85 Commits

This Branch

This Branch

All Branches

Author	SHA1	Message	Date
gongzhiyongandClaude Opus 4.6	0d97298b6a	feat: web-read-progress Gen-UI card + production mode + security fixes ... - Add web-read-progress Gen-UI card for real-time web reading feedback - Searcher tool-executor pushes loading→complete cards during web_read - ToolCallStatus supports expanding during loading state - Switch from langgraphjs dev to production mode (server.ts) - Add Hono middleware to strip sensitive request headers - Bundle splitting via Vite manualChunks (1.85MB → 8 chunks) - SPA fallback via staticwebapp.config.json - Agent model configs updated (sonnet 1M → standard sonnet) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-04-12 22:18:05 +08:00
gongzhiyongandClaude Sonnet 4.6	5503a7ca9d	fix: add web_search_deep to UI_NAME_MAP in ToolCallStatus and main.tsx ... Previously web_search_deep tool calls were missing from the UI_NAME_MAP, causing their search-result cards to render as standalone cards instead of being embedded within the ToolCallStatus component row. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>	2026-04-12 21:54:15 +08:00
gongzhiyongandClaude Sonnet 4.6	bb1885eca1	fix(searcher): prevent silent termination when MAX_ITERATIONS reached ... When tool rounds hit MAX_ITERATIONS, the router was jumping directly to END without giving the LLM a chance to produce a final text reply. Added a `force-summary` node that invokes the LLM without tools, ensuring a coherent Chinese summary is always generated after all search results are collected. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>	2026-04-12 21:42:59 +08:00
gongzhiyongandClaude Sonnet 4.6	60735de884	fix(security): strip Azure HTTP headers from thread config to prevent IP leak ... The LangGraph API server injects all x-* request headers into thread.config.configurable, which gets persisted and exposed via POST /threads/search. This leaked Azure-injected headers (x-forwarded-for, x-client-ip, x-arr-ssl, etc.) to any user. Adds a Hono middleware (http.app) that strips sensitive infrastructure headers at the HTTP layer before they reach the LangGraph API routes. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>	2026-04-12 20:48:06 +08:00
gongzhiyongandClaude Sonnet 4.6	569e157f25	fix: SPA 404 fallback + JS bundle code splitting + pre-existing build errors ... - Add public/staticwebapp.config.json with navigationFallback for Azure SWA SPA routing - Add manualChunks in vite.config.ts to split vendor deps (react/langgraph-sdk/assistant-ui/radix/markdown/charts/router) - Add property-information resolve alias to fix refractor→hastscript@6 CJS/ESM conflict (pre-existing build failure) - Exclude agent/ from tsconfig.app.json to fix hono missing type error (agent is backend-only, frontend CI uses pnpm vite build directly) Build output chunks: vendor-react: 12.30 kB (gzip: 4.35 kB) vendor-langgraph: 32.85 kB (gzip: 9.29 kB) vendor-markdown: 316.65 kB (gzip: 94.19 kB) vendor-charts: 428.01 kB (gzip: 113.50 kB) vendor-assistant-ui: 740.11 kB (gzip: 260.11 kB) index: 379.74 kB (gzip: 110.17 kB) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>	2026-04-12 20:45:54 +08:00
gongzhiyong	00f5dc3438	fix: flash via lastMessagesRef, web_read timeout 8s, searcher no parallel google+web_read	2026-04-12 17:29:16 +08:00
gongzhiyong	6f036abdaa	feat: CoT output — agents output thinking sentence before tool calls	2026-04-12 17:21:02 +08:00
gongzhiyong	782ef8ab95	style: redesign Stop button — border style, Loader2 spinner during stopping	2026-04-12 17:06:37 +08:00
gongzhiyong	004f26eb94	fix: Stop button uses cancelling poll flow, shows status text during stop	2026-04-12 16:45:36 +08:00
gongzhiyong	1034d4bbc8	fix: remove hardcoded secrets, centralize storage config, move dev deps to devDependencies	2026-04-12 16:16:11 +08:00
gongzhiyong	fb23b64df2	feat: Redis tool cache, Service Bus alerts, doc-creator agent for edit/translate	2026-04-12 16:15:45 +08:00
gongzhiyong	33cd268003	fix: jina partial_success, bash base64, ticket parallel search, ExecutionLogPanel toolCallId, sanitizeTitle regex	2026-04-12 16:06:21 +08:00
gongzhiyongandClaude Sonnet 4.6	73a7175051	feat: cancelling status, ThreadSidebar lastActive time, MessageBubble smart summary ... - 点5: cancelling 独立状态 UI，用轮询替换 300ms setTimeout，StreamStatusBar 加橙色视觉 - 点6: ThreadSidebar formatTime 支持毫秒时间戳，时间展示改用 lastActive - 点7: MessageBubble 智能摘要（关键词优先/标题次优/跳开场白），折叠阈值提升至25行/800字，展开后可收起 Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-12 15:44:58 +08:00
