diversifyByVendor(sorted []oci.GenAiModel, limit int) []oci.GenAiModel {
+	picked := make([]oci.GenAiModel, 0, limit)
+	used := make(map[int]bool)
+	seenVendor := make(map[string]bool)
+	for i, m := range sorted {
+		if len(picked) >= limit {
+			break
+		}
+		if v := modelVendor(m); !seenVendor[v] {
+			seenVendor[v] = true
+			used[i] = true
+			picked = append(picked, m)
+		}
+	}
+	for i, m := range sorted {
+		if len(picked) >= limit {
+			break
+		}
+		if !used[i] {
+			picked = append(picked, m)
+		}
+	}
+	return picked
+}
+
+// modelVendor 取厂商标识;OCI 未回填 vendor 时退化为模型名「.」前缀。
+func modelVendor(m oci.GenAiModel) string {
+	if m.Vendor != "" {
+		return strings.ToLower(m.Vendor)
+	}
+	name := strings.ToLower(m.Name)
+	if i := strings.Index(name, "."); i > 0 {
+		return name[:i]
+	}
+	return name
 }

 func probeScore(name string) int {
@@ -376,7 +483,7 @@ func truncateErr(msg string) string {
 	return msg
 }

-// SyncModels 重新拉取渠道区域的模型列表并覆盖缓存。
+// SyncModels 重新拉取渠道区域的模型列表并覆盖缓存,黑名单中的模型不入库。
 func (s *AiGatewayService) SyncModels(ctx context.Context, id uint) ([]model.AiModelCache, error) {
 	var ch model.AiChannel
 	if err := s.db.WithContext(ctx).First(&ch, id).Error; err != nil {
@@ -390,6 +497,10 @@ func (s *AiGatewayService) SyncModels(ctx context.Context, id uint) ([]model.AiM
 	if err != nil {
 		return nil, fmt.Errorf("同步模型失败:%s", oci.CompactError(err))
 	}
+	models = supportedGatewayModels(models)
+	if models, err = s.withoutBlacklisted(ctx, models); err != nil {
+		return nil, err
+	}
 	if err := s.replaceModels(ctx, id, models); err != nil {
 		return nil, err
 	}
@@ -421,7 +532,7 @@ func (s *AiGatewayService) channelModels(ctx context.Context, channelID uint) ([
 	return rows, err
 }

-// GatewayModels 聚合启用渠道的模型(按名称去重),供 /ai/v1/models;
+// GatewayModels 聚合启用渠道的可用模型(按名称去重),供 /ai/v1/models;
 // group 非空时仅聚合该分组渠道(与密钥分组路由口径一致)。
 func (s *AiGatewayService) GatewayModels(ctx context.Context, group string) (aiwire.ModelList, error) {
 	q := s.db.WithContext(ctx).
@@ -429,6 +540,9 @@ func (s *AiGatewayService) GatewayModels(ctx context.Context, group string) (aiw
 	if group != "" {
 		q = q.Where("ai_channels.channel_group = ?", group)
 	}
+	if s.FilterDeprecated() {
+		q = q.Where("ai_model_caches.deprecated_at IS NULL")
+	}
 	var rows []model.AiModelCache
 	err := q.Order("ai_model_caches.name ASC").Find(&rows).Error
 	list := aiwire.ModelList{Object: "list", Data: []aiwire.Model{}}
@@ -501,18 +615,134 @@ func (s *AiGatewayService) ProbeAll(ctx context.Context) (string, error) {
 	return msg, nil
 }

+// ---- 模型黑名单 ----
+
+// Blacklist 列出全部黑名单模型(按名称排序)。
+func (s *AiGatewayService) Blacklist(ctx context.Context) ([]model.AiModelBlacklist, error) {
+	var rows []model.AiModelBlacklist
+	err := s.db.WithContext(ctx).Order("name ASC").Find(&rows).Error
+	return rows, err
+}
+
+// AddBlacklist 把模型名加入黑名单并删除全部渠道缓存中的同名条目;
+// 该模型此后同步 / 探测均被过滤,直到移出黑名单后重新同步。
+func (s *AiGatewayService) AddBlacklist(ctx context.Context, name string) (*model.AiModelBlacklist, error) {
+	name = strings.TrimSpace(name)
+	if name == "" {
+		return nil, fmt.Errorf("模型名不能为空")
+	}
+	row := model.AiModelBlacklist{Name: name}
+	err := s.db.WithContext(ctx).Transaction(func(tx *gorm.DB) error {
+		var n int64
+		if err := tx.Model(&model.AiModelBlacklist{}).Where("name = ?", name).Count(&n).Error; err != nil {
+			return err
+		}
+		if n > 0 {
+			return fmt.Errorf("模型已在黑名单中")
+		}
+		if err := tx.Create(&row).Error; err != nil {
+			return err
+		}
+		return tx.Where("name = ?", name).Delete(&model.AiModelCache{}).Error
+	})
+	if err != nil {
+		return nil, err
+	}
+	return &row, nil
+}
+
+// RemoveBlacklist 把模型移出黑名单;缓存不回填,下次同步 / 探测自然恢复入池。
+func (s *AiGatewayService) RemoveBlacklist(ctx context.Context, id uint) error {
+	res := s.db.WithContext(ctx).Delete(&model.AiModelBlacklist{}, id)
+	if res.Error != nil {
+		return res.Error
+	}
+	if res.RowsAffected == 0 {
+		return fmt.Errorf("黑名单条目不存在")
+	}
+	return nil
+}
+
+// withoutBlacklisted 过滤掉黑名单中的模型,同步与探测入库前统一经此收口。
+// supportedGatewayModels 过滤模型目录:对话模型仅保留实测支持 OpenAI 兼容面的
+// 厂商(xai / meta / openai)——typed chat 面已剔除,google / cohere 对话模型无
+// 上游通路,不入目录(不出现在列表、路由与探测候选);EMBEDDING 模型不受影响。
+func supportedGatewayModels(models []oci.GenAiModel) []oci.GenAiModel {
+	out := make([]oci.GenAiModel, 0, len(models))
+	for _, m := range models {
+		if m.Capability == "CHAT" && !compatChatVendor(m.Name) {
+			continue
+		}
+		out = append(out, m)
+	}
+	return out
+}
+
+func compatChatVendor(model string) bool {
+	for _, prefix := range []string{"xai.", "meta.", "openai."} {
+		if strings.HasPrefix(model, prefix) {
+			return true
+		}
+	}
+	return false
+}
+
+func (s *AiGatewayService) withoutBlacklisted(ctx context.Context, models []oci.GenAiModel) ([]oci.GenAiModel, error) {
+	var names []string
+	if err := s.db.WithContext(ctx).Model(&model.AiModelBlacklist{}).Pluck("name", &names).Error; err != nil {
+		return nil, fmt.Errorf("读取模型黑名单: %w", err)
+	}
+	if len(names) == 0 {
+		return models, nil
+	}
+	banned := make(map[string]bool, len(names))
+	for _, n := range names {
+		banned[n] = true
+	}
+	out := make([]oci.GenAiModel, 0, len(models))
+	for _, m := range models {
+		if !banned[m.Name] {
+			out = append(out, m)
+		}
+	}
+	return out, nil
+}
+
 // ---- 调用日志 ----

 // LogCall 落一条调用日志(仅元数据与用量,永不含请求 / 响应正文),返回落库 ID 供内容日志关联(失败为 0)。
 func (s *AiGatewayService) LogCall(entry model.AiCallLog) uint {
 	entry.ErrMsg = truncateErr(entry.ErrMsg)
-	if err := s.db.Create(&entry).Error; err != nil {
+	stored := false
+	err := s.db.Transaction(func(tx *gorm.DB) error {
+		ok, err := lockAiLogParent(tx, &model.AiChannel{}, entry.ChannelID)
+		if err != nil || !ok {
+			return err
+		}
+		stored = true
+		return tx.Create(&entry).Error
+	})
+	if err != nil {
 		log.Printf("ai call log: %v", err)
 		return 0
 	}
+	if !stored {
+		return 0
+	}
 	return entry.ID
 }

+func lockAiLogParent(tx *gorm.DB, value any, id uint) (bool, error) {
+	if id == 0 {
+		return true, nil
+	}
+	err := tx.Clauses(clause.Locking{Strength: "UPDATE"}).Select("id").First(value, id).Error
+	if errors.Is(err, gorm.ErrRecordNotFound) {
+		return false, nil
+	}
+	return err == nil, err
+}
+
 // CallLogs 分页查询调用日志。
 func (s *AiGatewayService) CallLogs(ctx context.Context, page, size int) ([]model.AiCallLog, int64, error) {
 	if page < 1 {
@@ -555,9 +785,19 @@ func (s *AiGatewayService) UpdateKeyContentLog(ctx context.Context, id uint, hou

 // LogContent 写一条内容日志(调用方已确认密钥开启且未过期);正文截断至 64KB。
 func (s *AiGatewayService) LogContent(entry model.AiContentLog) {
+	if entry.CallLogID == 0 {
+		return
+	}
 	entry.RequestBody = truncateBody(entry.RequestBody)
 	entry.ResponseBody = truncateBody(entry.ResponseBody)
-	if err := s.db.Create(&entry).Error; err != nil {
+	err := s.db.Transaction(func(tx *gorm.DB) error {
+		ok, err := lockAiLogParent(tx, &model.AiCallLog{}, entry.CallLogID)
+		if err != nil || !ok {
+			return err
+		}
+		return tx.Create(&entry).Error
+	})
+	if err != nil {
 		log.Printf("ai content log: %v", err)
 	}
 }
@@ -3,6 +3,7 @@ package service
 import (
 	"context"
 	"errors"
+	"io"
 	"math/rand"
 	"time"

@@ -26,27 +27,31 @@ type aiCandidate struct {
 	modelOcid string
 }

-// Chat 编排非流式调用:选渠道 → 调用 → 可重试错误换渠道(整请求上限 3 次)。
-// group 非空时只在同分组渠道内路由(取自调用密钥)。
-func (s *AiGatewayService) Chat(ctx context.Context, ir aiwire.ChatRequest, group string) (*aiwire.ChatResponse, ChatMeta, error) {
+// RespPassthrough 编排一次非流式直通调用:选渠道(priority→加权随机)→ 调用 →
+// 可重试错误换渠道(整请求上限 3 次)并维护熔断;group 非空时只在同分组渠道内路由。
+// 上游为 OpenAI-compatible /actions/v1/responses(实测可用,无 Oracle 文档合同)。
+func (s *AiGatewayService) RespPassthrough(ctx context.Context, raw []byte, modelName, group string) ([]byte, ChatMeta, error) {
 	meta := ChatMeta{}
 	excluded := map[uint]bool{}
 	var lastErr error
 	for attempt := 0; attempt < 3; attempt++ {
-		cand, err := s.pick(ctx, ir.Model, group, "CHAT", excluded)
+		cand, err := s.pick(ctx, modelName, group, "CHAT", excluded)
 		if err != nil {
 			return nil, meta, firstErr(lastErr, err)
 		}
 		meta.ChannelID, meta.ChannelName = cand.ch.ID, cand.ch.Name
-		resp, err := s.callOnce(ctx, cand, ir)
+		payload, err := s.passthroughOnce(ctx, cand, raw)
 		if err == nil {
 			s.markSuccess(ctx, cand.ch.ID)
-			return resp, meta, nil
+			return payload, meta, nil
 		}
-		if !retryable(err) {
+		retry, penalize := switchable(err)
+		if !retry {
 			return nil, meta, err
 		}
+		if penalize {
 			s.markFailure(ctx, cand.ch.ID)
+		}
 		excluded[cand.ch.ID] = true
 		meta.Retries++
 		lastErr = err
@@ -54,22 +59,22 @@ func (s *AiGatewayService) Chat(ctx context.Context, ir aiwire.ChatRequest, grou
 	return nil, meta, lastErr
 }

-func (s *AiGatewayService) callOnce(ctx context.Context, cand *aiCandidate, ir aiwire.ChatRequest) (*aiwire.ChatResponse, error) {
+func (s *AiGatewayService) passthroughOnce(ctx context.Context, cand *aiCandidate, raw []byte) ([]byte, error) {
 	cred, err := s.configs.credentialsByID(ctx, cand.ch.OciConfigID)
 	if err != nil {
 		return nil, err
 	}
-	return s.client.GenAiChat(ctx, cred, cand.ch.Region, cand.modelOcid, ir)
+	return s.client.GenAiCompatResponses(ctx, cred, cand.ch.Region, raw)
 }

-// OpenStream 编排流式调用:流建立成功后即绑定渠道,建立失败可换渠道重试。
-// group 语义与 Chat 相同。
-func (s *AiGatewayService) OpenStream(ctx context.Context, ir aiwire.ChatRequest, group string) (oci.GenAiStream, ChatMeta, error) {
+// RespPassthroughStream 编排流式直通:流建立成功即绑定渠道,建立失败按 switchable
+// 换渠道重试;建立后的中断不重试、不计熔断(与 OpenStream 语义一致)。
+func (s *AiGatewayService) RespPassthroughStream(ctx context.Context, raw []byte, modelName, group string) (io.ReadCloser, ChatMeta, error) {
 	meta := ChatMeta{}
 	excluded := map[uint]bool{}
 	var lastErr error
 	for attempt := 0; attempt < 3; attempt++ {
-		cand, err := s.pick(ctx, ir.Model, group, "CHAT", excluded)
+		cand, err := s.pick(ctx, modelName, group, "CHAT", excluded)
 		if err != nil {
 			return nil, meta, firstErr(lastErr, err)
 		}
@@ -78,15 +83,18 @@ func (s *AiGatewayService) OpenStream(ctx context.Context, ir aiwire.ChatRequest
 		if err != nil {
 			return nil, meta, err
 		}
-		stream, err := s.client.GenAiChatStream(ctx, cred, cand.ch.Region, cand.modelOcid, ir)
+		stream, err := s.client.GenAiCompatResponsesStream(ctx, cred, cand.ch.Region, raw)
 		if err == nil {
 			s.markSuccess(ctx, cand.ch.ID)
 			return stream, meta, nil
 		}
-		if !retryable(err) {
+		retry, penalize := switchable(err)
+		if !retry {
 			return nil, meta, err
 		}
+		if penalize {
 			s.markFailure(ctx, cand.ch.ID)
+		}
 		excluded[cand.ch.ID] = true
 		meta.Retries++
 		lastErr = err
@@ -102,12 +110,17 @@ func firstErr(lastErr, pickErr error) error {
 	return pickErr
 }

-// retryable 判定是否换渠道重试:429 / 5xx / 网络错误可重试,其余 4xx 直接透传。
-func retryable(err error) bool {
-	if status, ok := oci.ServiceStatus(err); ok {
-		return status == 429 || status >= 500
+// switchable 判断调用失败是否换渠道重试、是否计入熔断:模型级不可用(微调基座
+// 400 / 实体不存在 404)换渠道但不计熔断——这是模型×区域供给问题而非渠道健康
+// 问题,持久解决靠加入模型黑名单;429 / 5xx / 网络错误换渠道并计熔断;其余 4xx 直接透传。
+func switchable(err error) (retry, penalize bool) {
+	if oci.IsModelUnavailable(err) {
+		return true, false
 	}
-	return true
+	if status, ok := oci.ServiceStatus(err); ok {
+		return status == 429 || status >= 500, true
+	}
+	return true, true
 }

