feat: integrate 5 feature branches + daemon/job 命令层级化 + 跨平台后台引擎 + TypeScript 错误修复

Squashed merge of:
1. fix/mcp-tsc-errors — 修复上游 MCP 重构后的 tsc 错误和测试失败
2. feat/pipe-mute-disconnect — Pipe IPC 逻辑断开、/lang 命令、mute 状态机
3. feat/stub-recovery-all — 实现全部 stub 恢复 (task 001-012)
4. feat/kairos-activation — KAIROS 激活解除阻塞 + 工具实现
5. codex/openclaw-autonomy-pr — 自治权限系统、运行记录、managed flows

Additional:
6. daemon/job 命令层级化重构 (subcommand 架构)
7. 跨平台后台引擎抽象 (detached/tmux engines)
8. 修复 src/ 中 43 个预存在的 TypeScript 类型错误
9. 修复 langfuse isolated test mock 完整性
10. 修复 CodeRabbit 审查的 Critical/Major/Minor 问题
11. remote-control-server logger 抽象 (测试 stderr 静默化)
12. /simplify 审查修复 (代码复用、质量、效率)
This commit is contained in:
unraid
2026-04-14 14:19:00 +08:00
parent dad3ad2b8d
commit 637c9081f6
123 changed files with 13536 additions and 1887 deletions

2
.gitignore vendored
View File

@@ -15,7 +15,7 @@ src/utils/vendor/
.claude/
.codex/
.omx/
.docs/task/
# Binary / screenshot files (root only)
/*.png
*.bmp