gongzhiyongandClaude Sonnet 4.6	1be8b2602d	refactor: preValidateToolCall rules array, extract getLastHumanContent helper ... Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-12 15:43:04 +08:00
gongzhiyongandClaude Sonnet 4.6	e4ba9f78b1	feat: add toolCallId to ExecutionLogEntry for precise frontend matching ... Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-12 15:28:29 +08:00
gongzhiyongandClaude Sonnet 4.6	107069ea06	feat: execution_log full fields in ToolCallStatus, index matching, ExecutionLogPanel overview badges ... Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-12 15:28:27 +08:00
gongzhiyongandClaude Sonnet 4.6	4895ff5465	feat: message queue upgrade, full payload snapshot, StreamStatusBar queue count, config badge ... - Replace single pendingSubmitRef with pendingQueueRef array supporting multiple queued messages - Add SubmitPayload interface to snapshot text/file/tools/modelMode at submission time - Rewrite handleSubmit/doSubmit to use payload snapshots instead of closures over mutable state - Rewrite handleConcurrentQueue with polling loop that drains the queue sequentially - Rewrite handleConcurrentInterrupt and handleConcurrentCancel for queue awareness - StreamStatusBar queued state now shows count (N 条消息已排队) and inline cancel button - Add key={concurrentStatus} to StreamStatusBar div for fade-in animation on status change - Config panel summary button shows numeric badge instead of dot when tools are active Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-12 15:28:16 +08:00
gongzhiyong	964e0dd305	fix: load historical messages when switching threads via client.threads.getState	2026-04-12 15:14:29 +08:00
gongzhiyongandClaude Sonnet 4.6	0d36c0fc0d	feat: execution log panel, stream status bar, config polish, card sorting, error recovery CTA ... - Point 5: dual-layer card sort (type priority + sort_key) - Point 3: config summary text, active button style, Escape key, open animation - Point 2: ConcurrentStatus state machine with StreamStatusBar - Point 1: ExecutionLogPanel component replaces text-based execution summary - Point 4: error-result gets fallback tool suggestions and direct-answer CTA Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-12 15:10:17 +08:00
gongzhiyongandClaude Sonnet 4.6	531197fd68	feat: add originalQuery to error-result props, semantic sort_key weights for card ordering ... - All error-result ui.push() calls now include originalQuery derived from the failed tool's input args - Added CARD_SORT_WEIGHT constant mapping card names to semantic sort priorities (error=100, chart=200, knowledge=300, ticket=400, search=500, code=600, doc=700, actions=900) - All sort_key values now computed as CARD_SORT_WEIGHT * 1_000_000 + timestamp for stable ordering Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-12 15:09:13 +08:00
gongzhiyong	66eaf7680d	fix: sanitize thread titles — reject UUID/hash/empty values, fallback to 新对话	2026-04-12 15:01:29 +08:00
gongzhiyong	b93f08f206	feat: soc-client timeout via toolConfig, 24 unit tests, fix abort early-check and placeholder regex	2026-04-12 14:57:06 +08:00
gongzhiyongandClaude Opus 4.6	38395226e5	feat: fix filterTools modelMode, tool call limit, temperature default, preValidate routing rules ... - filterTools now actually filters by modelMode (flash excludes search tools) - Hard limit of 3 tool calls per turn in preValidateToolCall - Temperature default changed from 0.5 to 0.2 to match MODEL_PRESETS.auto - ticket_list blocked when message contains specific ticket ID (TK-xxxx) - google_search/web_search_deep blocked for internal knowledge queries Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-04-12 14:49:35 +08:00
gongzhiyong	23e228b0f9	docs: fix CLAUDE.md — Supervisor uses AzureOpenAI not Gemini	2026-04-12 14:48:26 +08:00
gongzhiyongandClaude Sonnet 4.6	7167109414	feat: collapse model/tool config into expandable summary bar ... Replace flat tool toggles + model selector bar with a default-collapsed summary chip (e.g. "Auto · 已开启 3 个工具 ▾") that expands upward into a compact panel. Panel auto-closes on outside mousedown click. Options changes keep the panel open. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-12 14:40:18 +08:00
gongzhiyongandClaude Sonnet 4.6	d86da0e48b	feat: abort/cancel support in tool-executor and retry ... Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-12 14:33:47 +08:00