 // pick 选出支持该模型的最优渠道:能力匹配 → 启用 → 分组匹配 → 未熔断 → 最小优先级组 → 加权随机。
@@ -151,6 +164,9 @@ func (s *AiGatewayService) modelChannels(ctx context.Context, modelName, capabil
 	} else {
 		q = q.Where("capability = ?", capability)
 	}
+	if s.FilterDeprecated() {
+		q = q.Where("deprecated_at IS NULL")
+	}
 	var rows []model.AiModelCache
 	if err := q.Find(&rows).Error; err != nil {
 		return nil, nil, err
@@ -221,10 +237,13 @@ func (s *AiGatewayService) Embeddings(ctx context.Context, req aiwire.Embeddings
 			s.markSuccess(ctx, cand.ch.ID)
 			return resp, meta, nil
 		}
-		if !retryable(err) {
+		retry, penalize := switchable(err)
+		if !retry {
 			return nil, meta, err
 		}
+		if penalize {
 			s.markFailure(ctx, cand.ch.ID)
+		}
 		excluded[cand.ch.ID] = true
 		meta.Retries++
 		lastErr = err
@@ -0,0 +1,284 @@
+package service
+
+import (
+	"context"
+	"encoding/json"
+	"fmt"
+	"strings"
+	"time"
+
+	"oci-portal/internal/aiwire"
+	"oci-portal/internal/model"
+	"oci-portal/internal/oci"
+)
+
+// moderationsMaxInputs 限制单次审核条数,防止批量滥用拖垮渠道配额。
+const moderationsMaxInputs = 8
+
+// SpeechBodyNormalize 校验并规范化 TTS 请求体:model / input 必填;
+// xAI 上游把 language 当必填(实测缺失返回 422),缺省注入 "auto"。
+// 其余字段(voice / response_format / extra_body 平铺项)原样保留。
+func SpeechBodyNormalize(raw []byte) (string, []byte, error) {
+	dec := json.NewDecoder(strings.NewReader(string(raw)))
+	dec.UseNumber()
+	var body map[string]any
+	if err := dec.Decode(&body); err != nil {
+		return "", nil, fmt.Errorf("解析请求体: %w", err)
+	}
+	modelName, _ := body["model"].(string)
+	input, _ := body["input"].(string)
+	if strings.TrimSpace(modelName) == "" || strings.TrimSpace(input) == "" {
+		return "", nil, fmt.Errorf("model 与 input 不能为空")
+	}
+	if lang, ok := body["language"].(string); !ok || strings.TrimSpace(lang) == "" {
+		body["language"] = "auto"
+	}
+	out, err := json.Marshal(body)
+	if err != nil {
+		return "", nil, fmt.Errorf("重组请求体: %w", err)
+	}
+	return modelName, out, nil
+}
+
+// ttsDefaultModel 是 /ai/v1/tts 缺省注入的模型(xAI 官方格式无 model 字段)。
+const ttsDefaultModel = "xai.grok-tts"
+
+// TtsBodyConvert 把 xAI 官方 TTS 格式转换为上游 OpenAI 兼容 audio/speech 形态:
+// text→input、voice_id→voice,text 与 language 必填(对齐 xAI 官方);
+// model 为网关扩展字段,缺省注入 ttsDefaultModel;其余字段原样保留。
+func TtsBodyConvert(raw []byte) (string, []byte, error) {
+	dec := json.NewDecoder(strings.NewReader(string(raw)))
+	dec.UseNumber()
+	var body map[string]any
+	if err := dec.Decode(&body); err != nil {
+		return "", nil, fmt.Errorf("解析请求体: %w", err)
+	}
+	text, _ := body["text"].(string)
+	language, _ := body["language"].(string)
+	if strings.TrimSpace(text) == "" || strings.TrimSpace(language) == "" {
+		return "", nil, fmt.Errorf("text 与 language 不能为空")
+	}
+	modelName, _ := body["model"].(string)
+	if strings.TrimSpace(modelName) == "" {
+		modelName = ttsDefaultModel
+		body["model"] = modelName
+	}
+	delete(body, "text")
+	body["input"] = text
+	if voice, ok := body["voice_id"].(string); ok && strings.TrimSpace(voice) != "" {
+		body["voice"] = voice
+	}
+	delete(body, "voice_id")
+	out, err := json.Marshal(body)
+	if err != nil {
+		return "", nil, fmt.Errorf("重组请求体: %w", err)
+	}
+	return modelName, out, nil
+}
+
+// Speech 编排 TTS 调用:按 TTS 能力选渠道,可重试错误换渠道(上限 3 次)。
+func (s *AiGatewayService) Speech(ctx context.Context, modelName string, body []byte, group string) ([]byte, string, ChatMeta, error) {
+	meta := ChatMeta{}
+	excluded := map[uint]bool{}
+	var lastErr error
+	for attempt := 0; attempt < 3; attempt++ {
+		cand, err := s.pick(ctx, modelName, group, "TTS", excluded)
+		if err != nil {
+			return nil, "", meta, firstErr(lastErr, err)
+		}
+		meta.ChannelID, meta.ChannelName = cand.ch.ID, cand.ch.Name
+		cred, err := s.configs.credentialsByID(ctx, cand.ch.OciConfigID)
+		if err != nil {
+			return nil, "", meta, err
+		}
+		audio, contentType, err := s.client.GenAiCompatSpeech(ctx, cred, cand.ch.Region, body)
+		if err == nil {
+			s.markSuccess(ctx, cand.ch.ID)
+			return audio, contentType, meta, nil
+		}
+		if done := s.retryStep(ctx, cand.ch.ID, err, excluded, &meta, &lastErr); done {
+			return nil, "", meta, err
+		}
+	}
+	return nil, "", meta, lastErr
+}
+
+// Rerank 编排文档重排:按 RERANK 能力选渠道,装配 Jina 风格响应。
+func (s *AiGatewayService) Rerank(ctx context.Context, req aiwire.RerankRequest, group string) (*aiwire.RerankResponse, ChatMeta, error) {
+	meta := ChatMeta{}
+	excluded := map[uint]bool{}
+	var lastErr error
+	for attempt := 0; attempt < 3; attempt++ {
+		cand, err := s.pick(ctx, req.Model, group, "RERANK", excluded)
+		if err != nil {
+			return nil, meta, firstErr(lastErr, err)
+		}
+		meta.ChannelID, meta.ChannelName = cand.ch.ID, cand.ch.Name
+		cred, err := s.configs.credentialsByID(ctx, cand.ch.OciConfigID)
+		if err != nil {
+			return nil, meta, err
+		}
+		ranks, err := s.client.GenAiRerank(ctx, cred, cand.ch.Region, cand.modelOcid, req.Query, req.Documents, req.TopN)
+		if err == nil {
+			s.markSuccess(ctx, cand.ch.ID)
+			return rerankResponse(req, ranks), meta, nil
+		}
+		if done := s.retryStep(ctx, cand.ch.ID, err, excluded, &meta, &lastErr); done {
+			return nil, meta, err
+		}
+	}
+	return nil, meta, lastErr
+}
+
+// retryStep 统一处理换渠道重试:不可重试返回 true 终止,可重试记罚并排除渠道。
+func (s *AiGatewayService) retryStep(ctx context.Context, chID uint, err error, excluded map[uint]bool, meta *ChatMeta, lastErr *error) bool {
+	retry, penalize := switchable(err)
+	if !retry {
+		return true
+	}
+	if penalize {
+		s.markFailure(ctx, chID)
+	}
+	excluded[chID] = true
+	meta.Retries++
+	*lastErr = err
+	return false
+}
+
+// rerankResponse 把上游排序结果装配为对外响应;return_documents 时回填原文。
+func rerankResponse(req aiwire.RerankRequest, ranks []oci.RerankRank) *aiwire.RerankResponse {
+	out := &aiwire.RerankResponse{Model: req.Model, Results: make([]aiwire.RerankResult, 0, len(ranks))}
+	withDoc := req.ReturnDocuments != nil && *req.ReturnDocuments
+	for _, r := range ranks {
+		if r.Index < 0 || r.Index >= len(req.Documents) {
+			continue
+		}
+		item := aiwire.RerankResult{Index: r.Index, RelevanceScore: r.Score}
+		if withDoc {
+			item.Document = &aiwire.RerankDocument{Text: req.Documents[r.Index]}
+		}
+		out.Results = append(out.Results, item)
+	}
+	return out
+}
+
+// ModerationInputs 解析 moderations 的 input 字段:string 或 []string,
+// 条数与空值校验(上限 moderationsMaxInputs)。
+func ModerationInputs(raw json.RawMessage) ([]string, error) {
+	var single string
+	if err := json.Unmarshal(raw, &single); err == nil {
+		if strings.TrimSpace(single) == "" {
+			return nil, fmt.Errorf("input 不能为空")
+		}
+		return []string{single}, nil
+	}
+	var many []string
+	if err := json.Unmarshal(raw, &many); err != nil {
+		return nil, fmt.Errorf("input 需为字符串或字符串数组")
+	}
+	if len(many) == 0 || len(many) > moderationsMaxInputs {
+		return nil, fmt.Errorf("input 条数需在 1~%d 之间", moderationsMaxInputs)
+	}
+	for _, item := range many {
+		if strings.TrimSpace(item) == "" {
+			return nil, fmt.Errorf("input 含空条目")
+		}
+	}
+	return many, nil
+}
+
+// Moderations 编排内容审核:guardrails 为服务级 API 无模型维度,
+// 按分组选启用渠道(整批同渠道),可重试错误换渠道(上限 3 次)。
+func (s *AiGatewayService) Moderations(ctx context.Context, id string, inputs []string, group string) (*aiwire.ModerationsResponse, ChatMeta, error) {
+	meta := ChatMeta{}
+	excluded := map[uint]bool{}
+	var lastErr error
+	for attempt := 0; attempt < 3; attempt++ {
+		ch, err := s.pickGuardChannel(ctx, group, excluded)
+		if err != nil {
+			return nil, meta, firstErr(lastErr, err)
+		}
+		meta.ChannelID, meta.ChannelName = ch.ID, ch.Name
+		resp, err := s.moderateOnce(ctx, ch, id, inputs)
+		if err == nil {
+			s.markSuccess(ctx, ch.ID)
+			return resp, meta, nil
+		}
+		if done := s.retryStep(ctx, ch.ID, err, excluded, &meta, &lastErr); done {
+			return nil, meta, err
+		}
+	}
+	return nil, meta, lastErr
+}
+
+// pickGuardChannel 按分组挑一个启用且未熔断的渠道(权重加成,不查模型缓存)。
+func (s *AiGatewayService) pickGuardChannel(ctx context.Context, group string, excluded map[uint]bool) (model.AiChannel, error) {
+	q := s.db.WithContext(ctx).Where("enabled = ?", true)
+	if group != "" {
+		q = q.Where("channel_group = ?", group)
+	}
+	var channels []model.AiChannel
+	if err := q.Find(&channels).Error; err != nil {
+		return model.AiChannel{}, err
+	}
+	now := time.Now()
+	avail := channels[:0]
+	for _, ch := range channels {
+		if excluded[ch.ID] || (ch.DisabledUntil != nil && ch.DisabledUntil.After(now)) {
+			continue
+		}
+		avail = append(avail, ch)
+	}
+	if len(avail) == 0 {
+		return model.AiChannel{}, ErrAiNoChannel
+	}
+	return weightedPick(topPriority(avail)), nil
+}
+
+// moderateOnce 在单渠道上逐条审核并装配 OpenAI moderations 外壳。
+func (s *AiGatewayService) moderateOnce(ctx context.Context, ch model.AiChannel, id string, inputs []string) (*aiwire.ModerationsResponse, error) {
+	cred, err := s.configs.credentialsByID(ctx, ch.OciConfigID)
+	if err != nil {
+		return nil, err
+	}
+	out := &aiwire.ModerationsResponse{ID: id, Model: "oci-guardrails",
+		Results: make([]aiwire.ModerationResult, 0, len(inputs))}
+	for _, text := range inputs {
+		outcome, err := s.client.GenAiApplyGuardrails(ctx, cred, ch.Region, text)
+		if err != nil {
+			return nil, err
+		}
+		out.Results = append(out.Results, moderationResult(outcome))
+	}
+	return out, nil
+}
+
+// moderationFlagThreshold 是判定违规的得分阈值(上游 CM / PI 为二值得分)。
+const moderationFlagThreshold = 0.5
+
+// moderationResult 把 guardrails 结果映射为 OpenAI moderations 条目:
+// flagged 由内容审核与提示注入判定,PII 命中只作扩展信息不参与 flagged。
+func moderationResult(outcome *oci.GuardrailsOutcome) aiwire.ModerationResult {
+	res := aiwire.ModerationResult{Categories: map[string]bool{}, CategoryScores: map[string]float64{}}
+	for _, cat := range outcome.Categories {
+		key := strings.ToLower(cat.Name)
+		res.Categories[key] = cat.Score >= moderationFlagThreshold
+		res.CategoryScores[key] = cat.Score
+		if res.Categories[key] {
+			res.Flagged = true
+		}
+	}
+	if outcome.PromptInjectionScore != nil {
+		hit := *outcome.PromptInjectionScore >= moderationFlagThreshold
+		res.Categories["prompt_injection"] = hit
+		res.CategoryScores["prompt_injection"] = *outcome.PromptInjectionScore
+		if hit {
+			res.Flagged = true
+		}
+	}
+	for _, p := range outcome.Pii {
+		res.Pii = append(res.Pii, aiwire.ModerationPii{Text: p.Text, Label: p.Label,
+			Score: p.Score, Offset: p.Offset, Length: p.Length})
+	}
+	return res
+}
@@ -0,0 +1,222 @@
+package service
+
+import (
+	"context"
+	"encoding/json"
+	"errors"
+	"strings"
+	"testing"
+	"time"
+
+	"oci-portal/internal/aiwire"
+	"oci-portal/internal/model"
+	"oci-portal/internal/oci"
+)
+
+// TestSpeechBodyNormalize 断言 TTS 请求体校验与 language 缺省注入。
+func TestSpeechBodyNormalize(t *testing.T) {
+	tests := []struct {
+		name     string
+		raw      string
+		wantErr  bool
+		wantLang string
+	}{
+		{"缺 model 拒绝", `{"input":"你好"}`, true, ""},
+		{"缺 input 拒绝", `{"model":"xai.grok-tts"}`, true, ""},
+		{"language 缺省注入 auto", `{"model":"xai.grok-tts","input":"你好","voice":"ara"}`, false, "auto"},
+		{"language 已有保留", `{"model":"xai.grok-tts","input":"你好","language":"zh"}`, false, "zh"},
+		{"非 JSON 拒绝", `<html>`, true, ""},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			modelName, body, err := SpeechBodyNormalize([]byte(tt.raw))
+			if (err != nil) != tt.wantErr {
+				t.Fatalf("err = %v, wantErr %v", err, tt.wantErr)
+			}
+			if err != nil {
+				return
+			}
+			var out map[string]any
+			_ = json.Unmarshal(body, &out)
+			if modelName != "xai.grok-tts" || out["language"] != tt.wantLang {
+				t.Fatalf("model=%s language=%v, want %s", modelName, out["language"], tt.wantLang)
+			}
+		})
+	}
+}
+
+// TestTtsBodyConvert 断言 xAI 官方 TTS 格式到 OpenAI 兼容形态的转换。
+func TestTtsBodyConvert(t *testing.T) {
+	tests := []struct {
+		name      string
+		raw       string
+		wantErr   bool
+		wantModel string
+	}{
+		{"缺 text 拒绝", `{"language":"zh"}`, true, ""},
+		{"缺 language 拒绝", `{"text":"你好"}`, true, ""},
+		{"缺省注入默认模型", `{"text":"你好","language":"zh"}`, false, "xai.grok-tts"},
+		{"model 扩展字段可覆盖", `{"model":"xai.other-tts","text":"你好","language":"auto"}`, false, "xai.other-tts"},
+		{"非 JSON 拒绝", `<html>`, true, ""},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			modelName, _, err := TtsBodyConvert([]byte(tt.raw))
+			if (err != nil) != tt.wantErr {
+				t.Fatalf("err = %v, wantErr %v", err, tt.wantErr)
+			}
+			if err != nil {
+				return
+			}
+			if modelName != tt.wantModel {
+				t.Fatalf("model = %s, want %s", modelName, tt.wantModel)
+			}
+		})
+	}
+}
+
+// TestTtsBodyConvertMapping 断言字段映射与未知字段保留。
+func TestTtsBodyConvertMapping(t *testing.T) {
+	raw := `{"text":"你好","language":"zh","voice_id":"ara","speed":1.2,` +
+		`"output_format":{"codec":"mp3","sample_rate":44100}}`
+	_, body, err := TtsBodyConvert([]byte(raw))
+	if err != nil {
+		t.Fatalf("err = %v", err)
+	}
+	var out map[string]any
+	_ = json.Unmarshal(body, &out)
+	if out["input"] != "你好" || out["voice"] != "ara" {
+		t.Fatalf("input/voice 映射错误: %v", out)
+	}
+	if _, ok := out["text"]; ok {
+		t.Fatal("text 字段应被移除")
+	}
+	if _, ok := out["voice_id"]; ok {
+		t.Fatal("voice_id 字段应被移除")
+	}
+	of, _ := out["output_format"].(map[string]any)
+	if of == nil || of["codec"] != "mp3" {
+		t.Fatalf("output_format 应原样保留: %v", out["output_format"])
+	}
+	if out["language"] != "zh" || out["speed"] == nil {
+		t.Fatalf("language/speed 应保留: %v", out)
+	}
+}
+
+// TestModerationInputs 断言 input 的 string / []string 解析与边界校验。
+func TestModerationInputs(t *testing.T) {
+	tests := []struct {
+		name    string
+		raw     string
+		wantN   int
+		wantErr bool
+	}{
+		{"单字符串", `"hello"`, 1, false},
+		{"数组", `["a","b"]`, 2, false},
+		{"空字符串拒绝", `""`, 0, true},
+		{"空数组拒绝", `[]`, 0, true},
+		{"含空条目拒绝", `["a",""]`, 0, true},
+		{"超上限拒绝", `["1","2","3","4","5","6","7","8","9"]`, 0, true},
+		{"非法类型拒绝", `{"x":1}`, 0, true},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			got, err := ModerationInputs(json.RawMessage(tt.raw))
+			if (err != nil) != tt.wantErr || len(got) != tt.wantN {
+				t.Fatalf("got %v (err=%v), want n=%d wantErr=%v", got, err, tt.wantN, tt.wantErr)
+			}
+		})
+	}
+}
+
+// TestModerationResultMapping 断言 guardrails 结果到 OpenAI 外壳的映射与 flagged 判定。
+func TestModerationResultMapping(t *testing.T) {
+	one := 1.0
+	zero := 0.0
+	tests := []struct {
+		name        string
+		outcome     oci.GuardrailsOutcome
+		wantFlagged bool
+		wantPii     int
+	}{
+		{"内容审核命中", oci.GuardrailsOutcome{Categories: []oci.GuardrailCategory{{Name: "OVERALL", Score: 1}}}, true, 0},
+		{"提示注入命中", oci.GuardrailsOutcome{PromptInjectionScore: &one}, true, 0},
+		{"仅 PII 不 flag", oci.GuardrailsOutcome{Categories: []oci.GuardrailCategory{{Name: "OVERALL", Score: 0}},
+			PromptInjectionScore: &zero, Pii: []oci.GuardrailPiiHit{{Text: "Jane", Label: "PERSON", Score: 0.99}}}, false, 1},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			res := moderationResult(&tt.outcome)
+			if res.Flagged != tt.wantFlagged || len(res.Pii) != tt.wantPii {
+				t.Fatalf("res = %+v, want flagged=%v pii=%d", res, tt.wantFlagged, tt.wantPii)
+			}
+		})
+	}
+}
+
+// TestAiRerank 断言重排编排:能力路由、top_n 透传、return_documents 回填与越界防御。
+func TestAiRerank(t *testing.T) {
+	client := &gatewayStubClient{
+		fakeClient:  &fakeClient{tenancy: oci.TenancyInfo{Name: "t"}},
+		rerankRanks: []oci.RerankRank{{Index: 1, Score: 0.9}, {Index: 0, Score: 0.4}, {Index: 9, Score: 0.1}},
+	}
+	gw, svc := newTestGateway(t, client)
+	cfg := importAliveConfig(t, svc)
+	ch := seedChannel(t, gw, cfg.ID, "us-chicago-1", 1, 1)
+	gw.db.Create(&model.AiModelCache{ChannelID: ch.ID, ModelOcid: "ocid1..rr", Name: "cohere.rerank-v4.0-fast",
+		Vendor: "cohere", Capability: "RERANK", SyncedAt: time.Now()})
+	ctx := context.Background()
+	yes := true
+	req := aiwire.RerankRequest{Model: "cohere.rerank-v4.0-fast", Query: "q", Documents: []string{"d0", "d1"}, ReturnDocuments: &yes}
+	resp, _, err := gw.Rerank(ctx, req, "")
+	if err != nil || len(resp.Results) != 2 {
+		t.Fatalf("Rerank = %+v, %v(越界 index 应被丢弃)", resp, err)
+	}
+	if resp.Results[0].Index != 1 || resp.Results[0].Document == nil || resp.Results[0].Document.Text != "d1" {
+		t.Fatalf("results[0] = %+v", resp.Results[0])
+	}
+	// 对话模型名打 rerank:能力不匹配 → 未知模型
+	if _, _, err := gw.Rerank(ctx, aiwire.RerankRequest{Model: "meta.llama-3.3-70b-instruct", Query: "q", Documents: []string{"d"}}, ""); !errors.Is(err, ErrAiUnknownModel) {
+		t.Errorf("chat 模型走 rerank err = %v, want ErrAiUnknownModel", err)
+	}
+}
+
+// TestAiSpeech 断言 TTS 编排走 TTS 能力路由并透传音频与 Content-Type。
+func TestAiSpeech(t *testing.T) {
+	client := &gatewayStubClient{
+		fakeClient:  &fakeClient{tenancy: oci.TenancyInfo{Name: "t"}},
+		speechAudio: []byte{0xFF, 0xF3}, speechCT: "audio/mpeg",
+	}
+	gw, svc := newTestGateway(t, client)
+	cfg := importAliveConfig(t, svc)
+	ch := seedChannel(t, gw, cfg.ID, "us-chicago-1", 1, 1)
+	gw.db.Create(&model.AiModelCache{ChannelID: ch.ID, ModelOcid: "ocid1..tts", Name: "xai.grok-tts",
+		Vendor: "xai", Capability: "TTS", SyncedAt: time.Now()})
+	audio, ct, _, err := gw.Speech(context.Background(), "xai.grok-tts", []byte(`{"model":"xai.grok-tts","input":"你好","language":"auto"}`), "")
+	if err != nil || ct != "audio/mpeg" || len(audio) != 2 {
+		t.Fatalf("Speech = %d bytes, ct=%s, %v", len(audio), ct, err)
+	}
+}
+
+// TestAiModerations 断言审核编排:无模型维度按分组选渠道,多条输入逐条聚合。
+func TestAiModerations(t *testing.T) {
+	one := 1.0
+	client := &gatewayStubClient{
+		fakeClient:   &fakeClient{tenancy: oci.TenancyInfo{Name: "t"}},
+		guardOutcome: &oci.GuardrailsOutcome{Categories: []oci.GuardrailCategory{{Name: "OVERALL", Score: 1}}, PromptInjectionScore: &one},
+	}
+	gw, svc := newTestGateway(t, client)
+	cfg := importAliveConfig(t, svc)
+	seedChannel(t, gw, cfg.ID, "us-chicago-1", 1, 1)
+	resp, meta, err := gw.Moderations(context.Background(), "modr_test", []string{"a", "b"}, "")
+	if err != nil || len(resp.Results) != 2 || !resp.Results[0].Flagged || resp.ID != "modr_test" {
+		t.Fatalf("Moderations = %+v, meta=%+v, %v", resp, meta, err)
+	}
+	if !strings.Contains(resp.Model, "guardrails") {
+		t.Errorf("model = %s", resp.Model)
+	}
+	// 分组不匹配 → 无可用渠道
+	if _, _, err := gw.Moderations(context.Background(), "modr_x", []string{"a"}, "ghost-group"); !errors.Is(err, ErrAiNoChannel) {
+		t.Errorf("ghost 分组 err = %v, want ErrAiNoChannel", err)
+	}
+}
@@ -0,0 +1,29 @@
+package service
+
+import (
+	"strings"
+	"testing"
+
+	"oci-portal/internal/oci"
+)
+
+// TestSupportedGatewayModels 断言 CHAT 目录仅保留兼容面厂商,EMBEDDING 不受影响。
+func TestSupportedGatewayModels(t *testing.T) {
+	in := []oci.GenAiModel{
+		{Name: "xai.grok-4.3", Capability: "CHAT"},
+		{Name: "meta.llama-4-maverick-17b-128e", Capability: "CHAT"},
+		{Name: "openai.gpt-oss-120b", Capability: "CHAT"},
+		{Name: "google.gemini-2.5-flash", Capability: "CHAT"},
+		{Name: "cohere.command-a-03-2025", Capability: "CHAT"},
+		{Name: "cohere.embed-v4.0", Capability: "EMBEDDING"},
+	}
+	out := supportedGatewayModels(in)
+	names := make([]string, 0, len(out))
+	for _, m := range out {
+		names = append(names, m.Name)
+	}
+	want := "xai.grok-4.3,meta.llama-4-maverick-17b-128e,openai.gpt-oss-120b,cohere.embed-v4.0"
+	if got := strings.Join(names, ","); got != want {
+		t.Fatalf("过滤结果 = %s, want %s", got, want)
+	}
+}
@@ -1,8 +1,11 @@
 package service

 import (
+	"bytes"
 	"context"
 	"errors"
+	"fmt"
+	"io"
 	"strings"
 	"testing"
 	"time"
@@ -13,14 +16,28 @@ import (
 )

