feat: Agent 思考模式前端可控、子代理全链路持久化、权限系统、工具 ID 追踪体系、前端面板与文档架构重构

- AgentConfig/LlmClient 新增 enable_thinking 参数，前端 SSE 请求传递 thinking
  开关，仅千问/DashScope 时启用
  - 完善权限系统，支持细粒度的权限控制和用户权限申请
  - delegate_research 工具重命名为 subagent，SubAgentTool/SubAgentRunner 重构
  - 子代理消息（system/user/assistant/tool）持久化到 agent_messages 表，带 agent_name 标识
  - 子代理活动日志（工具调用列表+思考摘要）注入返回结果，Hooks 获得正确 session_id 和 subagent_name
  - LLM 工具调用 ID 回退生成 UUID（llm.rs），ToolCall/ToolResult SSE 事件增加 id/tool_call_id 双字段
  - ToolContext 扩展 session_id/sse_tx/enable_thinking 字段，executor 统一注入而非构造函数传参
  - agent_messages 新增 metadata+raw_json 列，agent_sessions 暴露 summary 字段
  - 删除文件级 transcript 快照（compact.rs），改为依赖 DB 持久化
  - ResearchAgentPanel 重写：TimelineItem 类型替代 StreamStep，支持会话历史回放
  - 新增 AgentMetricsPanel/AskUserQuestionCard/AuditLogViewer 三个前端组件，types.ts 完整类型定义
  - docs/architecture/ 分层重组：概览/核心模块/核心工作流 + agent/ 子目录 11 篇专题文档
  - docs/api.md 补充 RAG/Target/Agent 接口，docs/development.md 新建开发指南
  - .env.example 完全重写，补充 FALLBACK_MODEL 等变量说明

CLAUDE.md

@@ -68,14 +68,19 @@ src/
 │   ├── targets.rs   # Target query/associate/extract, RAG chat, figure chat
 │   └── helpers.rs   # Shared DB helpers, format conversion, path validation
 ├── agent/           # ReAct-based research agent (LLM-driven tool-use loop)
-│   ├── tools/       # AgentTool trait, ToolRegistry, tool implementations per domain file
-│   ├── runtime/     # ReAct loop engine, streaming, session management, context building
-│   ├── compact/     # Context compression (micro/auto/manual layers)
+│   ├── tools/       # AgentTool trait, ToolRegistry, 25+ tool implementations per domain file
+│   ├── runtime/     # ReAct loop, streaming, session/context, token budget, error recovery,
+│   │                #   permission checker, file cache, system prompt assembly, circuit breaker
+│   ├── compact/     # Context compression (micro/auto/manual layers + collapse)
+│   ├── memory/      # Persistent memory manager: extraction, dedup, decay, age, guardrails
 │   ├── hooks.rs     # Lifecycle events (PreToolUse/PostToolUse/Stop/etc.)
-│   ├── skills.rs    # SkillRegistry: loads skill SKILL.md files from skills/ directory
-│   ├── subagent.rs  # Context-isolated sub-agent runner for delegate_research
+│   ├── skills.rs    # SkillRegistry: hot-loads SKILL.md files from skills/ directory
+│   ├── subagent.rs  # Context-isolated sub-agent runner (subagent tool)
+│   ├── team/        # Multi-agent team: file-based inbox, lead/teammate coordination
 │   ├── background.rs# BgNotificationQueue for async slow-task (download/parse) notifications
-│   └── team/        # Multi-agent team: file-based inbox, lead/teammate coordination
+│   ├── task_board.rs# Persistent task board (agent_tasks table)
+│   ├── trajectory.rs# Session trajectory recording for audit/debug
+│   └── terminal.rs  # Escape sequence filter for ANSI-heavy tool outputs
 ├── clients/         # External API wrappers
 │   ├── llm.rs       # LlmClient (OpenAI-compatible chat + streaming), EmbeddingClient
 │   ├── ads.rs       # NASA ADS API
@@ -106,6 +111,7 @@ All handlers access state via `Arc<AppState>`. Key fields:
 - `llm: LlmClient` / `embedding: EmbeddingClient` — OpenAI-compatible LLM clients
 - `ads: AdsClient` / `arxiv: ArxivClient` — academic search clients
 - `skill_registry: Arc<RwLock<SkillRegistry>>` — hot-reloaded agent skills
+- `memory_manager: Arc<MemoryManager>` — persistent agent memory (MEMORY.md + decay)
 - `cancelled_runs: Arc<Mutex<HashSet<String>>>` — agent cancellation tokens
 - `harvest_status` / `batch_status` — async batch operation status tracking
 
@@ -114,14 +120,17 @@ All handlers access state via `Arc<AppState>`. Key fields:
 The agent (`src/agent/`) implements a **ReAct** (Thought → Action → Observation) loop:
 