gongzhiyongandClaude Sonnet 4.6	f375220eb6	feat: concurrent input dialog, unlock textarea during loading, fix TS errors ... - Add showConcurrentDialog modal when submitting while AI is loading - Offer interrupt (stop + submit) or queue (poll until done) options - Remove disabled={thread.isLoading} from textarea and FileUploadButton - Fix TS2352: cast BaseMessage to unknown first before Record - Fix TS6133: rename unused toolName param to _toolName - Fix TS2769: chart_type typed as union literal, not string - Fix TS2322: separate cmdMap lookup from const declaration in soc-client Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-12 14:31:29 +08:00
gongzhiyongandClaude Sonnet 4.6	e8a29e1f89	feat: toolConfig.ts, expanded preValidate rules, envelope unification, fallbackFrom, temperature pass-through ... - New src/agent/utils/toolConfig.ts: centralized per-tool timeout/retry config - preValidateToolCall: kb_search min query length + duplicate guard, ticket_list page range, ticket_detail placeholder ID detection - ExecutionLogEntry: added fallbackFrom and retryCount fields - retry.ts: added onAttempt callback for retry counting - tool-executor.ts: all 7 tool content returns unified to {ok, tool, summary, resultCount, data, error, fallback} envelope - supervisor/types.ts: temperature slider config field - enterprise/agent.ts: temperature pass-through to createLlm() Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-12 14:16:19 +08:00
gongzhiyongandClaude Sonnet 4.6	fc56503e22	feat: ToolCallStatus partial_success/fallback_success visual states ... - Add CornerDownRight, AlertCircle icons for fallback/partial states - New isPartial/isFallback props on ToolCallRow with hint banners - Read execution_log from thread.values in main.tsx, pass to ToolCallStatus - Match by tc.name === entry.tool to determine partial/fallback per row Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-12 14:13:14 +08:00
gongzhiyongandClaude Sonnet 4.6	2f201bf1f0	feat: expand preValidateToolCall rules, add traceId+batch summary, extend log protocol to all agents ... - Add validation for web_search/google_search (query length), sandbox_run/code_execute (empty code + dangerous commands), doc_create/report_generate/reply_draft (title/topic required) - Add traceId field to ExecutionLogEntry, generate per-request traceId in enterprise tool-executor with batch summary log - Add structured console.log tool_exec events to searcher, coder, writer tool-executors (no state changes) Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-12 13:55:19 +08:00
gongzhiyongandClaude Sonnet 4.6	b22d13415a	fix: make GOOGLE_API_KEY and DATABASE_URL optional in dev environment ... GOOGLE_API_KEY is not used at startup (no code calls config.google.apiKey). DATABASE_URL is only needed for production checkpointer. Both are now warned-but-not-fatal in dev (NODE_ENV != production) to allow local development without these credentials. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-12 13:06:13 +08:00
gongzhiyongandClaude Sonnet 4.6	9c96913b4d	feat: enterprise product sprint — log protocol, tool rules, UX polish ... Backend: - config.ts: unified startup env validation (throw on missing critical vars) - tool-executor: structured logToolCall() JSON, ToolStatus enum (success/partial_success/fallback_success/error), preValidateToolCall() hardcoded guards for ticket_detail/chart_generate, durationMs/inputSummary/resultCount on all execution log entries - agent.ts: tool selection decision tree, chart-as-default-path prompt rules, source labels [知识库][工单][网络][推断] - tool-defs.ts: applicable/not-applicable guidance on all 4 tools - soc-client.ts: sandbox hardening (10k char limit, 15s timeout, output truncation, error classification), config.* accessors - router.ts: preCheckRoute() rules — TK-xxx/工单/知识库 → enterprise direct; greetings → generalInput - supervisor/types.ts: removed dead config fields (model/temperature/maxTokens/systemPrompt) - Remove chat-agent (legacy entry point) Frontend: - MessageBubble: source badge rendering [知识库][工单][网络][推断], CitationChip [1][2] → clickable chips - ThreadSidebar: auto-title from first message, collapsible search, long title truncation - ToolCallStatus: tool-specific loading labels, collapse-all toggle for multi-tool - main.tsx: conclusion-first layout (AI text above artifacts), draft persistence, IME fix, auto chip, multi-tool AnalysisBlock container, sort_key ordering, retry via soc:retry-tool - chart-result: empty guard, chart/table toggle, multi-chart format support - ticket-summary/knowledge-result: work-card quick actions (prefill with context) - ActionBar: structured payload {text, taskType, sourceCardId}, source label UI - index.css: card-enter slide animation, dot-bounce loading Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-12 13:01:02 +08:00