 // stubServiceError 实现 common.ServiceError,用于模拟带状态码的 OCI 服务端错误。
-type stubServiceError struct{ status int }
+type stubServiceError struct {
+	status int
+	msg    string
+}

 func (e stubServiceError) Error() string          { return "stub service error" }
 func (e stubServiceError) GetHTTPStatusCode() int { return e.status }
-func (e stubServiceError) GetMessage() string      { return "stub" }
+func (e stubServiceError) GetMessage() string {
+	if e.msg != "" {
+		return e.msg
+	}
+	return "stub"
+}
 func (e stubServiceError) GetCode() string         { return "Stub" }
 func (e stubServiceError) GetOpcRequestID() string { return "req-1" }

+// finetuneBaseErr 模拟「微调基座模型不可按需调用」的 OCI 400。
+func finetuneBaseErr() stubServiceError {
+	return stubServiceError{status: 400,
+		msg: "Not allowed to call finetune base model ocid1.generativeaimodel.oc1.eu-frankfurt-1.tpel5q, use Endpoint: false"}
+}
+
 // gatewayStubClient 覆写 GenAI 四方法;chatErrs 逐次弹出以模拟先失败后成功。
 type gatewayStubClient struct {
 	*fakeClient
@@ -29,14 +46,67 @@ type gatewayStubClient struct {
 	modelsErr error
 	probeCode int
 	probeErr  error
-	chatResp  *aiwire.ChatResponse
-	chatErrs  []error
+	// probeSeq 非空时逐次弹出,模拟按候选依次试调;弹尽后回落 probeCode/probeErr
+	probeSeq []probeResult
+	// probedNames 记录试调过的模型名,供断言候选过滤
+	probedNames []string
 	chatCalls   int
 	regions     []string

 	embedVecs  [][]float32
 	embedUsage *aiwire.Usage
 	embedErr   error
+
+	passPayload []byte
+	passErrs    []error
+	passCalls   int
+	passRegions []string
+
+	speechAudio  []byte
+	speechCT     string
+	speechErr    error
+	rerankRanks  []oci.RerankRank
+	rerankErr    error
+	guardOutcome *oci.GuardrailsOutcome
+	guardErr     error
+}
+
+func (f *gatewayStubClient) GenAiCompatSpeech(ctx context.Context, cred oci.Credentials, region string, body []byte) ([]byte, string, error) {
+	return f.speechAudio, f.speechCT, f.speechErr
+}
+
+func (f *gatewayStubClient) GenAiRerank(ctx context.Context, cred oci.Credentials, region, modelOcid, query string, documents []string, topN *int) ([]oci.RerankRank, error) {
+	return f.rerankRanks, f.rerankErr
+}
+
+func (f *gatewayStubClient) GenAiApplyGuardrails(ctx context.Context, cred oci.Credentials, region, text string) (*oci.GuardrailsOutcome, error) {
+	return f.guardOutcome, f.guardErr
+}
+
+func (f *gatewayStubClient) GenAiCompatResponses(ctx context.Context, cred oci.Credentials, region string, body []byte) ([]byte, error) {
+	f.passCalls++
+	f.passRegions = append(f.passRegions, region)
+	if len(f.passErrs) > 0 {
+		err := f.passErrs[0]
+		f.passErrs = f.passErrs[1:]
+		if err != nil {
+			return nil, err
+		}
+	}
+	return f.passPayload, nil
+}
+
+func (f *gatewayStubClient) GenAiCompatResponsesStream(ctx context.Context, cred oci.Credentials, region string, body []byte) (io.ReadCloser, error) {
+	f.passCalls++
+	f.passRegions = append(f.passRegions, region)
+	if len(f.passErrs) > 0 {
+		err := f.passErrs[0]
+		f.passErrs = f.passErrs[1:]
+		if err != nil {
+			return nil, err
+		}
+	}
+	return io.NopCloser(bytes.NewReader(f.passPayload)), nil
 }

 func (f *gatewayStubClient) GenAiEmbed(ctx context.Context, cred oci.Credentials, region, modelOcid string, inputs []string, dimensions *int) ([][]float32, *aiwire.Usage, error) {
@@ -47,27 +117,26 @@ func (f *gatewayStubClient) ListGenAiModels(ctx context.Context, cred oci.Creden
 	return f.models, f.modelsErr
 }

-func (f *gatewayStubClient) GenAiProbeChat(ctx context.Context, cred oci.Credentials, region, modelOcid, modelName string) (int, error) {
-	return f.probeCode, f.probeErr
+// probeResult 是 gatewayStubClient.probeSeq 的单次探测结果。
+type probeResult struct {
+	code int
+	err  error
 }

-func (f *gatewayStubClient) GenAiChat(ctx context.Context, cred oci.Credentials, region, modelOcid string, ir aiwire.ChatRequest) (*aiwire.ChatResponse, error) {
-	f.chatCalls++
-	f.regions = append(f.regions, region)
-	if len(f.chatErrs) > 0 {
-		err := f.chatErrs[0]
-		f.chatErrs = f.chatErrs[1:]
-		if err != nil {
-			return nil, err
+func (f *gatewayStubClient) GenAiProbeChat(ctx context.Context, cred oci.Credentials, region, modelOcid, modelName string) (int, error) {
+	f.probedNames = append(f.probedNames, modelName)
+	if len(f.probeSeq) > 0 {
+		r := f.probeSeq[0]
+		f.probeSeq = f.probeSeq[1:]
+		return r.code, r.err
 	}
-	}
-	return f.chatResp, nil
+	return f.probeCode, f.probeErr
 }

 func newTestGateway(t *testing.T, client oci.Client) (*AiGatewayService, *OciConfigService) {
 	t.Helper()
 	svc := newTestService(t, client)
-	if err := svc.db.AutoMigrate(&model.AiKey{}, &model.AiChannel{}, &model.AiModelCache{}, &model.AiCallLog{}); err != nil {
+	if err := svc.db.AutoMigrate(&model.Setting{}, &model.AiKey{}, &model.AiChannel{}, &model.AiModelCache{}, &model.AiModelBlacklist{}, &model.AiCallLog{}); err != nil {
 		t.Fatalf("auto migrate ai tables: %v", err)
 	}
 	return NewAiGatewayService(svc.db, svc, client), svc
@@ -77,14 +146,14 @@ func TestAiKeyLifecycle(t *testing.T) {
 	gw, _ := newTestGateway(t, &gatewayStubClient{fakeClient: &fakeClient{tenancy: oci.TenancyInfo{Name: "t"}}})
 	ctx := context.Background()

-	raw, key, err := gw.CreateKey(ctx, "免费-ai-api-key", "test-api-key", "")
+	raw, key, err := gw.CreateKey(ctx, "免费-ai-api-key", "test-api-key", "", nil)
 	if err != nil || raw != "test-api-key" || key.Tail != "-key" {
 		t.Fatalf("CreateKey 自定义值 = %q %+v, %v", raw, key, err)
 	}
-	if _, _, err := gw.CreateKey(ctx, "short", "abc", ""); err == nil {
+	if _, _, err := gw.CreateKey(ctx, "short", "abc", "", nil); err == nil {
 		t.Error("过短自定义密钥应被拒绝")
 	}
-	auto, _, err := gw.CreateKey(ctx, "auto", "", "")
+	auto, _, err := gw.CreateKey(ctx, "auto", "", "", nil)
 	if err != nil || len(auto) < 40 || auto[:3] != "sk-" {
 		t.Fatalf("CreateKey 随机值 = %q, %v", auto, err)
 	}
@@ -96,7 +165,7 @@ func TestAiKeyLifecycle(t *testing.T) {
 		t.Errorf("错误密钥 err = %v", err)
 	}
 	off := false
-	_ = gw.UpdateKey(ctx, key.ID, "", &off, nil)
+	_ = gw.UpdateKey(ctx, key.ID, "", &off, nil, nil)
 	if _, err := gw.VerifyKey(ctx, "test-api-key"); !errors.Is(err, ErrAiKeyInvalid) {
 		t.Errorf("禁用后 VerifyKey err = %v", err)
 	}
@@ -159,48 +228,43 @@ func seedChannel(t *testing.T, gw *AiGatewayService, cfgID uint, region string,
 	return ch
 }

-func TestAiChatRetrySwitchesChannel(t *testing.T) {
-	client := &gatewayStubClient{
-		fakeClient: &fakeClient{tenancy: oci.TenancyInfo{Name: "t"}},
-		chatResp:   &aiwire.ChatResponse{Model: "meta.llama-3.3-70b-instruct", Choices: []aiwire.Choice{{Message: aiwire.ChatMessage{Role: "assistant", Content: aiwire.NewTextContent("hi")}, FinishReason: "stop"}}},
-		chatErrs:   []error{stubServiceError{status: 429}},
-	}
-	gw, svc := newTestGateway(t, client)
+// TestFilterDeprecatedModels 断言「过滤弃用模型」开关对列表 / 路由的过滤与持久化。
+func TestFilterDeprecatedModels(t *testing.T) {
+	gw, svc := newTestGateway(t, &gatewayStubClient{fakeClient: &fakeClient{tenancy: oci.TenancyInfo{Name: "t"}}})
 	cfg := importAliveConfig(t, svc)
-	// 两个同优先级渠道
-	seedChannel(t, gw, cfg.ID, "eu-frankfurt-1", 1, 1)
-	seedChannel(t, gw, cfg.ID, "us-chicago-1", 1, 1)
+	ch := seedChannel(t, gw, cfg.ID, "eu-frankfurt-1", 1, 1)
+	dep := time.Now().Add(-24 * time.Hour)
+	old := &model.AiModelCache{ChannelID: ch.ID, ModelOcid: "ocid1..dep", Name: "meta.llama-old",
+		Vendor: "meta", SyncedAt: time.Now(), DeprecatedAt: &dep}
+	if err := gw.db.Create(old).Error; err != nil {
+		t.Fatalf("seed deprecated cache: %v", err)
+	}
 	ctx := context.Background()

-	resp, meta, err := gw.Chat(ctx, aiwire.ChatRequest{Model: "meta.llama-3.3-70b-instruct", Messages: []aiwire.ChatMessage{{Role: "user", Content: aiwire.NewTextContent("你好")}}}, "")
-	if err != nil || resp == nil {
-		t.Fatalf("Chat = %v, %v", resp, err)
+	list, _ := gw.GatewayModels(ctx, "")
+	if len(list.Data) != 2 {
+		t.Fatalf("开关关:应含弃用模型,got %d", len(list.Data))
 	}
-	if meta.Retries != 1 || client.chatCalls != 2 {
-		t.Errorf("应换渠道重试一次: retries=%d calls=%d", meta.Retries, client.chatCalls)
+	if err := gw.SetFilterDeprecated(ctx, true); err != nil {
+		t.Fatalf("SetFilterDeprecated: %v", err)
 	}
-	if len(client.regions) != 2 && client.regions[0] == client.regions[1] {
-		t.Errorf("重试未换渠道: %v", client.regions)
+	list, _ = gw.GatewayModels(ctx, "")
+	if len(list.Data) != 1 || list.Data[0].ID != "meta.llama-3.3-70b-instruct" {
+		t.Fatalf("开关开:弃用模型应被过滤,got %+v", list.Data)
 	}
-	// 未知模型
-	if _, _, err := gw.Chat(ctx, aiwire.ChatRequest{Model: "no-such-model"}, ""); !errors.Is(err, ErrAiUnknownModel) {
-		t.Errorf("未知模型 err = %v", err)
+	if _, err := gw.pick(ctx, "meta.llama-old", "", "CHAT", map[uint]bool{}); !errors.Is(err, ErrAiUnknownModel) {
+		t.Errorf("开关开:弃用模型路由应不可达,err = %v", err)
 	}
+	// 持久化:重建 service 后开关仍生效
+	gw2 := NewAiGatewayService(gw.db, svc, &gatewayStubClient{})
+	if !gw2.FilterDeprecated() {
+		t.Error("重建后开关状态应保持开启")
 	}
-
-func TestAiChatNonRetryablePassThrough(t *testing.T) {
-	client := &gatewayStubClient{
-		fakeClient: &fakeClient{tenancy: oci.TenancyInfo{Name: "t"}},
-		chatErrs:   []error{stubServiceError{status: 400}},
+	if err := gw.SetFilterDeprecated(ctx, false); err != nil {
+		t.Fatalf("关闭开关: %v", err)
 	}
-	gw, svc := newTestGateway(t, client)
-	cfg := importAliveConfig(t, svc)
-	seedChannel(t, gw, cfg.ID, "eu-frankfurt-1", 1, 1)
-	seedChannel(t, gw, cfg.ID, "us-chicago-1", 1, 1)
-
-	_, meta, err := gw.Chat(context.Background(), aiwire.ChatRequest{Model: "meta.llama-3.3-70b-instruct"}, "")
-	if err == nil || meta.Retries != 0 || client.chatCalls != 1 {
-		t.Errorf("400 应直接透传不重试: err=%v retries=%d calls=%d", err, meta.Retries, client.chatCalls)
+	if list, _ = gw.GatewayModels(ctx, ""); len(list.Data) != 2 {
+		t.Errorf("开关关:弃用模型应恢复,got %d", len(list.Data))
 	}
 }

@@ -255,7 +319,7 @@ func TestMarkFailureBackoff(t *testing.T) {
 func TestAiGroupRouting(t *testing.T) {
 	client := &gatewayStubClient{
 		fakeClient:  &fakeClient{tenancy: oci.TenancyInfo{Name: "t"}},
-		chatResp:   &aiwire.ChatResponse{Model: "meta.llama-3.3-70b-instruct", Choices: []aiwire.Choice{{Message: aiwire.ChatMessage{Role: "assistant", Content: aiwire.NewTextContent("hi")}, FinishReason: "stop"}}},
+		passPayload: []byte(`{"id":"resp_1","usage":{"input_tokens":3,"output_tokens":1}}`),
 	}
 	gw, svc := newTestGateway(t, client)
 	cfg := importAliveConfig(t, svc)
@@ -264,17 +328,17 @@ func TestAiGroupRouting(t *testing.T) {
 	gw.db.Model(vip).Update("channel_group", "vip")
 	gw.db.Create(&model.AiModelCache{ChannelID: other.ID, ModelOcid: "ocid1..cr", Name: "cohere.command-r", Vendor: "cohere", SyncedAt: time.Now()})
 	ctx := context.Background()
-	req := aiwire.ChatRequest{Model: "meta.llama-3.3-70b-instruct", Messages: []aiwire.ChatMessage{{Role: "user", Content: aiwire.NewTextContent("你好")}}}
+	body := []byte(`{"model":"meta.llama-3.3-70b-instruct","input":"你好"}`)

 	// 分组密钥只落同分组渠道
 	for i := 0; i < 5; i++ {
-		_, meta, err := gw.Chat(ctx, req, "vip")
+		_, meta, err := gw.RespPassthrough(ctx, body, "meta.llama-3.3-70b-instruct", "vip")
 		if err != nil || meta.ChannelID != vip.ID {
 			t.Fatalf("vip 组第 %d 次落点 = %d, err %v, want %d", i, meta.ChannelID, err, vip.ID)
 		}
 	}
 	// 分组内无渠道 → 无可用渠道
-	if _, _, err := gw.Chat(ctx, req, "nope"); !errors.Is(err, ErrAiNoChannel) {
+	if _, _, err := gw.RespPassthrough(ctx, body, "meta.llama-3.3-70b-instruct", "nope"); !errors.Is(err, ErrAiNoChannel) {
 		t.Errorf("空分组 err = %v, want ErrAiNoChannel", err)
 	}
 	// 模型列表按分组过滤
@@ -287,12 +351,12 @@ func TestAiGroupRouting(t *testing.T) {
 		t.Errorf("不限组模型数 = %d, want 2", len(all.Data))
 	}
 	// 密钥分组落库
-	_, key, err := gw.CreateKey(ctx, "vip-key", "vip-secret-1234", "vip")
+	_, key, err := gw.CreateKey(ctx, "vip-key", "vip-secret-1234", "vip", nil)
 	if err != nil || key.Group != "vip" {
 		t.Fatalf("CreateKey group = %+v, %v", key, err)
 	}
 	empty := ""
-	_ = gw.UpdateKey(ctx, key.ID, "", nil, &empty)
+	_ = gw.UpdateKey(ctx, key.ID, "", nil, &empty, nil)
 	var fresh model.AiKey
 	gw.db.First(&fresh, key.ID)
 	if fresh.Group != "" {
@@ -400,6 +464,218 @@ func TestProbeCandidates(t *testing.T) {
 	}
 }