 1. **`AgentRuntime`** (`runtime/mod.rs`) orchestrates the loop: session create/resume → context build → ReAct loop → finalize
-2. **Streaming**: LLM response is streamed via SSE (`AgentStreamEvent`) — thought, tool_call, tool_result, text_delta, usage, error, done
-3. **Tools**: Each tool implements `AgentTool` trait (name, description, JSON Schema parameters, execute). 19 tools in default registry including read_file, grep_files, glob_files, run_bash, file_write, file_edit, search_papers, download_paper, parse_paper, get_paper_content, rag_search, query_target, save_note, todo_write, compress_context, load_skill, delegate_research, plus optional background and team tools
-4. **Parallel execution**: Same-turn tool calls execute concurrently via `executor::execute_parallel`
-5. **Context compression**: Three layers — micro (placeholder replacement), auto (LLM summarization when over threshold), manual (compress_context tool). Protected by `CompactionCircuitBreaker`
-6. **Skills** (`skills.rs`): Two-layer loading — system-reminder lists names (~20 tokens each), LLM calls `load_skill` to inject full SKILL.md content
-7. **Sub-agents** (`subagent.rs`): `delegate_research` spawns a context-isolated sub-agent with its own ReAct loop, returning only the final summary
-8. **Teams** (`team/`): File-based inbox directory per session for lead/teammate message passing
-9. **Background tasks** (`background.rs`): Slow ops (download, parse) can run async; results inject via `BgNotificationQueue` before next LLM call
+2. **Streaming**: LLM response is streamed via SSE (`AgentStreamEvent`) — thought, tool_call (with `id`), tool_result (with `tool_call_id`), text_delta, usage, error, done. Tool calls execute in parallel.
+3. **Tools**: Each tool implements `AgentTool` trait (name, description, JSON Schema parameters, execute). Core tools: read_file, grep_files, glob_files, run_bash, file_write, file_edit, search_papers, download_paper, parse_paper, get_paper_content, rag_search, query_target, save_note, todo_write, compress_context, load_skill, subagent, ask_user, save_memory. Plus background tools (bg_task_run, bg_task_check) and team tools (spawn_teammate, send_teammate_message, team_broadcast, check_team_inbox).
+4. **Thinking mode**: `enable_thinking` flag propagates from `AgentChatRequest` → `AgentConfig` → `ToolContext` → `LlmClient::chat_stream`. Only enabled for Qwen/DashScope backends; frontend-controlled via the `thinking` request field.
+5. **Tool call ID tracking**: LLM may not return tool_call IDs — `LlmClient` generates UUID fallbacks. `ToolCall` and `ToolResult` SSE events carry matching IDs for precise frontend pairing.
+6. **ToolContext** (`tools/mod.rs`): Injected into every tool execution — holds `app_state`, `session_id`, `sse_tx` (for intermediate events), `enable_thinking`, `read_file_state` (file cache for dedup), `silent` (sub-agents skip permission prompts).
+7. **Context compression**: Four layers — micro (placeholder replacement), snip (old-message truncation), auto (LLM summarization), aggro_micro (aggressive placeholder). Protected by `CompactionCircuitBreaker`. Transcripts persisted in `agent_messages` table, not filesystem snapshots.
+8. **Skills** (`skills.rs`): Two-layer loading — system-reminder lists names (~20 tokens each), LLM calls `load_skill` to inject full SKILL.md content
+9. **Sub-agents** (`subagent.rs`): `subagent` tool spawns a context-isolated sub-agent with its own ReAct loop. Sub-agent messages (system/user/assistant/tool) are persisted to `agent_messages` with `agent_name` identifier. Returns final summary + activity log. SSE progress forwarded to parent via ToolContext.
+10. **Memory** (`memory/`): File-based persistent memory (MEMORY.md). `MemoryManager` handles extraction from conversation, dedup, recency decay, age-based pruning, and guardrails. Tools: `save_memory`, `load_memory` (auto-injected in system prompt).
+11. **Teams** (`team/`): File-based inbox directory per session for lead/teammate message passing
+12. **Background tasks** (`background.rs`): Slow ops (download, parse) can run async; results inject via `BgNotificationQueue` before next LLM call
 
 Environment variables for agent tuning: `AGENT_MAX_STEPS` (default 8), `AGENT_TOOL_TIMEOUT_SECS` (default 120), `AGENT_MAX_TOOL_OUTPUT_CHARS` (default 4000), `AGENT_CONTEXT_CHAR_LIMIT` (default 16000), `AGENT_TOKEN_SOFT_LIMIT` / `AGENT_TOKEN_HARD_LIMIT`.
 
@@ -140,7 +149,7 @@ React 19 + TypeScript + Vite + Tailwind CSS 4. Features are organized by domain:
 - `features/reader/` — Bilingual reader with highlight annotations (KaTeX for math)
 - `features/citation/` — Canvas-based force-directed citation graph
 - `features/sync/` — Batch sync control panel
-- `features/agent/` — Agent chat interface (SSE event consumption)
+- `features/agent/` — Agent chat: ResearchAgentPanel (timeline view with thought/tool_call/answer/subagent), AgentMetricsPanel (tool stats), AskUserQuestionCard (interactive Q&A), AuditLogViewer
 - `features/settings/` — System configuration
 
 Dependencies: `react-markdown` + `rehype-katex` + `remark-math` for Markdown/LaTeX rendering, `framer-motion` for animations, `lucide-react` for icons.