gongzhiyongandClaude Sonnet 4.6	183794dc42	fix: sandbox friendly error card + ticket_list empty guard + search auto-supplement ... - coder/tool-executor: push sandbox-result card with friendly message on failure instead of raw error, guide user to local execution or retry - enterprise/tool-executor: skip ticket-summary card when ticket_list returns empty, only return hint in tool message to avoid blank UI - searcher/tool-executor: auto-supplement web_search_deep with google_search when results < 3, matching existing google_search behavior - ticket_detail UUID: already correctly uses find() match (no change needed) Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-11 22:35:52 +08:00
gongzhiyongandClaude Sonnet 4.6	19cb2e42bc	fix(ui): fix ToolCallStatus duplicate label + add fade-in to standalone cards ... - Fix ToolCallStatus label duplication: label span was shown twice when failed/done; consolidate into a single span with correct copy - Add animate-in fade-in duration-300 wrapper to standalone Gen-UI cards and streaming orphan cards in main.tsx (already present in ToolCallStatus expanded view) Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-11 22:34:59 +08:00
gongzhiyongandClaude Sonnet 4.6	3b1ce32bc4	feat(R3): deduplicate UI cards by card_id (loading → complete merge) ... Cards with the same card_id in props are deduplicated to show only the last (complete) version, replacing any earlier loading skeleton cards. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-11 22:33:45 +08:00
gongzhiyongandClaude Sonnet 4.6	c4c5c9916e	feat(R3): chart-result + reply-draft + next-actions + citations ... - enterprise tool-executor: push chart-result card after ticket_list with status/priority distribution pie+bar charts - enterprise tool-executor: push next-actions card after successful tool executions with contextual suggested actions - enterprise tool-executor: add citations array to kb_search results - searcher tool-executor: add citations array to google_search and web_search_deep results - writer tool-defs: add replyDraftSchema and reply_draft tool definition - writer tool-executor: add reply_draft case with LLM-generated drafts in customer/internal modes, push reply-draft Gen-UI card Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-11 22:33:35 +08:00
gongzhiyongandClaude Sonnet 4.6	e408ca9977	feat(R3): reply-draft + next-actions cards + citations in knowledge/search results ... - Add reply-draft card (customer/internal mode, inline edit, copy) - Add next-actions card (suggested action buttons with soc:prefill-input dispatch) - Register reply-draft and next-actions in agent-uis ComponentMap - Add citations[] prop to knowledge-result and search-result cards Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-11 22:32:25 +08:00
gongzhiyongandClaude Sonnet 4.6	f3cf9d027f	feat(R3): reply-draft/next-actions/chart prompt quality ... Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-11 22:31:13 +08:00
gongzhiyongandClaude Sonnet 4.6	397c2c1124	feat: Gen-UI卡片error态 + ToolCallStatus失败状态 + 骨架屏loading ... - 4个Gen-UI卡片增加errorMessage prop，错误时显示红色提示块，空结果时边框变红 - ToolCallStatus新增failedToolIds prop，失败时显示XCircle红色图标+失败文案 - main.tsx从tool messages status===error派生failedToolIds并传入ToolCallStatus - ToolCallRow展开无uiItem时显示animate-pulse骨架屏占位 - LoadExternalComponent外包animate-in fade-in过渡动画 Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-11 21:59:12 +08:00
gongzhiyongandClaude Sonnet 4.6	2430fd2174	fix(P1): ticket UUID mapping + sandbox/ticket/search fallback improvements ... - Fix ticket_detail UUID resolution: use find() to match ticketNumber instead of tickets[0] - Add sandbox fallback hint in coder tool-executor catch block - Add ticket_list empty result friendly hint with suggestions - Add google_search auto-supplement with Jina when results < 3 - Enhance enterprise tool-executor general catch with fallback_hint Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-11 21:57:56 +08:00
gongzhiyongandClaude Sonnet 4.6	979e900a17	feat(prompt): 4个Agent system prompt追加工具失败处理规范 ... 在 enterprise/coder/searcher/writer 的 agent.ts 中， 为 SYSTEM_PROMPT 追加"工具失败处理规范"段落， 指导 LLM 在工具调用失败时给出中文友好提示和替代建议， 不暴露技术报错细节。 Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-11 21:57:42 +08:00
gongzhiyongandClaude Sonnet 4.6	4ad169e891	fix(P0): Daytona sandbox use process/execute endpoint + add sourceType to writer doc_edit/translate ... - Fix Daytona code execution endpoint from process/code-run (404) to process/execute - Fix request body format to {command} with language-aware shell wrapping - Add missing sourceType/confidence props to writer doc_edit and doc_translate ui.push Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-11 20:01:33 +08:00