+func TestProbeCandidatesVendorDiversity(t *testing.T) {
+	// 部分区域单一厂商全为微调基座:候选须跨厂商分散,不能被 3 个 llama 占满前排
+	models := []oci.GenAiModel{
+		{Ocid: "l1", Name: "meta.llama-3-70b-instruct"},
+		{Ocid: "l2", Name: "meta.llama-3.1-405b-instruct"},
+		{Ocid: "l3", Name: "meta.llama-3.3-70b-instruct"},
+		{Ocid: "c1", Name: "cohere.command-a-03-2025"},
+		{Ocid: "g1", Name: "xai.grok-4"},
+	}
+	got := probeCandidates(models)
+	if len(got) != 5 || got[0].Name != "meta.llama-3-70b-instruct" {
+		t.Fatalf("上限内全量返回且最高分居首: %+v", got)
+	}
+	vendors := map[string]bool{}
+	for _, m := range got[:3] {
+		vendors[modelVendor(m)] = true
+	}
+	if len(vendors) != 3 {
+		t.Errorf("前 3 个候选应覆盖 3 个厂商: %+v", got)
+	}
+	// 超过上限时截断到 probeCandidateCap
+	var many []oci.GenAiModel
+	for i := 0; i < 12; i++ {
+		many = append(many, oci.GenAiModel{Ocid: fmt.Sprintf("m%d", i), Name: fmt.Sprintf("meta.llama-%d", i)})
+	}
+	if capped := probeCandidates(many); len(capped) != probeCandidateCap {
+		t.Errorf("候选应截断到 %d: got %d", probeCandidateCap, len(capped))
+	}
+}
+
+// entityNotFoundErr 模拟「实体不存在」404(模型在区域内无按需供给)。
+func entityNotFoundErr() stubServiceError {
+	return stubServiceError{status: 404,
+		msg: "Entity with key ocid1.generativeaimodel.oc1.eu-frankfurt-1.2flsfq not found"}
+}
+
+func TestProbeSkipsUnavailableModels(t *testing.T) {
+	// 首选 llama 不可按需调用(基座 400 / 实体 404)→ 换候选成功 → 渠道判可用;
+	// 模型全量入库不再自动标记,持久剔除交由用户手动拉黑
+	tests := []struct {
+		name string
+		bad  probeResult
+	}{
+		{"微调基座 400", probeResult{400, finetuneBaseErr()}},
+		{"实体不存在 404", probeResult{404, entityNotFoundErr()}},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			client := &gatewayStubClient{
+				fakeClient: &fakeClient{tenancy: oci.TenancyInfo{Name: "t"}},
+				models: []oci.GenAiModel{
+					{Ocid: "m1", Name: "meta.llama-3-70b-instruct", Vendor: "meta"},
+					{Ocid: "m2", Name: "meta.llama-3.1-70b-instruct", Vendor: "meta"},
+					{Ocid: "m3", Name: "cohere.command-a-03-2025", Vendor: "cohere"},
+				},
+				probeSeq: []probeResult{tt.bad, {200, nil}},
+			}
+			gw, svc := newTestGateway(t, client)
+			cfg := importAliveConfig(t, svc)
+			ctx := context.Background()
+
+			ch, err := gw.CreateChannel(ctx, ChannelInput{OciConfigID: cfg.ID, Region: "eu-frankfurt-1"})
+			if err != nil {
+				t.Fatalf("CreateChannel: %v", err)
+			}
+			probed, err := gw.ProbeChannel(ctx, ch.ID)
+			if err != nil || probed.ProbeStatus != "ok" {
+				t.Fatalf("坏候选后应换候选并判可用: %+v, %v", probed, err)
+			}
+			rows, _ := gw.channelModels(ctx, ch.ID)
+			if len(rows) != 3 {
+				t.Errorf("模型应全量入库, got %d", len(rows))
+			}
+		})
+	}
+}
+
+func TestProbeAuth404StillNoQuota(t *testing.T) {
+	// 鉴权类 404(NotAuthorizedOrNotFound)仍属租户级,直接定论 no_quota
+	client := &gatewayStubClient{
+		fakeClient: &fakeClient{tenancy: oci.TenancyInfo{Name: "t"}},
+		models:     []oci.GenAiModel{{Ocid: "m1", Name: "meta.llama-3.3-70b-instruct", Vendor: "meta"}},
+		probeCode:  404,
+		probeErr:   stubServiceError{status: 404, msg: "Authorization failed or requested resource not found."},
+	}
+	gw, svc := newTestGateway(t, client)
+	cfg := importAliveConfig(t, svc)
+	ctx := context.Background()
+
+	ch, _ := gw.CreateChannel(ctx, ChannelInput{OciConfigID: cfg.ID, Region: "eu-frankfurt-1"})
+	probed, _ := gw.ProbeChannel(ctx, ch.ID)
+	if probed.ProbeStatus != "no_quota" {
+		t.Errorf("鉴权 404 status = %q, want no_quota", probed.ProbeStatus)
+	}
+}
+
+func TestAiChatFinetuneSwitchesChannelWithoutPenalty(t *testing.T) {
+	// 微调基座 400 换渠道重试成功,且不计入熔断失败
+	client := &gatewayStubClient{
+		fakeClient:  &fakeClient{tenancy: oci.TenancyInfo{Name: "t"}},
+		passPayload: []byte(`{"id":"resp_1","usage":{"input_tokens":3,"output_tokens":1}}`),
+		passErrs:    []error{finetuneBaseErr()},
+	}
+	gw, svc := newTestGateway(t, client)
+	cfg := importAliveConfig(t, svc)
+	seedChannel(t, gw, cfg.ID, "eu-frankfurt-1", 1, 1)
+	seedChannel(t, gw, cfg.ID, "us-chicago-1", 1, 1)
+
+	resp, meta, err := gw.RespPassthrough(context.Background(), []byte(`{}`), "meta.llama-3.3-70b-instruct", "")
+	if err != nil || resp == nil {
+		t.Fatalf("RespPassthrough = %v, %v", resp, err)
+	}
+	if meta.Retries != 1 || client.passCalls != 2 {
+		t.Errorf("应换渠道重试一次: retries=%d calls=%d", meta.Retries, client.passCalls)
+	}
+	var chs []model.AiChannel
+	gw.db.Find(&chs)
+	for _, ch := range chs {
+		if ch.FailCount != 0 {
+			t.Errorf("微调基座 400 不应计入熔断: 渠道 %s failCount=%d", ch.Name, ch.FailCount)
+		}
+	}
+}
+
+func TestBlacklistLifecycle(t *testing.T) {
+	// 拉黑:全渠道同名缓存删除、列表与路由立即不可见;重复/空名被拒;
+	// 移出黑名单后重新同步恢复入池
+	client := &gatewayStubClient{
+		fakeClient: &fakeClient{tenancy: oci.TenancyInfo{Name: "t"}},
+		models:     []oci.GenAiModel{{Ocid: "ocid1..m-eu-frankfurt-1", Name: "meta.llama-3.3-70b-instruct", Vendor: "meta"}},
+	}
+	gw, svc := newTestGateway(t, client)
+	cfg := importAliveConfig(t, svc)
+	ch := seedChannel(t, gw, cfg.ID, "eu-frankfurt-1", 1, 1)
+	seedChannel(t, gw, cfg.ID, "us-chicago-1", 1, 1)
+	ctx := context.Background()
+
+	item, err := gw.AddBlacklist(ctx, " meta.llama-3.3-70b-instruct ")
+	if err != nil || item.Name != "meta.llama-3.3-70b-instruct" {
+		t.Fatalf("AddBlacklist = %+v, %v", item, err)
+	}
+	var left int64
+	gw.db.Model(&model.AiModelCache{}).Count(&left)
+	if left != 0 {
+		t.Errorf("拉黑应删除全部渠道的同名缓存, 剩 %d", left)
+	}
+	if _, _, err := gw.RespPassthrough(ctx, []byte(`{}`), "meta.llama-3.3-70b-instruct", ""); !errors.Is(err, ErrAiUnknownModel) {
+		t.Errorf("拉黑后路由应按未知模型拒绝: %v", err)
+	}
+	if _, err := gw.AddBlacklist(ctx, "meta.llama-3.3-70b-instruct"); err == nil {
+		t.Error("重复拉黑应被拒绝")
+	}
+	if _, err := gw.AddBlacklist(ctx, "  "); err == nil {
+		t.Error("空模型名应被拒绝")
+	}
+	rows, err := gw.Blacklist(ctx)
+	if err != nil || len(rows) != 1 {
+		t.Fatalf("Blacklist = %+v, %v", rows, err)
+	}
+	if err := gw.RemoveBlacklist(ctx, rows[0].ID); err != nil {
+		t.Fatalf("RemoveBlacklist: %v", err)
+	}
+	if err := gw.RemoveBlacklist(ctx, rows[0].ID); err == nil {
+		t.Error("移除不存在的条目应报错")
+	}
+	// 移出后重新同步:模型恢复入池
+	if _, err := gw.SyncModels(ctx, ch.ID); err != nil {
+		t.Fatalf("SyncModels: %v", err)
+	}
+	list, _ := gw.GatewayModels(ctx, "")
+	if len(list.Data) != 1 {
+		t.Errorf("移出黑名单并同步后应恢复入池: %+v", list.Data)
+	}
+}
+
+func TestSyncAndProbeFilterBlacklisted(t *testing.T) {
+	// 同步与探测都过滤黑名单模型:不入库、不进试调候选
+	client := &gatewayStubClient{
+		fakeClient: &fakeClient{tenancy: oci.TenancyInfo{Name: "t"}},
+		models: []oci.GenAiModel{
+			{Ocid: "m1", Name: "meta.llama-3-70b-instruct", Vendor: "meta"},
+			{Ocid: "m2", Name: "cohere.command-a-03-2025", Vendor: "cohere"},
+		},
+		probeCode: 200,
+	}
+	gw, svc := newTestGateway(t, client)
+	cfg := importAliveConfig(t, svc)
+	ctx := context.Background()
+	if _, err := gw.AddBlacklist(ctx, "meta.llama-3-70b-instruct"); err != nil {
+		t.Fatalf("AddBlacklist: %v", err)
+	}
+	ch, err := gw.CreateChannel(ctx, ChannelInput{OciConfigID: cfg.ID, Region: "eu-frankfurt-1"})
+	if err != nil {
+		t.Fatalf("CreateChannel: %v", err)
+	}
+	probed, err := gw.ProbeChannel(ctx, ch.ID)
+	if err != nil || probed.ProbeStatus != "ok" {
+		t.Fatalf("ProbeChannel = %+v, %v", probed, err)
+	}
+	rows, _ := gw.channelModels(ctx, ch.ID)
+	if len(rows) != 1 || rows[0].Name != "cohere.command-a-03-2025" {
+		t.Errorf("探测同步应过滤黑名单模型: %+v", rows)
+	}
+	if len(client.probedNames) != 1 || client.probedNames[0] != "cohere.command-a-03-2025" {
+		t.Errorf("试调候选不应包含黑名单模型: %v", client.probedNames)
+	}
+	models, err := gw.SyncModels(ctx, ch.ID)
+	if err != nil || len(models) != 1 || models[0].Name != "cohere.command-a-03-2025" {
+		t.Errorf("SyncModels 应过滤黑名单模型: %+v, %v", models, err)
+	}
+}
+
 func TestDeprecatingModels(t *testing.T) {
 	gw, _ := newTestGateway(t, &gatewayStubClient{fakeClient: &fakeClient{tenancy: oci.TenancyInfo{Name: "t"}}})
 	ctx := context.Background()
@@ -460,7 +736,7 @@ func TestAiEmbeddings(t *testing.T) {
 		t.Errorf("chat 模型走 embeddings err = %v, want ErrAiUnknownModel", err)
 	}
 	// embedding 模型名打 chat:同样未知模型
-	if _, _, err := gw.Chat(ctx, aiwire.ChatRequest{Model: "cohere.embed-v4.0", Messages: []aiwire.ChatMessage{{Role: "user", Content: aiwire.NewTextContent("x")}}}, ""); !errors.Is(err, ErrAiUnknownModel) {
+	if _, _, err := gw.RespPassthrough(ctx, []byte(`{}`), "cohere.embed-v4.0", ""); !errors.Is(err, ErrAiUnknownModel) {
 		t.Errorf("embedding 模型走 chat err = %v, want ErrAiUnknownModel", err)
 	}
 }
@@ -471,7 +747,7 @@ func TestAiContentLogSwitch(t *testing.T) {
 		t.Fatalf("migrate content log: %v", err)
 	}
 	ctx := context.Background()
-	_, key, err := gw.CreateKey(ctx, "k1", "content-key-1234", "")
+	_, key, err := gw.CreateKey(ctx, "k1", "content-key-1234", "", nil)
 	if err != nil {
 		t.Fatalf("CreateKey: %v", err)
 	}
@@ -488,6 +764,7 @@ func TestAiContentLogSwitch(t *testing.T) {
 		t.Error("超过 7 天上限应被拒绝")
 	}
 	// 写入与截断(带调用日志关联)
+	mustCreate(t, gw.db, &model.AiCallLog{ID: 42, KeyID: key.ID, ChannelID: 0})
 	gw.LogContent(model.AiContentLog{CallLogID: 42, KeyID: key.ID, KeyName: "k1", Endpoint: "openai", Model: "m", RequestBody: strings.Repeat("x", 70*1024)})
 	rows, total, err := gw.ContentLogs(ctx, key.ID, 0, 1, 20)
 	if err != nil || total != 1 || len(rows[0].RequestBody) != 64*1024 {
@@ -506,3 +783,131 @@ func TestAiContentLogSwitch(t *testing.T) {
 		t.Fatalf("关闭失败: %+v, %v", fresh.ContentLogUntil, err)
 	}
 }
+
+func TestRespPassthroughSwitchesChannel(t *testing.T) {
+	client := &gatewayStubClient{
+		fakeClient:  &fakeClient{tenancy: oci.TenancyInfo{Name: "t"}},
+		passPayload: []byte(`{"id":"resp_1","usage":{"input_tokens":9,"output_tokens":3}}`),
+		passErrs:    []error{stubServiceError{status: 429}},
+	}
+	gw, svc := newTestGateway(t, client)
+	cfg := importAliveConfig(t, svc)
+	seedChannel(t, gw, cfg.ID, "eu-frankfurt-1", 1, 1)
+	seedChannel(t, gw, cfg.ID, "us-chicago-1", 1, 1)
+
+	payload, meta, err := gw.RespPassthrough(context.Background(), []byte(`{}`), "meta.llama-3.3-70b-instruct", "")
+	if err != nil || len(payload) == 0 {
+		t.Fatalf("RespPassthrough = %v, %v", payload, err)
+	}
+	if meta.Retries != 1 || client.passCalls != 2 {
+		t.Errorf("应换渠道重试一次: retries=%d calls=%d", meta.Retries, client.passCalls)
+	}
+	if len(client.passRegions) == 2 && client.passRegions[0] == client.passRegions[1] {
+		t.Errorf("重试未换渠道: %v", client.passRegions)
+	}
+}
+
+func TestRespPassthroughNonRetryable(t *testing.T) {
+	client := &gatewayStubClient{
+		fakeClient: &fakeClient{tenancy: oci.TenancyInfo{Name: "t"}},
+		passErrs:   []error{stubServiceError{status: 400}},
+	}
+	gw, svc := newTestGateway(t, client)
+	cfg := importAliveConfig(t, svc)
+	seedChannel(t, gw, cfg.ID, "eu-frankfurt-1", 1, 1)
+	seedChannel(t, gw, cfg.ID, "us-chicago-1", 1, 1)
+
+	_, meta, err := gw.RespPassthrough(context.Background(), []byte(`{}`), "meta.llama-3.3-70b-instruct", "")
+	if err == nil || meta.Retries != 0 || client.passCalls != 1 {
+		t.Errorf("400 应直接透传不重试: err=%v retries=%d calls=%d", err, meta.Retries, client.passCalls)
+	}
+}
+
+func TestNormalizeKeyModels(t *testing.T) {
+	tests := []struct {
+		name string
+		in   []string
+		want []string
+	}{
+		{"nil 输入", nil, nil},
+		{"全空串", []string{"", "  "}, nil},
+		{"trim 与保序去重", []string{" a ", "b", "a", "", "b "}, []string{"a", "b"}},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			got := normalizeKeyModels(tt.in)
+			if len(got) != len(tt.want) {
+				t.Fatalf("normalizeKeyModels(%v) = %v, want %v", tt.in, got, tt.want)
+			}
+			for i := range tt.want {
+				if got[i] != tt.want[i] {
+					t.Fatalf("normalizeKeyModels(%v) = %v, want %v", tt.in, got, tt.want)
+				}
+			}
+		})
+	}
+}
+
+// TestAiKeyModelsPersistence 钉住 serializer:json 字段经 Create 与 Updates(map) 两条路径的往返。
+func TestAiKeyModelsPersistence(t *testing.T) {
+	gw, _ := newTestGateway(t, &gatewayStubClient{fakeClient: &fakeClient{tenancy: oci.TenancyInfo{Name: "t"}}})
+	ctx := context.Background()
+
+	_, key, err := gw.CreateKey(ctx, "m-key", "model-key-1234", "", []string{" x-model ", "x-model", "y-model", ""})
+	if err != nil {
+		t.Fatalf("CreateKey: %v", err)
+	}
+	fresh, err := gw.VerifyKey(ctx, "model-key-1234")
+	if err != nil || len(fresh.Models) != 2 || fresh.Models[0] != "x-model" || fresh.Models[1] != "y-model" {
+		t.Fatalf("创建后读回 models = %v, %v", fresh.Models, err)
+	}
+	set := []string{"z-model"}
+	if err := gw.UpdateKey(ctx, key.ID, "", nil, nil, &set); err != nil {
+		t.Fatalf("UpdateKey 覆盖: %v", err)
+	}
+	fresh, _ = gw.VerifyKey(ctx, "model-key-1234")
+	if len(fresh.Models) != 1 || fresh.Models[0] != "z-model" {
+		t.Fatalf("覆盖后 models = %v", fresh.Models)
+	}
+	if err := gw.UpdateKey(ctx, key.ID, "renamed", nil, nil, nil); err != nil {
+		t.Fatalf("UpdateKey 不传 models: %v", err)
+	}
+	fresh, _ = gw.VerifyKey(ctx, "model-key-1234")
+	if len(fresh.Models) != 1 {
+		t.Fatalf("未传 models 却被改动: %v", fresh.Models)
+	}
+	empty := []string{}
+	if err := gw.UpdateKey(ctx, key.ID, "", nil, nil, &empty); err != nil {
+		t.Fatalf("UpdateKey 清空: %v", err)
+	}
+	fresh, _ = gw.VerifyKey(ctx, "model-key-1234")
+	if len(fresh.Models) != 0 {
+		t.Fatalf("清空后 models = %v, want 空", fresh.Models)
+	}
+}
+
+// TestRespPassthroughStreamSwitchesChannel 断言流式直通建立失败按 switchable 换渠道。
+func TestRespPassthroughStreamSwitchesChannel(t *testing.T) {
+	client := &gatewayStubClient{
+		fakeClient:  &fakeClient{tenancy: oci.TenancyInfo{Name: "t"}},
+		passPayload: []byte("data: {\"type\":\"response.completed\"}\n\n"),
+		passErrs:    []error{stubServiceError{status: 503}},
+	}
+	gw, svc := newTestGateway(t, client)
+	cfg := importAliveConfig(t, svc)
+	seedChannel(t, gw, cfg.ID, "eu-frankfurt-1", 1, 1)
+	seedChannel(t, gw, cfg.ID, "us-chicago-1", 1, 1)
+
+	stream, meta, err := gw.RespPassthroughStream(context.Background(), []byte(`{}`), "meta.llama-3.3-70b-instruct", "")
+	if err != nil || stream == nil {
+		t.Fatalf("RespPassthroughStream = %v", err)
+	}
+	defer stream.Close()
+	if meta.Retries != 1 || client.passCalls != 2 {
+		t.Errorf("应换渠道重试一次: retries=%d calls=%d", meta.Retries, client.passCalls)
+	}
+	payload, _ := io.ReadAll(stream)
+	if !bytes.Contains(payload, []byte("response.completed")) {
+		t.Errorf("流内容未透传: %s", payload)
+	}
+}
@@ -8,38 +8,8 @@ import (
 	"oci-portal/internal/aiwire"
 )

-// ResponsesToIR 把 Responses API 请求转换为网关 IR(OpenAI Chat 形态)。
-// 有状态特性与内置工具不支持,直接报错(API 层 400)。
-func ResponsesToIR(req aiwire.RespRequest) (aiwire.ChatRequest, error) {
-	if err := respRejectUnsupported(req); err != nil {
-		return aiwire.ChatRequest{}, err
-	}
-	ir := aiwire.ChatRequest{
-		Model:       req.Model,
-		MaxTokens:   req.MaxOutputTokens,
-		Temperature: req.Temperature,
-		TopP:        req.TopP,
-		Stream:      req.Stream,
-		ToolChoice:  respToolChoice(req.ToolChoice),
-	}
-	if req.Reasoning != nil {
-		ir.ReasoningEffort = req.Reasoning.Effort
-	}
-	if req.Instructions != "" {
-		ir.Messages = append(ir.Messages, aiwire.ChatMessage{Role: "system", Content: aiwire.NewTextContent(req.Instructions)})
-	}
-	msgs, err := respInputToMessages(req.Input)
-	if err != nil {
-		return aiwire.ChatRequest{}, err
-	}
-	ir.Messages = append(ir.Messages, msgs...)
-	ir.Tools = respTools(req.Tools)
-	ir.ResponseFormat = respFormat(req.Text)
-	return ir, nil
-}
-
-// respRejectUnsupported 拒绝无状态网关无法承接的请求特性。
-func respRejectUnsupported(req aiwire.RespRequest) error {
+// respRejectStateful 拒绝有状态特性(网关无状态)。
+func respRejectStateful(req aiwire.RespRequest) error {
 	if req.PreviousResponseID != "" {
 		return fmt.Errorf("previous_response_id 不支持:网关不保存历史响应,请在 input 中自带完整上下文(store:false 模式)")
 	}
@@ -49,195 +19,97 @@ func respRejectUnsupported(req aiwire.RespRequest) error {
 	if req.Background != nil && *req.Background {
 		return fmt.Errorf("background 模式不支持")
 	}
+	return nil
+}
+
+// RespPassthroughValidate 校验直通请求:模型必填,有状态特性不支持,工具类型
+// 只放行 function 与 Oracle 文档化的服务端工具(web_search / x_search /
+// code_interpreter / mcp)。
+func RespPassthroughValidate(req aiwire.RespRequest) error {
+	if strings.TrimSpace(req.Model) == "" {
+		return fmt.Errorf("model 不能为空")
+	}
+	if err := respRejectStateful(req); err != nil {
+		return err
+	}
 	for _, t := range req.Tools {
-		if t.Type != "function" {
-			return fmt.Errorf("不支持的工具类型 %q:仅支持 function 工具", t.Type)
-		}
-	}
-	return nil
-}
-
-// respInputToMessages 把 input(string 或 item 数组)展开为 IR 消息序列。
-func respInputToMessages(in aiwire.RespInput) ([]aiwire.ChatMessage, error) {
-	if !in.IsArray {
-		if strings.TrimSpace(in.Text) == "" {
-			return nil, fmt.Errorf("input 不能为空")
-		}
-		return []aiwire.ChatMessage{{Role: "user", Content: aiwire.NewTextContent(in.Text)}}, nil
-	}
-	var msgs []aiwire.ChatMessage
-	for _, it := range in.Items {
-		out, err := respItemToMessages(it, msgs)
-		if err != nil {
-			return nil, err
-		}
-		msgs = out
-	}
-	return msgs, nil
-}
-
-// respItemToMessages 追加一个 input item;连续 function_call 合并进同一 assistant 消息。
-func respItemToMessages(it aiwire.RespItem, msgs []aiwire.ChatMessage) ([]aiwire.ChatMessage, error) {
-	switch it.Type {
-	case "", "message":
-		if it.Content.HasUnsupported() {
-			return nil, ErrAiUnsupportedBlock
-		}
-		role := respRole(it.Role)
-		return append(msgs, aiwire.ChatMessage{Role: role, Content: respContentToIR(it.Content)}), nil
-	case "function_call":
-		call := aiwire.ToolCall{ID: it.CallID, Type: "function",
-			Function: aiwire.FunctionCall{Name: it.Name, Arguments: it.Arguments}}
-		if n := len(msgs); n > 0 && msgs[n-1].Role == "assistant" && len(msgs[n-1].ToolCalls) > 0 {
-			msgs[n-1].ToolCalls = append(msgs[n-1].ToolCalls, call)
-			return msgs, nil
-		}
-		return append(msgs, aiwire.ChatMessage{Role: "assistant", ToolCalls: []aiwire.ToolCall{call}}), nil
-	case "function_call_output":
-		return append(msgs, aiwire.ChatMessage{Role: "tool", ToolCallID: it.CallID,
-			Content: aiwire.NewTextContent(it.OutputText())}), nil
-	case "reasoning":
-		return msgs, nil // 推理块不回灌上游
+		switch t.Type {
+		case "function", "web_search", "x_search", "code_interpreter", "mcp":
 		default:
-		return nil, fmt.Errorf("不支持的 input item 类型 %q", it.Type)
+			return fmt.Errorf("不支持的工具类型 %q:服务端工具仅支持 web_search / x_search / code_interpreter / mcp", t.Type)
 		}
 	}
-
-// respRole 归一 item 角色;developer 视为 system,缺省 user。
-func respRole(role string) string {
-	switch role {
-	case "system", "developer":
-		return "system"
-	case "assistant":
-		return "assistant"
-	default:
-		return "user"
-	}
-}
-
-// respContentToIR 把 Responses 内容转 IR;纯文本拍平,含图片时保留部件顺序。
-func respContentToIR(c aiwire.RespContent) aiwire.Content {
-	var parts []aiwire.ContentPart
-	hasImage := false
-	for _, p := range c.Parts {
-		switch p.Type {
-		case "input_image":
-			parts = append(parts, aiwire.ContentPart{Type: "image_url",
-				ImageURL: &aiwire.ImageURL{URL: p.ImageURL, Detail: p.Detail}})
-			hasImage = true
-		default:
-			if p.Text != "" {
-				parts = append(parts, aiwire.ContentPart{Type: "text", Text: p.Text})
-			}
-		}
-	}
-	if !c.IsArray || !hasImage {
-		return aiwire.NewTextContent(c.JoinText())
-	}
-	return aiwire.NewPartsContent(parts)
-}
-
-// respTools 把扁平工具定义还原为嵌套 Chat 形态;strict 无 OCI 对应,忽略。
-func respTools(tools []aiwire.RespTool) []aiwire.Tool {
-	if len(tools) == 0 {
 	return nil
 }
-	out := make([]aiwire.Tool, 0, len(tools))
-	for _, t := range tools {
-		out = append(out, aiwire.Tool{Type: "function",
-			Function: aiwire.FunctionDef{Name: t.Name, Description: t.Description, Parameters: t.Parameters}})
+
+// RespPassthroughBody 以原始请求体为基构造上游 body:强制 store:false(禁上游
+// 存态),stream 原样保留(流式直通);用 json.Number 保真未知字段与数值。
+func RespPassthroughBody(raw []byte) ([]byte, error) {
+	dec := json.NewDecoder(strings.NewReader(string(raw)))
+	dec.UseNumber()
+	var body map[string]any
+	if err := dec.Decode(&body); err != nil {
+		return nil, fmt.Errorf("解析请求体: %w", err)
 	}
-	return out
+	body["store"] = false
+	return json.Marshal(body)
 }

-// respToolChoice 转换 tool_choice:字符串直通;{type:function,name} 转嵌套;
-// allowed_tools 等对象形态降级 auto。
-func respToolChoice(raw json.RawMessage) json.RawMessage {
-	if len(raw) == 0 {
+// RespPassthroughUsage 从直通响应提取用量;缺失时返回 nil(日志记零)。
+func RespPassthroughUsage(payload []byte) *aiwire.Usage {
+	var root struct {
+		Usage *aiwire.RespUsage `json:"usage"`
+	}
+	if json.Unmarshal(payload, &root) != nil {
 		return nil
 	}
-	t := strings.TrimSpace(string(raw))
-	if strings.HasPrefix(t, "\"") {
-		return raw
-	}
-	var obj struct {
-		Type string `json:"type"`
-		Name string `json:"name"`
-	}
-	if err := json.Unmarshal(raw, &obj); err != nil || obj.Type != "function" || obj.Name == "" {
-		return json.RawMessage(`"auto"`)
-	}
-	out, _ := json.Marshal(map[string]any{"type": "function", "function": map[string]string{"name": obj.Name}})
-	return out
+	return usageFromResp(root.Usage)
 }

-// respFormat 把 text.format 转成 response_format;text 形态无需显式指定。
-func respFormat(text *aiwire.RespText) *aiwire.ResponseFormat {
-	if text == nil || text.Format == nil {
-		return nil
-	}
-	switch text.Format.Type {
-	case "json_object":
-		return &aiwire.ResponseFormat{Type: "json_object"}
-	case "json_schema":
-		spec, _ := json.Marshal(map[string]any{
-			"name": text.Format.Name, "schema": json.RawMessage(text.Format.Schema), "strict": text.Format.Strict,
-		})
-		return &aiwire.ResponseFormat{Type: "json_schema", JSONSchema: spec}
-	default:
-		return nil
-	}
-}
-
-// IRRespToResponses 把 IR 非流式响应装配为 Response 对象。
-func IRRespToResponses(resp *aiwire.ChatResponse, id string, created int64) aiwire.Response {
-	out := aiwire.Response{ID: id, Object: "response", CreatedAt: created,
-		Status: "completed", Model: resp.Model, Output: []aiwire.RespOutItem{}, Store: false}
-	if len(resp.Choices) == 0 {
-		return out
-	}
-	choice := resp.Choices[0]
-	if text := choice.Message.Content.JoinText(); text != "" {
-		out.Output = append(out.Output, respMessageItem(id, 0, text))
-	}
-	for i, tc := range choice.Message.ToolCalls {
-		out.Output = append(out.Output, aiwire.RespOutItem{Type: "function_call",
-			ID: respItemID(id, "fc", len(out.Output)+i), Status: "completed",
-			CallID: tc.ID, Name: tc.Function.Name, Arguments: respArgs(tc.Function.Arguments)})
-	}
-	if choice.FinishReason == "length" {
-		out.Status = "incomplete"
-		out.IncompleteDetails = &aiwire.RespIncomplete{Reason: "max_output_tokens"}
-	}
-	out.Usage = respUsage(resp.Usage)
-	return out
-}
-
-func respMessageItem(respID string, idx int, text string) aiwire.RespOutItem {
-	return aiwire.RespOutItem{Type: "message", ID: respItemID(respID, "msg", idx),
-		Status: "completed", Role: "assistant",
-		Content: []aiwire.RespOutPart{{Type: "output_text", Text: text, Annotations: []any{}}}}
-}
-
-// respItemID 从响应 ID 派生确定性 item ID。
-func respItemID(respID, kind string, idx int) string {
-	return fmt.Sprintf("%s_%s_%d", kind, strings.TrimPrefix(respID, "resp_"), idx)
-}
-
-// respArgs 保证 arguments 是合法 JSON 字符串(空实参回退 {})。
-func respArgs(args string) string {
-	if strings.TrimSpace(args) == "" {
-		return "{}"
-	}
-	return args
-}
-
-// respUsage 把 IR 用量改写为 Responses 命名口径。
-func respUsage(u *aiwire.Usage) *aiwire.RespUsage {
+// usageFromResp 换算 Responses usage 为 OpenAI 口径(缓存命中透传);nil 原样返回。
+func usageFromResp(u *aiwire.RespUsage) *aiwire.Usage {
 	if u == nil {
 		return nil
 	}
-	return &aiwire.RespUsage{InputTokens: u.PromptTokens, OutputTokens: u.CompletionTokens,
-		TotalTokens:        u.TotalTokens,
-		InputTokensDetails: aiwire.RespInDetails{CachedTokens: u.CachedTokens()}}
+	usage := &aiwire.Usage{PromptTokens: u.InputTokens,
+		CompletionTokens: u.OutputTokens, TotalTokens: u.TotalTokens}
+	if cached := u.InputTokensDetails.CachedTokens; cached > 0 {
+		usage.PromptTokensDetails = &aiwire.PromptTokensDetails{CachedTokens: cached}
+	}
+	return usage
+}
+
+// RespStreamCompletedUsage 从一行 SSE data JSON 中提取 response.completed 事件的
+// usage;非 completed 事件或解析失败返回 nil。流式直通逐行喂入,最后一次非 nil 生效。
+func RespStreamCompletedUsage(data []byte) *aiwire.Usage {
+	var ev struct {
+		Type     string          `json:"type"`
+		Response json.RawMessage `json:"response"`
+	}
+	if json.Unmarshal(data, &ev) != nil || ev.Type != "response.completed" || len(ev.Response) == 0 {
+		return nil
+	}
+	return RespPassthroughUsage(ev.Response)
+}
+
+// RespStreamErrorMsg 从一行 SSE data JSON 中提取 error / response.failed 事件的
+// 错误消息;非错误事件返回空串。流式直通据此把上游错误写入调用日志。
+func RespStreamErrorMsg(data []byte) string {
+	var ev respStreamEvent
+	if json.Unmarshal(data, &ev) != nil {
+		return ""
+	}
+	switch ev.Type {
+	case "error":
+		if ev.Message != "" {
+			return ev.Message
+		}
+		return "上游返回错误事件 error"
+	case "response.failed":
+		if m := respErrorMsg(ev.Response); m != "" {
+			return m
+		}
+		return "上游返回错误事件 response.failed"
+	}
+	return ""
 }
@@ -1,199 +0,0 @@
-package service
-
-import (
-	"strings"
-
-	"oci-portal/internal/aiwire"
-)
-
-// RespEvent 是一条 Responses SSE 语义事件(event 名 + data 载荷)。
-type RespEvent struct {
-	Event string
-	Data  map[string]any
-}
-
-// RespStream 把 IR chunk 流聚合为 Responses 语义事件序列。
-// 与 AnthStream 同构,但终态事件须携带完整 Response 快照,故全程缓冲文本与实参。
-type RespStream struct {
-	id      string
-	model   string
-	created int64
-	seq     int
-	started bool
-	items   []aiwire.RespOutItem
-	msgOpen bool
-	text    strings.Builder
-	tool    aiwire.RespOutItem
-	toolIdx int
-	tOpen   bool
-	args    strings.Builder
-	finish  string
-	usage   *aiwire.Usage
-}
-
-// NewRespStream 构造状态机;id 形如 resp_*,created 为响应时间戳。
-func NewRespStream(id, model string, created int64) *RespStream {
-	return &RespStream{id: id, model: model, created: created}
-}
-
-// Feed 消化一个上游 chunk,返回应立即下发的事件。
-func (s *RespStream) Feed(chunk aiwire.ChatChunk) []RespEvent {
-	var evs []RespEvent
-	if !s.started {
-		s.started = true
-		evs = append(evs, s.respEvent("response.created", "in_progress"),
-			s.respEvent("response.in_progress", "in_progress"))
-	}
-	if chunk.Usage != nil {
-		s.usage = chunk.Usage
-	}
-	if len(chunk.Choices) == 0 {
-		return evs
-	}
-	choice := chunk.Choices[0]
-	if choice.Delta.Content != "" {
-		evs = append(evs, s.feedText(choice.Delta.Content)...)
-	}
-	for _, tc := range choice.Delta.ToolCalls {
-		evs = append(evs, s.feedTool(tc)...)
-	}
-	if choice.FinishReason != nil {
-		s.finish = *choice.FinishReason
-	}
-	return evs
-}
-
-// Finish 关闭未闭合的块并产出终态事件(带完整 Response 快照)。
-func (s *RespStream) Finish() []RespEvent {
-	var evs []RespEvent
-	if !s.started {
-		s.started = true
-		evs = append(evs, s.respEvent("response.created", "in_progress"))
-	}
-	evs = append(evs, s.closeMsg()...)
-	evs = append(evs, s.closeTool()...)
-	status := "completed"
-	if s.finish == "length" {
-		status = "incomplete"
-	}
-	return append(evs, s.respEvent("response."+status, status))
-}
-
-// Usage 返回聚合到的用量(可能为 nil),供调用日志。
-func (s *RespStream) Usage() *aiwire.Usage { return s.usage }
-
-// ev 构造带自增 sequence_number 的事件。
-func (s *RespStream) ev(typ string, kv map[string]any) RespEvent {
-	s.seq++
-	data := map[string]any{"type": typ, "sequence_number": s.seq}
-	for k, v := range kv {
-		data[k] = v
-	}
-	return RespEvent{Event: typ, Data: data}
-}
-
-// respEvent 构造携带 Response 快照的生命周期事件。
-func (s *RespStream) respEvent(typ, status string) RespEvent {
-	return s.ev(typ, map[string]any{"response": s.snapshot(status)})
-}
-
-// snapshot 组装当前累计状态的 Response 对象。
-func (s *RespStream) snapshot(status string) aiwire.Response {
-	resp := aiwire.Response{ID: s.id, Object: "response", CreatedAt: s.created,
-		Status: status, Model: s.model, Store: false,
-		Output: append([]aiwire.RespOutItem{}, s.items...)}
-	resp.Usage = respUsage(s.usage)
-	if status == "incomplete" {
-		resp.IncompleteDetails = &aiwire.RespIncomplete{Reason: "max_output_tokens"}
-	}
-	return resp
-}
-
-// feedText 处理文本增量:必要时开 message item 与 content part。
-func (s *RespStream) feedText(delta string) []RespEvent {
-	evs := s.closeTool()
-	if !s.msgOpen {
-		s.msgOpen = true
-		s.text.Reset()
-		item := aiwire.RespOutItem{Type: "message", ID: s.itemID("msg"), Status: "in_progress",
-			Role: "assistant", Content: []aiwire.RespOutPart{}}
-		evs = append(evs, s.ev("response.output_item.added", map[string]any{
-			"output_index": len(s.items), "item": item}))
-		evs = append(evs, s.ev("response.content_part.added", map[string]any{
-			"item_id": item.ID, "output_index": len(s.items), "content_index": 0,
-			"part": aiwire.RespOutPart{Type: "output_text", Text: "", Annotations: []any{}}}))
-	}
-	s.text.WriteString(delta)
-	return append(evs, s.ev("response.output_text.delta", map[string]any{
-		"item_id": s.itemID("msg"), "output_index": len(s.items), "content_index": 0, "delta": delta}))
-}
-
-// closeMsg 闭合当前 message item(text done → part done → item done)。
-func (s *RespStream) closeMsg() []RespEvent {
-	if !s.msgOpen {
-		return nil
-	}
-	s.msgOpen = false
-	id, idx, full := s.itemID("msg"), len(s.items), s.text.String()
-	part := aiwire.RespOutPart{Type: "output_text", Text: full, Annotations: []any{}}
-	item := aiwire.RespOutItem{Type: "message", ID: id, Status: "completed",
-		Role: "assistant", Content: []aiwire.RespOutPart{part}}
-	evs := []RespEvent{
-		s.ev("response.output_text.done", map[string]any{"item_id": id, "output_index": idx, "content_index": 0, "text": full}),
-		s.ev("response.content_part.done", map[string]any{"item_id": id, "output_index": idx, "content_index": 0, "part": part}),
-		s.ev("response.output_item.done", map[string]any{"output_index": idx, "item": item}),
-	}
-	s.items = append(s.items, item)
-	return evs
-}
-
-// feedTool 处理工具调用增量:新 Index 先闭合旧调用再开新 item。
-func (s *RespStream) feedTool(tc aiwire.ToolCallDelta) []RespEvent {
-	evs := s.closeMsg()
-	if s.tOpen && tc.Index != s.toolIdx {
-		evs = append(evs, s.closeTool()...)
-	}
-	if !s.tOpen {
-		s.tOpen, s.toolIdx = true, tc.Index
-		s.args.Reset()
-		s.tool = aiwire.RespOutItem{Type: "function_call", ID: s.itemID("fc"), Status: "in_progress",
-			CallID: tc.ID, Name: tc.Function.Name}
-		evs = append(evs, s.ev("response.output_item.added", map[string]any{
-			"output_index": len(s.items), "item": s.tool}))
-	}
-	if tc.ID != "" && s.tool.CallID == "" {
-		s.tool.CallID = tc.ID
-	}
-	if tc.Function.Name != "" && s.tool.Name == "" {
-		s.tool.Name = tc.Function.Name
-	}
-	if tc.Function.Arguments == "" {
-		return evs
-	}
-	s.args.WriteString(tc.Function.Arguments)
-	return append(evs, s.ev("response.function_call_arguments.delta", map[string]any{
-		"item_id": s.tool.ID, "output_index": len(s.items), "delta": tc.Function.Arguments}))
-}
-
-// closeTool 闭合当前 function_call item(arguments done → item done)。
-func (s *RespStream) closeTool() []RespEvent {
-	if !s.tOpen {
-		return nil
-	}
-	s.tOpen = false
-	s.tool.Arguments = respArgs(s.args.String())
-	s.tool.Status = "completed"
-	idx := len(s.items)
-	evs := []RespEvent{
-		s.ev("response.function_call_arguments.done", map[string]any{
-			"item_id": s.tool.ID, "output_index": idx, "arguments": s.tool.Arguments}),
-		s.ev("response.output_item.done", map[string]any{"output_index": idx, "item": s.tool}),
-	}
-	s.items = append(s.items, s.tool)
-	return evs
-}
-
-// itemID 按当前 output 序号派生确定性 item ID。
-func (s *RespStream) itemID(kind string) string {
-	return respItemID(s.id, kind, len(s.items))
-}
@@ -17,47 +17,6 @@ func respReq(t *testing.T, raw string) aiwire.RespRequest {
 	return req
 }