gongzhiyongandClaude Sonnet 4.6	0e92444112	feat: prompt optimization — intent routing, source annotation, enterprise output template ... - supervisor router: intent-based routing (no longer keyword-dependent) - enterprise/searcher/writer/coder agents: source annotation [内部知识][工单数据][网络搜索][模型推断] - enterprise agent: enterprise output template (结论→依据→建议→下一步) - enterprise agent: cross-source synthesis prompt for multi-tool results - writer agent: proactive enterprise closure actions (report generation hints) Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-11 20:00:21 +08:00
gongzhiyongandClaude Sonnet 4.6	8c5890d66a	feat: Task#2 ActionBar+SourceBadge+状态语言统一+长消息分层 ... - 新增 ActionBar 组件：快速操作按钮，通过 soc:prefill-input 事件填充输入框 - 新增 SourceBadge 组件：数据来源标注（内部知识库/工单系统/外部搜索/代码执行） - 4个 Gen-UI 卡片集成 ActionBar 和 SourceBadge - main.tsx 添加 soc:prefill-input 事件监听 - ToolCallStatus 状态文案统一为中文（正在xxx.../已完成·xxx） - MessageBubble 长消息（>800字）显示摘要+展开全文按钮 Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-11 19:59:24 +08:00
gongzhiyongandClaude Sonnet 4.6	e16b4ce92e	fix: P0/P1/P2 + sourceType/confidence + kb fallback + report_generate ... - Daytona: fix /workspace→/sandbox + toolbox proxy exec endpoint - KB Agent: timeout 15s→45s, exponential retry (3x, 1s/2s/4s) - ticket_detail: auto-resolve TK-xxxx ticketNumber to UUID - retry.ts: add exponential backoff support - All ui.push(): add sourceType + confidence fields - kb_search: fallback to google_search on failure with error card - writer: add report_generate tool with structured Markdown template Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-11 19:59:13 +08:00
gongzhiyongandClaude Opus 4.6	b8fcc50d73	feat: optimize prompts — intent-based routing, source annotation, enterprise output format ... - Supervisor router: rewrite to intent-based routing instead of tool-name matching - Enterprise agent: add structured source annotation ([内部知识]/[工单数据]/[网络搜索]/[模型推断]) - Enterprise agent: add enterprise-style output format (结论/依据/建议) for responses >200 chars - Enterprise agent: add multi-source synthesis guidelines for cross-tool analysis - Writer agent: add enterprise document closure prompts (auto-suggest reports after queries) - Supervisor index: update ALL_TOOL_DESCRIPTIONS to match intent-based approach Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>	2026-04-11 19:57:55 +08:00
gongzhiyongandClaude Opus 4.6	ba59259458	docs: 产品 Agent MD 写入 8 项优化方案评审结果 ... Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-04-11 19:14:33 +08:00
gongzhiyongandClaude Opus 4.6	d6946be6c0	docs: 产品验收测试报告 — 97.2% 通过率，评分 8.0/10 ... Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-04-11 14:45:02 +08:00
gongzhiyongandClaude Opus 4.6	a41f2fbf44	feat: P1 滚动到底部浮动按钮 ... Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-04-11 14:44:14 +08:00
gongzhiyongandClaude Opus 4.6	dd70cadcc1	feat: P0 前端交互完善 + 测试报告更新 ... - ToolCallStatus 组件增强 - main.tsx P0 交互改进（textarea/光标/编辑消息） - 更新测试报告：9/12 通过，3 个外部服务问题 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-04-11 14:42:56 +08:00
gongzhiyongandClaude Opus 4.6	86cfa317b6	feat: ChatGPT Enterprise 对标优化 — Prompt 重构 + 前端 P0 交互 + 错误处理 ... - 重构所有 Agent System Prompt，参考 ChatGPT 风格：自然对话、匹配用户沟通风格、避免翻译腔 - Router 路由规则结构化，新增 Decision Rules 和分组示例，提升路由准确率 - 前端 P0：textarea 多行输入 + Shift+Enter 换行、流式打字光标动画、编辑已发送消息 - 新增 ToolCallStatus 组件、process-attachment 文件处理节点、retry 重试工具 - 工具调用错误响应友好化（formatToolError 中文提示） - Enterprise tool descriptions 增加使用场景说明 - 上下文截断提示改为中文 - 新增 soc-product-agent 产品经理 Agent 定义 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-04-11 14:35:45 +08:00
gongzhiyong	6400d85027	fix: use direct host IP http://192.168.30.30:2024 for LangGraph API	2026-04-11 13:36:37 +08:00
gongzhiyong	216650837a	fix: use window.location.origin as fallback URL for Vite proxy in Docker	2026-04-11 12:49:26 +08:00
gongzhiyongandClaude Sonnet 4.6	a505e441ad	feat: file upload + context compression + fix maxTokens ... - Add file-service.ts: Azure Blob upload + PDF/Excel/image parsing - Add truncate-messages.ts: context compression for long conversations - Fix maxTokens: flash 500→2000, auto 2048→4096, pro 4096→8192 - Integrate truncateMessages in general-input and router nodes - Add @azure/storage-blob, pdf-parse, xlsx dependencies Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>	2026-04-11 12:31:55 +08:00