-func TestResponsesToIR(t *testing.T) {
-	req := respReq(t, `{
-		"model": "meta.llama-3.3-70b-instruct",
-		"instructions": "你是助手",
-		"max_output_tokens": 128,
-		"input": [
-			{"role": "user", "content": "查天气"},
-			{"type": "function_call", "call_id": "call_1", "name": "get_weather", "arguments": "{\"city\":\"上海\"}"},
-			{"type": "function_call_output", "call_id": "call_1", "output": "{\"temp\":31}"},
-			{"type": "message", "role": "assistant", "content": [{"type": "output_text", "text": "31 度"}]}
-		],
-		"tools": [{"type": "function", "name": "get_weather", "description": "查天气", "parameters": {"type": "object"}, "strict": true}],
-		"tool_choice": {"type": "function", "name": "get_weather"},
-		"text": {"format": {"type": "json_schema", "name": "out", "schema": {"type": "object"}}}
-	}`)
-	ir, err := ResponsesToIR(req)
-	if err != nil {
-		t.Fatalf("ResponsesToIR: %v", err)
-	}
-	roles := make([]string, 0, len(ir.Messages))
-	for _, m := range ir.Messages {
-		roles = append(roles, m.Role)
-	}
-	want := []string{"system", "user", "assistant", "tool", "assistant"}
-	if strings.Join(roles, ",") != strings.Join(want, ",") {
-		t.Errorf("roles = %v, want %v", roles, want)
-	}
-	if ir.Messages[2].ToolCalls[0].ID != "call_1" || ir.Messages[3].ToolCallID != "call_1" {
-		t.Errorf("tool call 链接错误: %+v", ir.Messages)
-	}
-	if deref(ir.MaxTokens) != 128 || len(ir.Tools) != 1 || ir.Tools[0].Function.Name != "get_weather" {
-		t.Errorf("参数映射错误: max=%v tools=%+v", ir.MaxTokens, ir.Tools)
-	}
-	if !strings.Contains(string(ir.ToolChoice), `"function"`) || !strings.Contains(string(ir.ToolChoice), "get_weather") {
-		t.Errorf("tool_choice = %s", ir.ToolChoice)
-	}
-	if ir.ResponseFormat == nil || ir.ResponseFormat.Type != "json_schema" {
-		t.Errorf("response_format = %+v", ir.ResponseFormat)
-	}
-}
-
 func deref(p *int) int {
 	if p == nil {
 		return 0
@@ -65,112 +24,105 @@ func deref(p *int) int {
 	return *p
 }

-func TestResponsesToIRRejects(t *testing.T) {
-	cases := []struct{ name, raw string }{
-		{"previous_response_id", `{"model":"m","input":"hi","previous_response_id":"resp_x"}`},
-		{"conversation", `{"model":"m","input":"hi","conversation":{"id":"conv_1"}}`},
-		{"background", `{"model":"m","input":"hi","background":true}`},
-		{"builtin tool", `{"model":"m","input":"hi","tools":[{"type":"web_search"}]}`},
-		{"file_id image", `{"model":"m","input":[{"role":"user","content":[{"type":"input_image","file_id":"file_1"}]}]}`},
-		{"unknown item", `{"model":"m","input":[{"type":"item_reference","id":"x"}]}`},
+func TestRespPassthroughValidate(t *testing.T) {
+	prev := "resp_1"
+	bg := true
+	tests := []struct {
+		name    string
+		req     aiwire.RespRequest
+		wantErr bool
+	}{
+		{"合法", aiwire.RespRequest{Model: "m", Tools: []aiwire.RespTool{{Type: "web_search"}}}, false},
+		{"混用function", aiwire.RespRequest{Model: "m", Tools: []aiwire.RespTool{{Type: "web_search"}, {Type: "function", Name: "f"}}}, false},
+		{"缺model", aiwire.RespRequest{Tools: []aiwire.RespTool{{Type: "web_search"}}}, true},
+		{"服务端工具流式放行", aiwire.RespRequest{Model: "m", Stream: true, Tools: []aiwire.RespTool{{Type: "web_search"}}}, false},
+		{"无工具流式放行", aiwire.RespRequest{Model: "m", Stream: true}, false},
+		{"function工具流式放行", aiwire.RespRequest{Model: "m", Stream: true, Tools: []aiwire.RespTool{{Type: "function", Name: "f"}}}, false},
+		{"code_interpreter放行", aiwire.RespRequest{Model: "m", Tools: []aiwire.RespTool{{Type: "code_interpreter"}}}, false},
+		{"mcp流式放行", aiwire.RespRequest{Model: "m", Stream: true, Tools: []aiwire.RespTool{{Type: "mcp"}}}, false},
+		{"有状态拒绝", aiwire.RespRequest{Model: "m", PreviousResponseID: prev}, true},
+		{"background拒绝", aiwire.RespRequest{Model: "m", Background: &bg}, true},
+		{"未知工具拒绝", aiwire.RespRequest{Model: "m", Tools: []aiwire.RespTool{{Type: "web_search"}, {Type: "file_search"}}}, true},
 	}
-	for _, c := range cases {
-		if _, err := ResponsesToIR(respReq(t, c.raw)); err == nil {
-			t.Errorf("%s 应被拒绝", c.name)
+	for _, test := range tests {
+		t.Run(test.name, func(t *testing.T) {
+			err := RespPassthroughValidate(test.req)
+			if (err != nil) != test.wantErr {
+				t.Fatalf("err = %v, wantErr %v", err, test.wantErr)
+			}
+		})
 	}
 }
-	// input 字符串形态 + store/reasoning 忽略项不报错
-	req := respReq(t, `{"model":"m","input":"你好","store":false,"reasoning":{"effort":"low"},"parallel_tool_calls":true}`)
-	ir, err := ResponsesToIR(req)
-	if err != nil || len(ir.Messages) != 1 || ir.Messages[0].Role != "user" || ir.ReasoningEffort != "low" {
-		t.Errorf("字符串 input = %+v, %v", ir.Messages, err)
-	}
-	// input_image(url 形态)放行并保留图文顺序
-	req2 := respReq(t, `{"model":"m","input":[{"role":"user","content":[{"type":"input_text","text":"看图"},{"type":"input_image","image_url":"https://x/1.png","detail":"low"}]}]}`)
-	ir2, err := ResponsesToIR(req2)
+
+func TestRespPassthroughBody(t *testing.T) {
+	raw := []byte(`{"model":"m","input":"hi","stream":true,"store":true,"max_output_tokens":128,"custom_field":{"a":1.5}}`)
+	out, err := RespPassthroughBody(raw)
 	if err != nil {
-		t.Fatalf("input_image 应放行: %v", err)
+		t.Fatalf("RespPassthroughBody: %v", err)
 	}
-	parts := ir2.Messages[0].Content.Parts
-	if len(parts) != 2 || parts[1].Type != "image_url" || parts[1].ImageURL == nil || parts[1].ImageURL.URL != "https://x/1.png" {
-		t.Errorf("parts = %+v", parts)
+	var body map[string]any
+	if err := json.Unmarshal(out, &body); err != nil {
+		t.Fatalf("unmarshal out: %v", err)
+	}
+	if body["store"] != false {
+		t.Errorf("store 应强制 false, got %v", body["store"])
+	}
+	if body["stream"] != true {
+		t.Errorf("stream 应原样保留, got %v", body["stream"])
+	}
+	if string(out) == "" || !strings.Contains(string(out), `"max_output_tokens":128`) {
+		t.Errorf("数值字段应保真: %s", out)
+	}
+	if !strings.Contains(string(out), `"custom_field"`) {
+		t.Errorf("未知字段应保留: %s", out)
+	}
+	if _, err := RespPassthroughBody([]byte("not-json")); err == nil {
+		t.Error("非法 JSON 应报错")
 	}
 }

-func TestIRRespToResponses(t *testing.T) {
-	resp := &aiwire.ChatResponse{Model: "m", Choices: []aiwire.Choice{{
-		Message: aiwire.ChatMessage{Role: "assistant", Content: aiwire.NewTextContent("你好"),
-			ToolCalls: []aiwire.ToolCall{{ID: "call_9", Type: "function",
-				Function: aiwire.FunctionCall{Name: "f", Arguments: ""}}}},
-		FinishReason: "length",
-	}}, Usage: &aiwire.Usage{PromptTokens: 10, CompletionTokens: 5, TotalTokens: 15}}
-	out := IRRespToResponses(resp, "resp_abc", 1751966400)
-	if out.Status != "incomplete" || out.IncompleteDetails == nil || out.IncompleteDetails.Reason != "max_output_tokens" {
-		t.Errorf("length 应映射 incomplete: %+v", out)
+func TestRespPassthroughUsage(t *testing.T) {
+	tests := []struct {
+		name           string
+		payload        string
+		wantNil        bool
+		prompt, cached int
+	}{
+		{"完整", `{"usage":{"input_tokens":9,"output_tokens":3,"total_tokens":12,"input_tokens_details":{"cached_tokens":5}}}`, false, 9, 5},
+		{"无细分", `{"usage":{"input_tokens":9,"output_tokens":3,"total_tokens":12}}`, false, 9, 0},
+		{"无usage", `{"id":"resp_1"}`, true, 0, 0},
+		{"非法JSON", `xx`, true, 0, 0},
 	}
-	if len(out.Output) != 2 || out.Output[0].Type != "message" || out.Output[1].Type != "function_call" {
-		t.Fatalf("output = %+v", out.Output)
+	for _, test := range tests {
+		t.Run(test.name, func(t *testing.T) {
+			usage := RespPassthroughUsage([]byte(test.payload))
+			if (usage == nil) != test.wantNil {
+				t.Fatalf("usage = %v, wantNil %v", usage, test.wantNil)
 			}
-	if out.Output[0].Content[0].Text != "你好" || out.Output[1].CallID != "call_9" || out.Output[1].Arguments != "{}" {
-		t.Errorf("item 装配错误: %+v", out.Output)
+			if usage == nil {
+				return
 			}
-	if out.Usage.InputTokens != 10 || out.Usage.OutputTokens != 5 || out.Store {
-		t.Errorf("usage/store 错误: %+v", out)
+			if usage.PromptTokens != test.prompt || usage.CachedTokens() != test.cached {
+				t.Fatalf("prompt=%d cached=%d, want %d/%d", usage.PromptTokens, usage.CachedTokens(), test.prompt, test.cached)
+			}
+		})
 	}
 }

-func chunkText(text string) aiwire.ChatChunk {
-	return aiwire.ChatChunk{Choices: []aiwire.ChunkChoice{{Delta: aiwire.Delta{Content: text}}}}
+// TestRespStreamCompletedUsage 断言流式 usage 只从 completed 事件提取。
+func TestRespStreamCompletedUsage(t *testing.T) {
+	completed := []byte(`{"type":"response.completed","response":{"usage":{"input_tokens":10,"output_tokens":5,"total_tokens":15,"input_tokens_details":{"cached_tokens":4}}}}`)
+	if u := RespStreamCompletedUsage(completed); u == nil || u.PromptTokens != 10 || u.CompletionTokens != 5 ||
+		u.PromptTokensDetails == nil || u.PromptTokensDetails.CachedTokens != 4 {
+		t.Fatalf("completed 事件 usage 解析失败: %+v", u)
 	}
-
-func TestRespStreamTextAndTool(t *testing.T) {
-	st := NewRespStream("resp_x", "m", 1751966400)
-	var types []string
-	collect := func(evs []RespEvent) {
-		for _, ev := range evs {
-			types = append(types, ev.Event)
+	for _, data := range []string{
+		`{"type":"response.output_text.delta","delta":"hi"}`,
+		`{"type":"response.completed"}`,
+		`not-json`,
+	} {
+		if u := RespStreamCompletedUsage([]byte(data)); u != nil {
+			t.Fatalf("非 completed 或缺 response 应返回 nil: %s", data)
 		}
 	}
-	collect(st.Feed(chunkText("你")))
-	collect(st.Feed(chunkText("好")))
-	collect(st.Feed(aiwire.ChatChunk{Choices: []aiwire.ChunkChoice{{Delta: aiwire.Delta{
-		ToolCalls: []aiwire.ToolCallDelta{{Index: 0, ID: "call_1", Function: aiwire.FunctionCallDelta{Name: "f", Arguments: "{\"a\""}}}}}}}))
-	collect(st.Feed(aiwire.ChatChunk{Choices: []aiwire.ChunkChoice{{Delta: aiwire.Delta{
-		ToolCalls: []aiwire.ToolCallDelta{{Index: 0, Function: aiwire.FunctionCallDelta{Arguments: ":1}"}}}}}},
-		Usage: &aiwire.Usage{PromptTokens: 3, CompletionTokens: 7, TotalTokens: 10}}))
-	finish := st.Finish()
-	collect(finish)
-	want := []string{
-		"response.created", "response.in_progress",
-		"response.output_item.added", "response.content_part.added", "response.output_text.delta",
-		"response.output_text.delta",
-		"response.output_text.done", "response.content_part.done", "response.output_item.done",
-		"response.output_item.added", "response.function_call_arguments.delta",
-		"response.function_call_arguments.delta",
-		"response.function_call_arguments.done", "response.output_item.done",
-		"response.completed",
-	}
-	if strings.Join(types, "\n") != strings.Join(want, "\n") {
-		t.Errorf("事件序列 =\n%s\nwant\n%s", strings.Join(types, "\n"), strings.Join(want, "\n"))
-	}
-	last := finish[len(finish)-1].Data
-	resp := last["response"].(aiwire.Response)
-	if len(resp.Output) != 2 || resp.Output[0].Content[0].Text != "你好" || resp.Output[1].Arguments != "{\"a\":1}" {
-		t.Errorf("终态快照 = %+v", resp.Output)
-	}
-	if resp.Usage == nil || resp.Usage.TotalTokens != 10 {
-		t.Errorf("终态 usage = %+v", resp.Usage)
-	}
-	// sequence_number 严格递增
-	if seq, ok := last["sequence_number"].(int); !ok || seq != len(want) {
-		t.Errorf("最终 sequence_number = %v, want %d", last["sequence_number"], len(want))
-	}
-}
-
-func TestRespStreamEmpty(t *testing.T) {
-	st := NewRespStream("resp_e", "m", 1)
-	evs := st.Finish()
-	if len(evs) != 2 || evs[0].Event != "response.created" || evs[1].Event != "response.completed" {
-		t.Errorf("空流事件 = %+v", evs)
-	}
 }
@@ -0,0 +1,528 @@
+package service
+
+import (
+	"encoding/json"
+	"fmt"
+	"strings"
+
+	"oci-portal/internal/aiwire"
+)
+
+// Anthropic Messages ↔ OCI OpenAI 兼容面(/actions/v1/responses)直通转换。
+// typed chat 面剔除后 Messages 入口的唯一上游通路。语义损失(README 已披露):
+// stop_sequences / top_k / metadata / thinking 无对应字段,忽略;上游 reasoning
+// 输出项与增量事件丢弃(Anthropic thinking 块含签名语义,不伪造)。
+
+// ErrAiUnsupportedBlock 表示请求含网关无法承接的内容块(仅支持文本/图片/工具块)。
+var ErrAiUnsupportedBlock = fmt.Errorf("暂不支持文本与图片以外的内容块")
+
+// AnthropicToResponsesBody 把 Messages 请求转为直通 body(强制 store:false)。
+func AnthropicToResponsesBody(req aiwire.MessagesRequest) ([]byte, error) {
+	input, err := anthInputItems(req.Messages)
+	if err != nil {
+		return nil, err
+	}
+	body := map[string]any{"model": req.Model, "max_output_tokens": req.MaxTokens,
+		"input": input, "store": false}
+	if sys := req.SystemText(); sys != "" {
+		body["instructions"] = sys
+	}
+	if req.Temperature != nil {
+		body["temperature"] = *req.Temperature
+	}
+	if req.TopP != nil {
+		body["top_p"] = *req.TopP
+	}
+	if tools := anthRespTools(req.Tools); tools != nil {
+		body["tools"] = tools
+	}
+	if tc := anthRespToolChoice(req.ToolChoice); tc != nil {
+		body["tool_choice"] = tc
+	}
+	if req.OutputConfig != nil && req.OutputConfig.Effort != "" {
+		body["reasoning"] = map[string]string{"effort": strings.ToLower(req.OutputConfig.Effort)}
+	}
+	if req.Stream {
+		body["stream"] = true
+	}
+	return json.Marshal(body)
+}
+
+// anthInputItems 把消息序列展开为 Responses input 项:tool_use / tool_result 为
+// 独立 function_call / function_call_output 项,其余聚合为 message 项;遇独立项时
+// 先冲刷已聚合部件,保持块间相对顺序。
+func anthInputItems(messages []aiwire.AnthMessage) ([]any, error) {
+	var items []any
+	for _, m := range messages {
+		var parts []map[string]any
+		flush := func() {
+			if len(parts) > 0 {
+				items = append(items, map[string]any{"role": m.Role, "content": parts})
+				parts = nil
+			}
+		}
+		for _, b := range m.Content.AllBlocks() {
+			item, part, err := anthBlockItem(m.Role, b)
+			if err != nil {
+				return nil, err
+			}
+			if part != nil {
+				parts = append(parts, part)
+			}
+			if item != nil {
+				flush()
+				items = append(items, item)
+			}
+		}
+		flush()
+	}
+	return items, nil
+}
+
+// anthBlockItem 把单个内容块转为独立项或消息部件(thinking 忽略,未知块拒绝)。
+func anthBlockItem(role string, b aiwire.AnthBlock) (item, part map[string]any, err error) {
+	switch b.Type {
+	case "text":
+		return nil, respTextPart(role, b.Text), nil
+	case "image":
+		part, err = anthImageInput(b.Source)
+		return nil, part, err
+	case "tool_use":
+		return map[string]any{"type": "function_call", "call_id": b.ID,
+			"name": b.Name, "arguments": string(b.Input)}, nil, nil
+	case "tool_result":
+		return map[string]any{"type": "function_call_output",
+			"call_id": b.ToolUseID, "output": b.ResultText()}, nil, nil
+	case "thinking", "redacted_thinking":
+		return nil, nil, nil
+	default:
+		return nil, nil, ErrAiUnsupportedBlock
+	}
+}
+
+// respTextPart 按角色选择 Responses 文本部件类型(assistant 历史为 output_text);
+// Anthropic 与 Chat Completions 两条转换链共用。
+func respTextPart(role, text string) map[string]any {
+	if role == "assistant" {
+		return map[string]any{"type": "output_text", "text": text}
+	}
+	return map[string]any{"type": "input_text", "text": text}
+}
+
+// anthImageInput 把 Anthropic image source 转为 Responses input_image 部件。
+func anthImageInput(source json.RawMessage) (map[string]any, error) {
+	var src struct {
+		Type      string `json:"type"`
+		MediaType string `json:"media_type"`
+		Data      string `json:"data"`
+		URL       string `json:"url"`
+	}
+	if err := json.Unmarshal(source, &src); err != nil {
+		return nil, fmt.Errorf("image source 解析失败: %w", err)
+	}
+	switch src.Type {
+	case "base64":
+		if src.MediaType == "" || src.Data == "" {
+			return nil, fmt.Errorf("image source 缺少 media_type 或 data")
+		}
+		return map[string]any{"type": "input_image",
+			"image_url": "data:" + src.MediaType + ";base64," + src.Data}, nil
+	case "url":
+		if src.URL == "" {
+			return nil, fmt.Errorf("image source 缺少 url")
+		}
+		return map[string]any{"type": "input_image", "image_url": src.URL}, nil
+	}
+	return nil, fmt.Errorf("不支持的 image source 类型 %q", src.Type)
+}
+
+func anthRespTools(tools []aiwire.AnthTool) []map[string]any {
+	if len(tools) == 0 {
+		return nil
+	}
+	out := make([]map[string]any, 0, len(tools))
+	for _, t := range tools {
+		out = append(out, map[string]any{"type": "function", "name": t.Name,
+			"description": t.Description, "parameters": t.InputSchema})
+	}
+	return out
+}
+
+// anthRespToolChoice 映射 {type:auto|any|tool|none,name} → Responses 形态。
+func anthRespToolChoice(raw json.RawMessage) any {
+	if len(raw) == 0 {
+		return nil
+	}
+	var tc struct {
+		Type string `json:"type"`
+		Name string `json:"name"`
+	}
+	if json.Unmarshal(raw, &tc) != nil {
+		return nil
+	}
+	switch tc.Type {
+	case "auto":
+		return "auto"
+	case "any":
+		return "required"
+	case "none":
+		return "none"
+	case "tool":
+		return map[string]string{"type": "function", "name": tc.Name}
+	}
+	return nil
+}
+
+// respPayload 是直通响应中本转换关心的子集(未知字段忽略)。
+type respPayload struct {
+	Model             string             `json:"model"`
+	Status            string             `json:"status"`
+	IncompleteDetails *respIncomplete    `json:"incomplete_details"`
+	Output            []respOutputItem   `json:"output"`
+	Usage             *aiwire.RespUsage  `json:"usage"`
+	Error             *map[string]string `json:"error"`
+}
+
+type respIncomplete struct {
+	Reason string `json:"reason"`
+}
+
+type respOutputItem struct {
+	Type      string `json:"type"`
+	CallID    string `json:"call_id"`
+	Name      string `json:"name"`
+	Arguments string `json:"arguments"`
+	Content   []struct {
+		Type string `json:"type"`
+		Text string `json:"text"`
+	} `json:"content"`
+}
+
+// ResponsesToAnthropic 把直通非流式响应转为 Anthropic Messages 响应。
+func ResponsesToAnthropic(payload []byte, msgID string) (*aiwire.MessagesResponse, error) {
+	var resp respPayload
+	if err := json.Unmarshal(payload, &resp); err != nil {
+		return nil, fmt.Errorf("解析上游响应: %w", err)
+	}
+	out := &aiwire.MessagesResponse{ID: msgID, Type: "message", Role: "assistant",
+		Model: resp.Model, Content: []aiwire.AnthBlock{}, StopReason: "end_turn"}
+	for _, item := range resp.Output {
+		switch item.Type {
+		case "message":
+			for _, part := range item.Content {
+				if part.Type == "output_text" && part.Text != "" {
+					out.Content = append(out.Content, aiwire.AnthBlock{Type: "text", Text: part.Text})
+				}
+			}
+		case "function_call":
+			out.Content = append(out.Content, aiwire.AnthBlock{Type: "tool_use", ID: item.CallID,
+				Name: item.Name, Input: argsToJSON(item.Arguments)})
+			out.StopReason = "tool_use"
+		}
+	}
+	if respTruncated(&resp) {
+		out.StopReason = "max_tokens"
+	}
+	out.Usage = anthUsageFromResp(resp.Usage)
+	return out, nil
+}
+
+func anthUsageFromResp(u *aiwire.RespUsage) aiwire.AnthUsage {
+	if u == nil {
+		return aiwire.AnthUsage{}
+	}
+	return aiwire.AnthUsage{InputTokens: u.InputTokens, OutputTokens: u.OutputTokens,
+		CacheReadInputTokens: u.InputTokensDetails.CachedTokens}
+}
+
+// argsToJSON 保证 tool_use.input 是合法 JSON 对象(模型可能产出非法片段)。
+func argsToJSON(args string) json.RawMessage {
+	trimmed := strings.TrimSpace(args)
+	if trimmed == "" {
+		return json.RawMessage(`{}`)
+	}
+	if json.Valid([]byte(trimmed)) {
+		return json.RawMessage(trimmed)
+	}
+	b, _ := json.Marshal(map[string]string{"_raw": args})
+	return b
+}
+
+// ---- Anthropic 流式桥:Responses SSE 事件 → Anthropic 事件序列 ----
+
+// AnthEvent 是一条待写出的 Anthropic SSE 事件。
+type AnthEvent struct {
+	Event string
+	Data  any
+}
+
+// AnthRespBridge 把直通 SSE 事件流桥接为 Anthropic 事件序列:
+// message_start → content_block_start/delta/stop(text 与 tool_use 分块)→ message_delta → message_stop。
+// reasoning 系列事件丢弃;上游 error / response.failed 转 Anthropic error 事件透传。
+type AnthRespBridge struct {
+	id, model   string
+	started     bool
+	blockOpen   bool
+	blockIsTool bool
+	blockIndex  int
+	stopReason  string
+	usage       aiwire.AnthUsage
+	// sawTerminal 标记收到过终态事件(completed/incomplete/failed);
+	// 上游流提前 EOF 时据此发 error 事件而非伪装正常结束
+	sawTerminal bool
+	// errMsg 记录上游错误消息,非空即本流已失败(供调用日志)
+	errMsg string
+}
+
+// NewAnthRespBridge 构造桥;id 为响应消息 ID。
+func NewAnthRespBridge(id, model string) *AnthRespBridge {
+	return &AnthRespBridge{id: id, model: model, blockIndex: -1, stopReason: "end_turn"}
+}
+
+// respStreamEvent 是直通 SSE data JSON 中桥关心的子集。
+type respStreamEvent struct {
+	Type     string          `json:"type"`
+	Delta    string          `json:"delta"`
+	Message  string          `json:"message"`
+	Item     *respOutputItem `json:"item"`
+	Response *respPayload    `json:"response"`
+}
+
+// Feed 消费一行 SSE data JSON,返回应立即写出的事件。
+func (st *AnthRespBridge) Feed(data []byte) []AnthEvent {
+	var ev respStreamEvent
+	if json.Unmarshal(data, &ev) != nil {
+		return nil
+	}
+	if ev.Type == "error" || ev.Type == "response.failed" {
+		return st.failWith(ev)
+	}
+	// message_start 延迟到首个可见输出事件:created / reasoning 阶段不向客户端
+	// 写任何字节,上游此段断流时 handler 才有降级非流式重做的无感窗口
+	var events []AnthEvent
+	switch ev.Type {
+	case "response.output_item.added":
+		if ev.Item != nil && ev.Item.Type == "function_call" {
+			events = append(events, st.ensureStarted()...)
+			events = append(events, st.openBlock(true, ev.Item.CallID, ev.Item.Name)...)
+			st.stopReason = "tool_use"
+		}
+	case "response.output_text.delta":
+		events = append(events, st.ensureStarted()...)
+		events = append(events, st.textDelta(ev.Delta)...)
+	case "response.function_call_arguments.delta":
+		events = append(events, st.argsDelta(ev.Delta)...)
+	case "response.completed", "response.incomplete":
+		st.sawTerminal = true
+		st.finishFrom(ev.Response)
+	}
+	return events
+}
+
+// ensureStarted 在首个可见输出事件前补发 message_start(仅一次)。
+func (st *AnthRespBridge) ensureStarted() []AnthEvent {
+	if st.started {
+		return nil
+	}
+	st.started = true
+	return []AnthEvent{st.startEvent()}
+}
+
+// failWith 记录上游错误并产出 Anthropic error 事件(每流至多一次)。
+func (st *AnthRespBridge) failWith(ev respStreamEvent) []AnthEvent {
+	if st.errMsg != "" {
+		return nil
+	}
+	st.sawTerminal = true
+	msg := ev.Message
+	if ev.Type == "response.failed" {
+		st.finishFrom(ev.Response)
+		if m := respErrorMsg(ev.Response); m != "" {
+			msg = m
+		}
+	}
+	if msg == "" {
+		msg = "上游返回错误事件 " + ev.Type
+	}
+	st.errMsg = msg
+	return []AnthEvent{st.errorEvent()}
+}
+
+// respErrorMsg 提取 response.failed 载荷中的错误消息。
+func respErrorMsg(resp *respPayload) string {
+	if resp == nil || resp.Error == nil {
+		return ""
+	}
+	return (*resp.Error)["message"]
+}
+
+// errorEvent 按 Anthropic 流式协议产出 error 事件。
+func (st *AnthRespBridge) errorEvent() AnthEvent {
+	return AnthEvent{Event: "error", Data: map[string]any{
+		"type":  "error",
+		"error": map[string]string{"type": "api_error", "message": st.errMsg},
+	}}
+}
+
+// Err 返回上游错误消息;空串表示流正常(供调用日志)。
+func (st *AnthRespBridge) Err() string { return st.errMsg }
+
+// SawTerminal 报告是否收到过终态事件;false 即上游流提前终止。
+func (st *AnthRespBridge) SawTerminal() bool { return st.sawTerminal }
+
+// AnthMessageEvents 把完整 Messages 响应展开为标准事件序列,
+// 供流式上游断流后的非流式降级结果推送(客户端协议不变)。
+func AnthMessageEvents(m *aiwire.MessagesResponse) []AnthEvent {
+	events := []AnthEvent{{Event: "message_start", Data: map[string]any{
+		"type": "message_start",
+		"message": map[string]any{
+			"id": m.ID, "type": "message", "role": m.Role, "model": m.Model,
+			"content": []any{}, "stop_reason": nil,
+			"usage": map[string]int{"input_tokens": 0, "output_tokens": 0},
+		},
+	}}}
+	for i, b := range m.Content {
+		events = append(events, anthBlockEvents(i, b)...)
+	}
+	usage := map[string]int{"input_tokens": m.Usage.InputTokens, "output_tokens": m.Usage.OutputTokens}
+	if m.Usage.CacheReadInputTokens > 0 {
+		usage["cache_read_input_tokens"] = m.Usage.CacheReadInputTokens
+	}
+	events = append(events,
+		AnthEvent{Event: "message_delta", Data: map[string]any{
+			"type":  "message_delta",
+			"delta": map[string]any{"stop_reason": m.StopReason, "stop_sequence": nil},
+			"usage": usage,
+		}},
+		AnthEvent{Event: "message_stop", Data: map[string]any{"type": "message_stop"}})
+	return events
+}
+
+// anthBlockEvents 把一个内容块展开为 start / delta / stop 三事件。
+func anthBlockEvents(index int, b aiwire.AnthBlock) []AnthEvent {
+	var start, delta map[string]any
+	if b.Type == "tool_use" {
+		start = map[string]any{"type": "tool_use", "id": b.ID, "name": b.Name, "input": map[string]any{}}
+		delta = map[string]any{"type": "input_json_delta", "partial_json": string(b.Input)}
+	} else {
+		start = map[string]any{"type": "text", "text": ""}
+		delta = map[string]any{"type": "text_delta", "text": b.Text}
+	}
+	return []AnthEvent{
+		{Event: "content_block_start", Data: map[string]any{
+			"type": "content_block_start", "index": index, "content_block": start}},
+		{Event: "content_block_delta", Data: map[string]any{
+			"type": "content_block_delta", "index": index, "delta": delta}},
+		{Event: "content_block_stop", Data: map[string]any{
+			"type": "content_block_stop", "index": index}},
+	}
+}
+
+func (st *AnthRespBridge) textDelta(delta string) []AnthEvent {
+	var events []AnthEvent
+	if !st.blockOpen || st.blockIsTool {
+		events = append(events, st.openBlock(false, "", "")...)
+	}
+	events = append(events, AnthEvent{Event: "content_block_delta", Data: map[string]any{
+		"type": "content_block_delta", "index": st.blockIndex,
+		"delta": map[string]string{"type": "text_delta", "text": delta},
+	}})
+	return events
+}
+
+func (st *AnthRespBridge) argsDelta(delta string) []AnthEvent {
+	if !st.blockOpen || !st.blockIsTool {
+		return nil
+	}
+	return []AnthEvent{{Event: "content_block_delta", Data: map[string]any{
+		"type": "content_block_delta", "index": st.blockIndex,
+		"delta": map[string]string{"type": "input_json_delta", "partial_json": delta},
+	}}}
+}
+
+// finishFrom 记录终态:usage 与 stop_reason(max_output_tokens 截断 → max_tokens)。
+func (st *AnthRespBridge) finishFrom(resp *respPayload) {
+	if resp == nil {
+		return
+	}
+	st.usage = anthUsageFromResp(resp.Usage)
+	if respTruncated(resp) {
+		st.stopReason = "max_tokens"
+	}
+}
+
+func (st *AnthRespBridge) startEvent() AnthEvent {
+	return AnthEvent{Event: "message_start", Data: map[string]any{
+		"type": "message_start",
+		"message": map[string]any{
+			"id": st.id, "type": "message", "role": "assistant", "model": st.model,
+			"content": []any{}, "stop_reason": nil,
+			"usage": map[string]int{"input_tokens": 0, "output_tokens": 0},
+		},
+	}}
+}
+
+// openBlock 关闭当前块并打开新块(text 或 tool_use)。
+func (st *AnthRespBridge) openBlock(isTool bool, toolID, toolName string) []AnthEvent {
+	var events []AnthEvent
+	if st.blockOpen {
+		events = append(events, st.closeBlockEvent())
+	}
+	st.blockOpen, st.blockIsTool = true, isTool
+	st.blockIndex++
+	block := map[string]any{"type": "text", "text": ""}
+	if isTool {
+		block = map[string]any{"type": "tool_use", "id": toolID, "name": toolName, "input": map[string]any{}}
+	}
+	events = append(events, AnthEvent{Event: "content_block_start", Data: map[string]any{
+		"type": "content_block_start", "index": st.blockIndex, "content_block": block,
+	}})
+	return events
+}
+
+func (st *AnthRespBridge) closeBlockEvent() AnthEvent {
+	return AnthEvent{Event: "content_block_stop", Data: map[string]any{
+		"type": "content_block_stop", "index": st.blockIndex,
+	}}
+}
+
+// Finish 在上游流结束后收尾:关块 → message_delta(stop_reason+usage)→ message_stop。
+// 已发过 error 事件的流不再补终态;未见终态事件即 EOF 视为上游提前终止,
+// 发 error 事件而非伪装正常结束(否则客户端拿到"成功的空消息")。
+func (st *AnthRespBridge) Finish() []AnthEvent {
+	if st.errMsg != "" {
+		return nil
+	}
+	if !st.sawTerminal {
+		st.errMsg = "上游流提前终止,未返回终态事件"
+		return []AnthEvent{st.errorEvent()}
+	}
+	var events []AnthEvent
+	if !st.started {
+		st.started = true
+		events = append(events, st.startEvent())
+	}
+	if st.blockOpen {
+		events = append(events, st.closeBlockEvent())
+		st.blockOpen = false
+	}
+	usage := map[string]int{"output_tokens": st.usage.OutputTokens}
+	if st.usage.InputTokens > 0 {
+		usage["input_tokens"] = st.usage.InputTokens
+	}
+	if st.usage.CacheReadInputTokens > 0 {
+		usage["cache_read_input_tokens"] = st.usage.CacheReadInputTokens
+	}
+	events = append(events,
+		AnthEvent{Event: "message_delta", Data: map[string]any{
+			"type":  "message_delta",
+			"delta": map[string]any{"stop_reason": st.stopReason, "stop_sequence": nil},
+			"usage": usage,
+		}},
+		AnthEvent{Event: "message_stop", Data: map[string]any{"type": "message_stop"}},
+	)
+	return events
+}
+
+// Usage 返回聚合到的用量(供调用日志)。
+func (st *AnthRespBridge) Usage() aiwire.AnthUsage { return st.usage }
@@ -0,0 +1,268 @@
+package service
+
+import (
+	"encoding/json"
+	"strings"
+	"testing"
+
+	"oci-portal/internal/aiwire"
+)
+
+func mustAnthReq(t *testing.T, raw string) aiwire.MessagesRequest {
+	t.Helper()
+	var req aiwire.MessagesRequest
+	if err := json.Unmarshal([]byte(raw), &req); err != nil {
+		t.Fatalf("解析请求: %v", err)
+	}
+	return req
+}
+
+// TestAnthropicToResponsesBody 断言 system/消息/工具/effort 的直通装配与 store 强制。
+func TestAnthropicToResponsesBody(t *testing.T) {
+	req := mustAnthReq(t, `{
+		"model": "xai.grok-4.3", "max_tokens": 128, "system": "你是助手", "stream": true,
+		"temperature": 0.5,
+		"messages": [
+			{"role": "user", "content": "东京天气?"},
+			{"role": "assistant", "content": [
+				{"type": "text", "text": "查询中"},
+				{"type": "tool_use", "id": "t1", "name": "get_weather", "input": {"city": "东京"}}
+			]},
+			{"role": "user", "content": [
+				{"type": "tool_result", "tool_use_id": "t1", "content": "晴 25 度"},
+				{"type": "text", "text": "继续"}
+			]}
+		],
+		"tools": [{"name": "get_weather", "description": "查天气", "input_schema": {"type": "object"}}],
+		"tool_choice": {"type": "any"},
+		"output_config": {"effort": "HIGH"}
+	}`)
+	payload, err := AnthropicToResponsesBody(req)
+	if err != nil {
+		t.Fatalf("AnthropicToResponsesBody: %v", err)
+	}
+	var body map[string]any
+	if err := json.Unmarshal(payload, &body); err != nil {
+		t.Fatalf("unmarshal: %v", err)
+	}
+	if body["model"] != "xai.grok-4.3" || body["max_output_tokens"] != float64(128) ||
+		body["instructions"] != "你是助手" || body["store"] != false || body["stream"] != true {
+		t.Fatalf("顶层字段装配错误: %v", body)
+	}
+	if body["tool_choice"] != "required" {
+		t.Fatalf("tool_choice = %v, want required", body["tool_choice"])
+	}
+	reasoning, _ := body["reasoning"].(map[string]any)
+	if reasoning["effort"] != "high" {
+		t.Fatalf("effort = %v, want high(小写透传)", reasoning)
+	}
+	input, _ := body["input"].([]any)
+	// user 消息、assistant 文本消息、function_call、function_call_output、末条 user 文本
+	if len(input) != 5 {
+		t.Fatalf("input 项数 = %d, want 5: %s", len(input), payload)
+	}
+	kinds := make([]string, 0, len(input))
+	for _, it := range input {
+		m := it.(map[string]any)
+		if ty, ok := m["type"].(string); ok {
+			kinds = append(kinds, ty)
+		} else {
+			kinds = append(kinds, "message/"+m["role"].(string))
+		}
+	}
+	want := "message/user,message/assistant,function_call,function_call_output,message/user"
+	if got := strings.Join(kinds, ","); got != want {
+		t.Fatalf("input 顺序 = %s, want %s", got, want)
+	}
+	if !strings.Contains(string(payload), `"output_text"`) {
+		t.Fatalf("assistant 历史文本应为 output_text 部件: %s", payload)
+	}
+}
+
+// TestAnthropicToResponsesBodyRejects 断言不支持的内容块拒绝与图片装配。
+func TestAnthropicToResponsesBodyRejects(t *testing.T) {
+	bad := mustAnthReq(t, `{"model":"m","max_tokens":8,"messages":[
+		{"role":"user","content":[{"type":"document","source":{}}]}]}`)
+	if _, err := AnthropicToResponsesBody(bad); err == nil {
+		t.Fatal("document 块应拒绝")
+	}
+	img := mustAnthReq(t, `{"model":"m","max_tokens":8,"messages":[
+		{"role":"user","content":[{"type":"image","source":{"type":"base64","media_type":"image/png","data":"QUJD"}}]}]}`)
+	payload, err := AnthropicToResponsesBody(img)
+	if err != nil || !strings.Contains(string(payload), "data:image/png;base64,QUJD") {
+		t.Fatalf("图片应转 data URI: %v %s", err, payload)
+	}
+}
+
+// TestResponsesToAnthropic 断言输出块、stop_reason 与 usage 的回转。
+func TestResponsesToAnthropic(t *testing.T) {
+	payload := []byte(`{"model":"xai.grok-4.3","status":"completed","output":[
+		{"type":"reasoning","summary":[]},
+		{"type":"message","content":[{"type":"output_text","text":"你好"}]},
+		{"type":"function_call","call_id":"c1","name":"get_weather","arguments":"{\"city\":\"东京\"}"}],
+		"usage":{"input_tokens":10,"output_tokens":5,"total_tokens":15,"input_tokens_details":{"cached_tokens":4}}}`)
+	out, err := ResponsesToAnthropic(payload, "msg_1")
+	if err != nil {
+		t.Fatalf("ResponsesToAnthropic: %v", err)
+	}
+	if len(out.Content) != 2 || out.Content[0].Type != "text" || out.Content[0].Text != "你好" {
+		t.Fatalf("content 装配错误: %+v", out.Content)
+	}
+	if out.Content[1].Type != "tool_use" || out.Content[1].ID != "c1" || string(out.Content[1].Input) != `{"city":"东京"}` {
+		t.Fatalf("tool_use 装配错误: %+v", out.Content[1])
+	}
+	if out.StopReason != "tool_use" {
+		t.Fatalf("stop_reason = %s, want tool_use", out.StopReason)
+	}
+	if out.Usage.InputTokens != 10 || out.Usage.OutputTokens != 5 || out.Usage.CacheReadInputTokens != 4 {
+		t.Fatalf("usage = %+v", out.Usage)
+	}
+
+	trunc := []byte(`{"model":"m","status":"incomplete","incomplete_details":{"reason":"max_output_tokens"},
+		"output":[{"type":"message","content":[{"type":"output_text","text":"半"}]}]}`)
+	out2, err := ResponsesToAnthropic(trunc, "msg_2")
+	if err != nil || out2.StopReason != "max_tokens" {
+		t.Fatalf("截断 stop_reason = %v, %v", out2, err)
+	}
+}
+
+func bridgeEventTypes(events []AnthEvent) string {
+	kinds := make([]string, 0, len(events))
+	for _, ev := range events {
+		kinds = append(kinds, ev.Event)
+	}
+	return strings.Join(kinds, ",")
+}
+
+// TestAnthRespBridge 断言流桥:文本增量、工具调用切块、reasoning 丢弃与收尾事件。
+func TestAnthRespBridge(t *testing.T) {
+	st := NewAnthRespBridge("msg_1", "m1")
+	var events []AnthEvent
+	feed := func(lines ...string) {
+		for _, l := range lines {
+			events = append(events, st.Feed([]byte(l))...)
+		}
+	}
+	feed(`{"type":"response.created","response":{"model":"m1"}}`,
+		`{"type":"response.reasoning_text.delta","delta":"思考中"}`,
+		`{"type":"response.output_text.delta","delta":"你"}`,
+		`{"type":"response.output_text.delta","delta":"好"}`,
+		`{"type":"response.output_item.added","item":{"type":"function_call","call_id":"c1","name":"f"}}`,
+		`{"type":"response.function_call_arguments.delta","delta":"{\"a\":1}"}`,
+		`{"type":"response.completed","response":{"status":"completed","usage":{"input_tokens":8,"output_tokens":4}}}`)
+	events = append(events, st.Finish()...)
+
+	got := bridgeEventTypes(events)
+	want := "message_start,content_block_start,content_block_delta,content_block_delta," +
+		"content_block_stop,content_block_start,content_block_delta,content_block_stop," +
+		"message_delta,message_stop"
+	if got != want {
+		t.Fatalf("事件序列:\n got %s\nwant %s", got, want)
+	}
+	if st.Usage().InputTokens != 8 || st.Usage().OutputTokens != 4 {
+		t.Fatalf("usage = %+v", st.Usage())
+	}
+}
+
+// TestAnthRespBridgeFailure 断言异常流的错误透传:上游 error / response.failed
+// 事件转 Anthropic error 事件,提前 EOF(未见终态)不再伪装正常结束。
+func TestAnthRespBridgeFailure(t *testing.T) {
+	cases := []struct {
+		name      string
+		lines     []string
+		wantKinds string
+		wantErr   string
+	}{
+		{
+			name:      "上游 error 事件透传",
+			lines:     []string{`{"type":"error","message":"model overloaded"}`},
+			wantKinds: "error",
+			wantErr:   "model overloaded",
+		},
+		{
+			name: "response.failed 提取错误消息",
+			lines: []string{
+				`{"type":"response.output_text.delta","delta":"你"}`,
+				`{"type":"response.failed","response":{"status":"failed","error":{"message":"content filtered"}}}`,
+			},
+			wantKinds: "message_start,content_block_start,content_block_delta,error",
+			wantErr:   "content filtered",
+		},
+		{
+			name:      "空流提前终止",
+			lines:     nil,
+			wantKinds: "error",
+			wantErr:   "上游流提前终止,未返回终态事件",
+		},
+		{
+			name: "输出中途 EOF 无终态",
+			lines: []string{
+				`{"type":"response.output_text.delta","delta":"你"}`,
+			},
+			wantKinds: "message_start,content_block_start,content_block_delta,error",
+			wantErr:   "上游流提前终止,未返回终态事件",
+		},
+	}
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			st := NewAnthRespBridge("msg_1", "m1")
+			var events []AnthEvent
+			for _, l := range tc.lines {
+				events = append(events, st.Feed([]byte(l))...)
+			}
+			events = append(events, st.Finish()...)
+			if got := bridgeEventTypes(events); got != tc.wantKinds {
+				t.Fatalf("事件序列 = %s, want %s", got, tc.wantKinds)
+			}
+			if st.Err() != tc.wantErr {
+				t.Fatalf("Err() = %q, want %q", st.Err(), tc.wantErr)
+			}
+		})
+	}
+}
+
+// TestRespStreamErrorMsg 断言直通流错误事件消息提取。
+func TestRespStreamErrorMsg(t *testing.T) {
+	cases := []struct {
+		name string
+		data string
+		want string
+	}{
+		{name: "error 事件", data: `{"type":"error","message":"boom"}`, want: "boom"},
+		{name: "error 无消息用占位", data: `{"type":"error"}`, want: "上游返回错误事件 error"},
+		{name: "failed 事件", data: `{"type":"response.failed","response":{"error":{"message":"bad"}}}`, want: "bad"},
+		{name: "正常事件返回空", data: `{"type":"response.completed","response":{}}`, want: ""},
+		{name: "非 JSON 返回空", data: `<html>`, want: ""},
+	}
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			if got := RespStreamErrorMsg([]byte(tc.data)); got != tc.want {
+				t.Fatalf("RespStreamErrorMsg = %q, want %q", got, tc.want)
+			}
+		})
+	}
+}
+
+// TestAnthMessageEvents 断言非流式降级结果展开的事件序列与 usage。
+func TestAnthMessageEvents(t *testing.T) {
+	msg := &aiwire.MessagesResponse{
+		ID: "msg_1", Type: "message", Role: "assistant", Model: "m1",
+		Content: []aiwire.AnthBlock{
+			{Type: "text", Text: "好"},
+			{Type: "tool_use", ID: "c1", Name: "f", Input: json.RawMessage(`{"a":1}`)},
+		},
+		StopReason: "tool_use",
+		Usage:      aiwire.AnthUsage{InputTokens: 9, OutputTokens: 3, CacheReadInputTokens: 5},
+	}
+	events := AnthMessageEvents(msg)
+	want := "message_start,content_block_start,content_block_delta,content_block_stop," +
+		"content_block_start,content_block_delta,content_block_stop,message_delta,message_stop"
+	if got := bridgeEventTypes(events); got != want {
+		t.Fatalf("事件序列:\n got %s\nwant %s", got, want)
+	}
+	delta := events[len(events)-2].Data.(map[string]any)
+	usage := delta["usage"].(map[string]int)
+	if usage["input_tokens"] != 9 || usage["output_tokens"] != 3 || usage["cache_read_input_tokens"] != 5 {
+		t.Fatalf("usage = %+v", usage)
+	}
+}
@@ -2,17 +2,19 @@ package service