gongzhiyong	9b5b98f675	feat: add FileUploadButton and FileAttachmentPreview components	2026-04-11 12:31:24 +08:00
gongzhiyong	d878135443	fix: Vite proxy for Docker — frontend routes API through same-origin proxy to langgraph-dev:2024	2026-04-11 12:28:52 +08:00
gongzhiyongandClaude Opus 4.6	3eb70fec59	fix: coder agent must always call code_execute, never return plain text code ... Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-04-11 11:03:50 +08:00
gongzhiyongandClaude Opus 4.6	e4760ca66e	fix: CanvasPanel returns null when no doc, no longer occupies space ... Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-04-11 02:24:11 +08:00
gongzhiyongandClaude Opus 4.6	8a5d1c0ea9	fix: router coder/writer distinction + writer must use doc_create tool ... - Router prompt now clearly distinguishes "写代码" (execute code → coder) from "写文档" (create document → writer) - Writer prompt strengthened to always use doc_create, never plain text Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-04-11 02:20:22 +08:00
gongzhiyongandClaude Opus 4.6	77e64d945e	feat: Phase 2 - Deep Search, Code Interpreter, Writer agents + CanvasPanel ... - Add searcher agent with Jina deep search + KB search tools - Add coder agent with Daytona sandbox code execution - Add writer agent with canvas document creation/editing tools - Integrate all sub-agents into supervisor router - Add CanvasPanel drawer UI component for document rendering - Refactor enterprise tools into shared soc-client module Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-04-11 01:52:10 +08:00
gongzhiyongandClaude Sonnet 4.6	08add81409	feat: Phase 2 - CanvasPanel drawer + sidebar search & date groups ... - Add CanvasPanel.tsx: right-side 480px drawer, renders markdown via MessageBubble or code via SyntaxHighlighter, with Copy/Download/Close header; listens to window 'open-canvas' CustomEvent dispatched by Gen-UI cards - Update ThreadSidebar.tsx: add search input below new-chat button with live filtering; group filtered threads by date (今天/昨天/本周/更早) with section headings; internal state only, Props interface unchanged - Update main.tsx: import CanvasPanel, add canvasDoc state, register open-canvas event listener on mount, render CanvasPanel alongside the main chat column inside a flex-row wrapper Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>	2026-04-11 01:48:25 +08:00
gongzhiyongandClaude Opus 4.6	251c6586f4	feat: Phase 1 — ReAct loop + Markdown rendering + Docker deployment ... Backend: - Enterprise Agent refactored from single-round to ReAct multi-turn loop - New agent.ts (LLM decision node) + tool-executor.ts (tool execution + Gen-UI) - tool-defs.ts extracted for shared tool schemas - MAX_ITERATIONS=6 safeguard against infinite loops Frontend: - MessageBubble: Markdown + code highlighting + LaTeX + tables - ThemeToggle: light/dark/system theme cycling - chart-result Gen-UI card: recharts bar/line/pie/area charts Infrastructure: - Docker Compose (lightweight): only LangGraph + Frontend, Azure cloud for PG/Redis/Blob - Dockerfiles for dev (hot reload) and prod - Makefile with dev/prod/down/logs commands - Updated CLAUDE.md and agent definitions for LangGraph.js architecture Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-04-11 01:29:56 +08:00
gongzhiyongandClaude Opus 4.6	b30cb0e758	fix: add .npmrc and CUSTOM_BUILD_COMMAND to resolve Oryx npm peer dep conflict ... Oryx runs `npm install` which fails on peer dependency conflicts between @assistant-ui/react@^0.8.0 and @assistant-ui/react-syntax-highlighter@^0.7.2. Adding .npmrc with legacy-peer-deps=true and CUSTOM_BUILD_COMMAND for pnpm. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-04-10 13:36:30 +08:00
gongzhiyongandClaude Opus 4.6	45685e8b4a	fix: resolve Web App startup timeout — exclude node_modules from zip, add startup.sh, increase timeout to 600s ... - Remove node_modules from deploy zip (302MB -> ~1MB), let Oryx build on Azure - Add startup.sh: creates .env/.langgraph_api prereqs, installs pnpm, runs langgraphjs dev - Azure config: WEBSITES_CONTAINER_START_TIME_LIMIT=600, SCM_DO_BUILD_DURING_DEPLOYMENT=true, ENABLE_ORYX_BUILD=true, PRE_BUILD_COMMAND="npm install -g pnpm@10" - Delete auto-generated main_soc-langgraph.yml (conflicts with deploy-langgraph.yml) - Enable Always On to avoid cold starts Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-04-10 06:17:47 +08:00
gongzhiyongandClaude Opus 4.6	32f83d714e	refactor: migrate LangGraph server from Container App to Azure Web App ... ## Changes - Delete Dockerfile and .dockerignore (no longer needed) - Rewrite deploy-langgraph.yml: zip deploy to Web App (not ACR build) - Update deploy-langgraph-ui.yml: VITE_LANGGRAPH_URL → soc-langgraph.azurewebsites.net ## Azure Resources - CREATED: soc-langgraph Web App (Node.js 20, B1 Linux, southeastasia) - CREATED: soc-langgraph-plan (App Service Plan, B1, southeastasia) - DELETED: soc-langgraph Container App - DELETED: soc-cae Container App Environment - DELETED: socsocacr ACR - All env vars migrated to Web App - Startup: npx langgraphjs dev --host 0.0.0.0 --port 2024 --no-browser - WEBSITES_PORT=2024 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-04-10 05:47:32 +08:00
gongzhiyong	4a485a9c5b	fix(dockerfile): pre-create .langgraph_api to prevent tsx watch restart loop	2026-04-10 05:23:07 +08:00
gongzhiyong	3d35ca43be	fix(dockerfile): create empty .env file — langgraphjs dev requires it to exist	2026-04-10 05:05:17 +08:00
gongzhiyong	df2b72c53c	fix(dockerfile): explicit port binding for Container App health probe	2026-04-10 04:54:47 +08:00
gongzhiyong	f8a73e461c	fix(ci): point app_location directly to langgraph/dist for SWA upload	2026-04-10 04:44:57 +08:00
gongzhiyong	14a7e79792	fix(ci): use pnpm vite build then SWA upload from repo root with langgraph/dist	2026-04-10 04:43:02 +08:00
gongzhiyong	3edeafa062	fix(ci): let SWA action handle Vite build instead of manual pnpm build	2026-04-10 04:40:14 +08:00
gongzhiyong	f43d239e6c	fix(main.tsx): cast message prop to any in LoadExternalComponent	2026-04-10 04:37:21 +08:00
gongzhiyong	8d30e0cca0	fix(main.tsx): rename UIMessage→UIMsgLocal to avoid SDK type conflict	2026-04-10 04:35:15 +08:00
gongzhiyong	4f75b65d52	fix(main.tsx): add type field to local UIMessage type	2026-04-10 04:32:06 +08:00
gongzhiyong	844aaa8dcc	fix(main.tsx): resolve TypeScript errors — UIMessage type, role→type, ComponentMap cast	2026-04-10 04:28:28 +08:00
gongzhiyongandClaude Sonnet 4.6	7a3cc140b0	refactor: replace SOC system with LangGraph.js gen-ui — full cleanup ... ## Removed (old SOC system) - backend/ — Python FastAPI + LangGraph Python ReAct agent - frontend/ — Next.js Gemini-style UI - config/, doc/ — old documentation - .github/workflows/deploy-backend.yml - .github/workflows/deploy-frontend.yml ## Added (new LangGraph.js system) - langgraph/src/agent/enterprise/ — enterprise agent with 6 tools - kb_search → KnowledgeResultCard - ticket_list/detail → TicketSummaryCard / TicketDetailCard - web_search (Jina Search+Reader) → SearchResultCard - sandbox_run (Daytona REST) → SandboxResultCard - langgraph/src/agent-uis/enterprise/ — UI card components - .github/workflows/deploy-langgraph-ui.yml — Vite SPA → Azure Static Web App ## Azure Resources - soc-backend webapp: DELETED - soc-frontend Static Web App (eastasia): DELETED - soc-langgraph-ui Static Web App (eastasia): CREATED URL: salmon-mushroom-0d8872e00.7.azurestaticapps.net Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-10 04:23:51 +08:00
gongzhiyongandClaude Sonnet 4.6	645f1ecaae	feat(langgraph): add enterprise UI card components and implement main.tsx Chat entry ... - Create 5 enterprise UI card components under src/agent-uis/enterprise/: knowledge-result, ticket-summary, ticket-detail, search-result, sandbox-result - Update ComponentMap in src/agent-uis/index.tsx to replace () => null placeholders with real component imports - Implement src/main.tsx with useStream + LoadExternalComponent for full Chat UI, connecting to LangGraph agent with VITE_LANGGRAPH_URL env var Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-10 04:23:32 +08:00
gongzhiyong	b621fbae92	fix(langgraph): add missing pnpm-lock.yaml for Docker build	2026-04-10 04:04:27 +08:00
gongzhiyongandClaude Sonnet 4.6	ac388b3c07	feat: add langgraph/ — LangGraph.js gen-ui server with CI/CD to Azure Container App ... ## Azure Resources Created - ACR: socsocacr.azurecr.io (Operation, southeastasia) - Container App Environment: soc-cae (Operation, southeastasia) - Container App: soc-langgraph (Operation, southeastasia, port 2024) - Scale: 0-2 replicas, 0.5 vCPU / 1GB - Registry: socsocacr.azurecr.io (system-assigned managed identity) - OIDC: oidc-msi-8ac6 granted Contributor on Operation resource group ## What's included - Full langgraphjs-gen-ui-examples codebase under langgraph/ - Dockerfile: node:20-slim + pnpm, exposes port 2024 - .dockerignore, .gitignore - GitHub Actions: .github/workflows/deploy-langgraph.yml - Trigger: push to main, paths langgraph/** - Uses az acr build (cloud build, no local Docker needed) - Deploys to soc-langgraph Container App Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-10 04:01:06 +08:00