 import (
 	"context"
+	"encoding/base64"
 	"encoding/json"
 	"errors"
+	"strconv"
 	"time"

 	"oci-portal/internal/oci"
 )

-// 审计查询时间窗（小时）上下限。
+// 批式懒加载的单批条数上下限与缺省值。
 const (
-	minAuditHours = 1
-	maxAuditHours = 72
+	auditLimitDefault = 100
+	auditLimitMax     = 200
 )

 // 审计原始事件缓存:约 3KB/条,上限 2000 条 ≈ 6MB;TTL 内详情秒开。
@@ -21,39 +23,32 @@ const (
 	auditRawTTL = 10 * time.Minute
 )

-// ErrInvalidAuditHours 表示审计时间窗参数越界，handler 据此返回 400。
-var ErrInvalidAuditHours = errors.New("audit events: hours must be between 1 and 72")
-
-// ErrInvalidAuditWindow 表示续查的绝对时间窗非法(格式/顺序/跨度),handler 返回 400。
-var ErrInvalidAuditWindow = errors.New("audit events: invalid start/end window")
+// ErrInvalidAuditCursor 表示续查游标不可解析(过期格式/被篡改),handler 返回 400。
+var ErrInvalidAuditCursor = errors.New("audit events: invalid cursor, refresh to restart")

 // ErrAuditEventGone 表示原始事件缓存过期且小窗重查未找回;handler 映射 404。
 var ErrAuditEventGone = errors.New("原始事件已不可取回,请刷新列表后重试")

-// AuditQuery 是审计查询参数:Hours 为首查(相对当前时刻);
-// Start/End(RFC3339)为续查绝对窗,配合 Page 从截断游标断点续翻。
+// AuditQuery 是批式懒加载查询参数:Cursor 为空表示自当前时刻首查,
+// 非空则从上次响应的游标位置继续向更早回溯;Limit 为单批目标条数。
 type AuditQuery struct {
 	Region string
-	Hours  int
-	Start  string
-	End    string
-	Page   string
+	Cursor string
+	Limit  int
 }

-// AuditEventsView 是审计查询响应:列表不含 raw(详情接口取回);
-// Start/End 回传本次实际使用的绝对窗,截断续查必须原样带回(游标绑定查询参数)。
+// AuditEventsView 是批式查询响应:列表不含 raw(详情接口取回);
+// Cursor 供下一批续查原样带回,空且 Exhausted 表示已到 365 天保留期尽头。
 type AuditEventsView struct {
 	Items     []oci.AuditEvent `json:"items"`
-	Truncated bool             `json:"truncated"`
-	NextPage  string           `json:"nextPage,omitempty"`
-	Start     time.Time        `json:"start"`
-	End       time.Time        `json:"end"`
+	Cursor    string           `json:"cursor,omitempty"`
+	Exhausted bool             `json:"exhausted"`
 }

 // AuditEvents 实时查询租户 OCI 审计事件,纯透传不入库;region 为空时用配置
 // 默认区域。原始事件剥离进进程内缓存,响应体瘦身约 10 倍(调研④方案 A1+B1)。
 func (s *OciConfigService) AuditEvents(ctx context.Context, id uint, q AuditQuery) (AuditEventsView, error) {
-	start, end, err := auditWindow(q)
+	cur, err := decodeAuditCursor(q.Cursor)
 	if err != nil {
 		return AuditEventsView{}, err
 	}
@@ -61,53 +56,73 @@ func (s *OciConfigService) AuditEvents(ctx context.Context, id uint, q AuditQuer
 	if err != nil {
 		return AuditEventsView{}, err
 	}
-	res, err := s.client.ListAuditEvents(ctx, cred, q.Region, start, end, q.Page)
+	res, err := s.client.ListAuditEventsBatch(ctx, cred, q.Region, cur, auditLimit(q.Limit))
 	if err != nil {
 		return AuditEventsView{}, err
 	}
-	return AuditEventsView{
-		Items:     s.stripAuditRaw(res.Items),
-		Truncated: res.Truncated,
-		NextPage:  res.NextPage,
-		// 与实际请求同粒度(分钟),续查回传时窗口逐字节一致
-		Start: start.UTC().Truncate(time.Minute),
-		End:   end.UTC().Truncate(time.Minute),
-	}, nil
+	view := AuditEventsView{Items: s.stripAuditRaw(id, res.Items), Exhausted: res.Exhausted}
+	if res.Cursor != nil {
+		view.Cursor = encodeAuditCursor(*res.Cursor)
+	}
+	return view, nil
 }

-// auditWindow 解析查询窗口:带 start/end/page 走绝对窗校验,否则按 hours 相对窗。
-func auditWindow(q AuditQuery) (time.Time, time.Time, error) {
-	if q.Start != "" || q.End != "" || q.Page != "" {
-		start, err1 := time.Parse(time.RFC3339, q.Start)
-		end, err2 := time.Parse(time.RFC3339, q.End)
-		if err1 != nil || err2 != nil || !start.Before(end) ||
-			end.Sub(start) > maxAuditHours*time.Hour+time.Minute {
-			return time.Time{}, time.Time{}, ErrInvalidAuditWindow
+// auditLimit 归一单批条数:缺省 100,上限 200(响应体量与页预算的平衡)。
+func auditLimit(limit int) int {
+	if limit <= 0 {
+		return auditLimitDefault
 	}
-		return start, end, nil
-	}
-	if q.Hours < minAuditHours || q.Hours > maxAuditHours {
-		return time.Time{}, time.Time{}, ErrInvalidAuditHours
-	}
-	end := time.Now().UTC()
-	return end.Add(-time.Duration(q.Hours) * time.Hour), end, nil
+	return min(limit, auditLimitMax)
 }

-// stripAuditRaw 把每条原始事件按 eventId 放进缓存并从列表剥离。
-func (s *OciConfigService) stripAuditRaw(items []oci.AuditEvent) []oci.AuditEvent {
+// encodeAuditCursor 把游标序列化为不透明字符串(base64url JSON)。
+// 内容仅时间窗与 OCI 翻页令牌,伪造只影响自己的查询范围,无需签名。
+func encodeAuditCursor(cur oci.AuditCursor) string {
+	b, _ := json.Marshal(cur)
+	return base64.RawURLEncoding.EncodeToString(b)
+}
+
+// decodeAuditCursor 解析续查游标;空串返回自当前时刻起的首查游标。
+func decodeAuditCursor(s string) (oci.AuditCursor, error) {
+	if s == "" {
+		return oci.NewAuditCursor(time.Now()), nil
+	}
+	b, err := base64.RawURLEncoding.DecodeString(s)
+	if err != nil {
+		return oci.AuditCursor{}, ErrInvalidAuditCursor
+	}
+	var cur oci.AuditCursor
+	if err := json.Unmarshal(b, &cur); err != nil || cur.Start.IsZero() || !cur.Start.Before(cur.End) {
+		return oci.AuditCursor{}, ErrInvalidAuditCursor
+	}
+	return cur, nil
+}
+
+// auditRawKey 组装租户隔离的原始事件缓存键。
+func auditRawKey(configID uint, eventID string) string {
+	return strconv.FormatUint(uint64(configID), 10) + ":" + eventID
+}
+
+// stripAuditRaw 把每条原始事件按租户与 eventId 放进缓存并从列表剥离。
+func (s *OciConfigService) stripAuditRaw(configID uint, items []oci.AuditEvent) []oci.AuditEvent {
 	for i := range items {
 		if items[i].EventId != "" && items[i].Raw != nil {
-			s.auditRaw.Set(items[i].EventId, items[i].Raw, auditRawTTL)
+			s.auditRaw.Set(auditRawKey(configID, items[i].EventId), items[i].Raw, auditRawTTL)
 		}
 		items[i].Raw = nil
 	}
 	return items
 }

+// InvalidateAuditCache 删除指定租户的全部审计原始事件缓存。
+func (s *OciConfigService) InvalidateAuditCache(configID uint) {
+	s.auditRaw.DeletePrefix(strconv.FormatUint(uint64(configID), 10) + ":")
+}
+
 // AuditEventDetail 取回单条事件的原始 JSON:缓存命中即回;miss 按事件时间
 // 所在分钟起 2 分钟小窗重查兜底(调研④方案 A2),仍未命中报 ErrAuditEventGone。
 func (s *OciConfigService) AuditEventDetail(ctx context.Context, id uint, region, eventID string, eventTime time.Time) (json.RawMessage, error) {
-	if raw, ok := s.auditRaw.Get(eventID); ok {
+	if raw, ok := s.auditRaw.Get(auditRawKey(id, eventID)); ok {
 		return raw.(json.RawMessage), nil
 	}
 	cred, err := s.credentialsByID(ctx, id)
@@ -124,7 +139,7 @@ func (s *OciConfigService) AuditEventDetail(ctx context.Context, id uint, region
 		if ev.EventId == "" || ev.Raw == nil {
 			continue
 		}
-		s.auditRaw.Set(ev.EventId, ev.Raw, auditRawTTL)
+		s.auditRaw.Set(auditRawKey(id, ev.EventId), ev.Raw, auditRawTTL)
 		if ev.EventId == eventID {
 			found = ev.Raw
 		}
--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
wangdefa	b4ef98a25e	AI网关新增xAI格式TTS端点与过滤弃用模型开关 CI / test (push) Successful in 32s Details Release / release (push) Successful in 54s Details - POST /ai/v1/tts:xAI 官方 TTS 格式转换层(text/voice_id→input/voice),复用 Speech 编排,model 缺省 xai.grok-tts;实测 output_format/speed 透传生效 - 「过滤弃用模型」开关(GET/PUT /api/v1/ai-settings,settings 表持久化):开启后已宣布弃用(未退役)模型从列表与路由排除;同步入库与退役提醒不受影响 - AI 网关文档更名 docs/AI网关.md,补 tts 矩阵段与开关说明;README 引用同步 - CHANGELOG 0.5.0(0.4.0 已发版冻结);DASH_VERSION v0.5.0	2026-07-14 09:49:28 +08:00
wangdefa	0a86b5a291	AI网关新增TTS/重排/审核端点,xAI工具扩展,swagger修缺 CI / test (push) Successful in 32s Details Release / release (push) Successful in 1m4s Details - 新端点 /ai/v1/audio/speech(xai.grok-tts)、/rerank(cohere.rerank-v4)、/moderations(OCI Guardrails) - Responses 放行 code_interpreter 与远程 mcp 工具,web_search/x_search 解除仅非流式限制 - 模型能力映射扩展:TEXT_RERANK→RERANK、TEXT_TO_AUDIO→TTS - AI 网关文档独立 docs/ai-gateway.md,字段兼容矩阵只列支持项;README 精简引用 - swagger 修缺:135 处响应注解具体化,RawMessage/联合类型统一渲染 AnyJSON,overrides 迁至 docs/.swaggo - CHANGELOG 0.4.0,版本段不再记日期;DASH_VERSION v0.4.0	2026-07-13 20:17:06 +08:00
wangdefa	9309ad1ffc	修复 AI 网关流式断流:错误透传、自动降级,max_tokens 可缺省 CI / test (push) Successful in 33s Details Release / release (push) Successful in 56s Details	2026-07-13 15:06:01 +08:00
wangdefa	81d4650f3d	代理批量关联租户接口、关于页运行时与资源指标 CI / test (push) Successful in 31s Details Release / release (push) Successful in 57s Details	2026-07-13 12:23:18 +08:00
wangdefa	c7cc5616ed	发布 0.3.0：恢复 Chat 接口并增强云端事件通知 CI / test (push) Successful in 31s Details Release / release (push) Successful in 48s Details	2026-07-13 10:13:44 +08:00
wangdefa	489cb49cb3	AI 网关切换 OpenAI 兼容面并移除 chat 端点,新增模型黑白名单 CI / test (push) Successful in 31s Details Release / release (push) Successful in 54s Details	2026-07-12 17:48:28 +08:00
wangdefa	7706f59549	发布 0.2.0:模型池自愈、探测修正、任务异步触发与删除加固 CI / test (push) Successful in 32s Details Release / release (push) Successful in 52s Details	2026-07-10 20:25:37 +08:00
Wang Defa	dbba1f4905	发布 0.1.0:通知渠道、告警规则、令牌版本与安全加固 CI / test (push) Successful in 30s Details Release / release (push) Successful in 49s Details	2026-07-10 17:38:34 +08:00
@@ -1 +1 @@
 .6.5
 .6.6