gongzhiyongandClaude Sonnet 4.6	c6ca9dc126	fix: restore workspace components accidentally dropped from git index ... workspace/ files existed on disk but were not included in previous incremental commit, causing git to record them as deleted. Re-adding all workspace card components, AgentWorkspace, ActivityTimeline, and WorkspaceCardRenderer to properly track them. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-10 03:09:41 +08:00
gongzhiyongandClaude Sonnet 4.6	47921d1a45	fix: resolve 6 gen-ui issues — timeline completeness, regen sync, localStorage persistence ... ## Fixes - [#3] Status events now add/update a status ActivityNode in timeline (upsertWsStatusNode) - [#4] Done callback adds done node; error callback adds error node + sets session.status="error" - [#5] handleRegenerate fully synced with workspace: status/tool/card/done/error all handled - [#1][#2] localStorage persistence: completed/error sessions auto-saved, lazy-loaded on demand - [#1] handleSelectConversation preloads workspace sessions from localStorage for history messages - Refactored completeWsSession to include done ActivityNode in timeline - Added errorWsSession, upsertWsStatusNode, saveWsToStorage, loadWsFromStorage helpers ## Known limitation - [#7] workspace_card merge:true not yet used by backend (all cards are append-only for now) - History workspace recovery depends on localStorage (browser-local, not cross-device) Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-10 03:08:23 +08:00
gongzhiyongandClaude Sonnet 4.6	0a2b339550	feat: implement Generative UI — Agent Workspace + workspace_card SSE protocol ... ## Backend - Add workspace_card SSE event protocol: {id, name, props, merge} - Add _extract_llm_text / _maybe_emit_workspace_card helpers in chat.py - Refactor all tools to dual-output format: {llm_text, ui: {name, props}} - kb_search → KnowledgeResultCard - ticket_list/detail → TicketSummaryCard / TicketDetailCard - web_search → SearchResultCard - generate_document → DocumentResultCard - sandbox_run → SandboxResultCard - Update SYSTEM_PROMPT: instruct LLM not to repeat tool data (UI shows it) ## Frontend - Three-column layout: sidebar + chat + Agent Workspace (360px right panel) - WorkspaceSession state model with ActivityNode + WorkspaceCard - New components/workspace/: AgentWorkspace, ActivityTimeline, WorkspaceCardRenderer - 6 card components: Knowledge/Ticket/Search/Document/Sandbox/ErrorCard - GeminiChat: workspace state management, SSE routing for workspace_card events - GeminiMessage: replace TracePanel with lightweight activity summary line - lib/api.ts: add WorkspaceSession/ActivityNode/WorkspaceCard types Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-10 02:57:09 +08:00
gongzhiyongandClaude Sonnet 4.6	1ad75a3602	fix: prevent duplicate AI message when switching conversations during stream ... Two bugs caused the same AI response to appear twice: 1. handleSelectConversation would fetch conversation history even while a stream was in progress (isLoading). If the fetch resolved mid-stream, fetchConversation overwrote the conversation messages unconditionally, writing the server-persisted AI message (with a real UUID) alongside the still-streaming client-side message (with a temp ID), producing a duplicate. Fix: guard the fetch with !isLoading, and add a second check inside the .then callback so setConversations only writes history if the messages array is still empty (protects against slow-resolving network requests). 2. onDone in streamChat could theoretically be called a second time if the ReadableStream completed without a "done" SSE event. Fix: call reader.cancel() immediately after onDone so the while loop exits cleanly. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-09 22:58:47 +08:00
gongzhiyongandClaude Sonnet 4.6	d86d9b8b32	fix(backend): handle MCP tool output format and read_url list input ... - Add _extract_output_str() to unwrap MCP ToolMessage content objects instead of calling str() on the raw object - Handle read_url url param when LLM passes a list instead of a string Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-09 22:46:58 +08:00
gongzhiyongandClaude Sonnet 4.6	fe8c361127	feat(input): replace model selector buttons with Select dropdown ... Replace the three pill buttons (Auto/Flash/Pro) in GeminiInput with a single shadcn/ui Select component. Selection logic and onSelectedModelChange callback are unchanged; only the UI widget is swapped. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>	2026-04-08 23:47:35 +08:00