204
02-kairos (1).md Normal file
View File

@@ -0,0 +1,204 @@
# KAIROS — 永不关机的 Claude
> 源码位置:`src/assistant/`、`src/proactive/`、`src/services/autoDream/`
> 编译开关:`feature('KAIROS')`、`feature('KAIROS_BRIEF')`、`feature('KAIROS_CHANNELS')`
> 远程开关GrowthBook `tengu_kairos`
关掉终端 Claude 还在运行的持久助手模式。KAIROS 是 Claude Code 中最复杂的隐藏功能之一。
---
## 核心概念
KAIROS 让 Claude 从"一次性对话工具"变成"持久运行的 AI 助手"
- 关闭终端后 Claude 仍在后台运行
- 每天自动写日志
- 晚上自动"做梦"整理记忆
- 没人说话时自己找活干
- 命令超 15 秒自动丢后台
---
## 激活流程
定义在 `src/main.tsx`(约第 1054-1092 行),需要通过五层检查:
```
1. feature('KAIROS') ← 编译时 flag
2. settings.assistant: true ← .claude/settings.json
3. 目录信任状态检查 ← 防恶意仓库劫持
4. tengu_kairos ← GrowthBook 远程开关
5. setKairosActive(true) ← 全局状态激活
```
`--assistant` CLI 参数可跳过远程开关检查(用于 Agent SDK daemon 模式)。
全局状态存储在 `src/bootstrap/state.ts`
- `kairosActive: boolean`(默认 `false`
- `getKairosActive()` / `setKairosActive(true)`
---
## 跨会话持久运行
### 会话恢复
`src/utils/conversationRecovery.ts` 中使用 `feature('KAIROS')` 条件导入 `BriefTool``SendUserFileTool`。在反序列化会话时识别这些工具的结果为"终端工具结果",判断 turn 是正常完成还是被中断。
### 持久 Cron 任务
关键在 `.claude/scheduled_tasks.json`。标记为 `permanent: true` 的任务不受 7 天自动过期限制:
- `catch-up`:恢复中断的工作
- `morning-checkin`:每日早间签到
- `dream`:记忆整合
### 会话历史 API
`src/assistant/sessionHistory.ts` 通过 OAuth API 加载远程会话历史,使用 `v1/sessions/{sessionId}/events` 端点,支持分页拉取。
---
## 做梦机制Dream
KAIROS 最精巧的子系统——后台运行的子代理,将分散的会话记忆整合为持久的结构化知识。
### 触发条件(三层门控,由廉到贵)
定义在 `src/services/autoDream/autoDream.ts`
```
1. 时间门控:距上次整合超过 24 小时minHours
2. 会话门控:至少 5 个新会话minSessions
3. 锁门控:没有其他进程正在整合
```
阈值通过 GrowthBook `tengu_onyx_plover` 远程配置动态控制。
### 四阶段整合流程
定义在 `src/services/autoDream/consolidationPrompt.ts`
| 阶段 | 动作 |
|------|------|
| **Orient** | 列出记忆目录、读取 `MEMORY.md` 索引、浏览已有主题文件 |
| **Gather** | 从每日日志、已有记忆、JSONL transcript 中搜集新信号 |
| **Consolidate** | 合并新信号到主题文件,转换相对日期为绝对日期,删除过时事实 |
| **Prune** | 更新 `MEMORY.md` 索引,保持在行数和大小限制内 |
### 锁机制
`src/services/autoDream/consolidationLock.ts`
- 使用 `.consolidate-lock` 文件
- 文件 mtime = `lastConsolidatedAt`
- 文件内容 = 持有者 PID
- 支持 PID 存活检查1 小时超时)
- double-write 后 re-read 验证防竞争
### 每日日志
路径由 `src/memdir/paths.ts``getAutoMemDailyLogPath()` 计算:
```
<autoMemPath>/logs/YYYY/MM/YYYY-MM-DD.md
```
### UI 呈现
- Footer pill 标签显示 **"dreaming"**
- `src/components/tasks/DreamDetailDialog.tsx` 提供专门的详情对话框
- 支持查看实时进度和手动中止
- `Shift+Down` 打开后台任务对话框
---
## 主动模式Proactive Mode
没人说话时 Claude 自己找活干。
### 核心状态
`src/proactive/index.ts` 维护三个状态:
| 状态 | 说明 |
|------|------|
| `active` | 是否激活 |
| `paused` | 是否暂停(用户按 Esc 取消时暂停,下次输入恢复) |
| `contextBlocked` | API 错误时阻塞 tick防止 tick-error-tick 死循环 |
### 激活方式
- `--proactive` CLI 参数
- `CLAUDE_CODE_PROACTIVE` 环境变量
-`feature('PROACTIVE') || feature('KAIROS')` 保护
### 系统提示
激活后追加:
```
# Proactive Mode
You are in proactive mode. Take initiative -- explore, act, and make progress
without waiting for instructions.
Start by briefly greeting the user.
You will receive periodic <tick> prompts. These are check-ins. Do whatever
seems most useful, or call Sleep if there's nothing to do.
```
### SleepTool 集成
设置中的 `minSleepDurationMs``maxSleepDurationMs` 控制 Sleep 持续时间范围,节流 proactive tick 频率。没活干就 Sleep 等着。
---
## 后台任务管理
### Cron 调度器
`src/utils/cronScheduler.ts`
- 每 1 秒 tick 一次(`CHECK_INTERVAL_MS = 1000`
- 使用 chokidar 监视 `.claude/scheduled_tasks.json`
- 支持调度器锁(`src/utils/cronTasksLock.ts`),防止多实例重复触发
- 锁探测间隔 5 秒,持有者崩溃时自动接管
### 任务类型
| 类型 | 说明 |
|------|------|
| 一次性(`recurring: false` | 触发后自动删除,支持错过任务检测 |
| 循环(`recurring: true` | 触发后重新调度,默认 7 天过期 |
| 永久(`permanent: true` | 不受过期限制KAIROS 专用) |
| 会话级(`durable: false` | 仅内存中,进程退出即消失 |
### Jitter 防雷群机制
`src/utils/cronJitterConfig.ts`
- 循环任务:基于 taskId 的确定性延迟interval 的 10%,上限 15 分钟)
- 一次性任务:在 :00 和 :30 施加最多 90 秒提前量
- 运维可在事故期间推送配置变更60 秒内全客户端生效
---
## 关键源码文件
| 文件 | 职责 |
|------|------|
| `src/bootstrap/state.ts` | KAIROS 全局状态 |
| `src/assistant/index.ts` | 助手模式入口 |
| `src/assistant/sessionHistory.ts` | 远程会话历史 API |
| `src/proactive/index.ts` | 主动模式状态管理 |
| `src/services/autoDream/autoDream.ts` | Auto-Dream 引擎 |
| `src/services/autoDream/consolidationPrompt.ts` | 整合提示(四阶段) |
| `src/services/autoDream/consolidationLock.ts` | 整合锁 |
| `src/services/autoDream/config.ts` | Dream 配置 |
| `src/tasks/DreamTask/DreamTask.ts` | Dream 任务定义 |
| `src/utils/cronScheduler.ts` | Cron 调度器 |
| `src/utils/cronTasks.ts` | Cron 任务持久化 |
| `src/skills/bundled/dream.ts` | `/dream` Skill存根 |

283
AGENTS.md Normal file
View File

@@ -0,0 +1,283 @@
# AGENTS.md
This file provides guidance to Codex (Codex.ai/code) when working with code in this repository.
## Project Overview
This is a **reverse-engineered / decompiled** version of Anthropic's official Codex CLI tool. The goal is to restore core functionality while trimming secondary capabilities. Many modules are stubbed or feature-flagged off. TypeScript strict mode is enforced — **`bunx tsc --noEmit` must pass with zero errors**.
## Git Commit Message Convention
使用 **Conventional Commits** 规范:
```
<type>: <描述>
```
常见 type`feat``fix``docs``chore``refactor`
示例:
- `feat: 添加模型 1M 上下文切换`
- `fix: 修复初次登陆的校验问题`
- `chore: remove prefetchOfficialMcpUrls call on startup`
## Commands
```bash
# Install dependencies
bun install
# Dev mode (runs cli.tsx with MACRO defines injected via -d flags)
bun run dev
# Dev mode with debugger (set BUN_INSPECT=9229 to pick port)
bun run dev:inspect
# Pipe mode
echo "say hello" | bun run src/entrypoints/cli.tsx -p
# Build (code splitting, outputs dist/cli.js + chunk files)
bun run build
# Test
bun test # run all tests (2453 tests / 137 files / 0 fail)
bun test src/utils/__tests__/hash.test.ts # run single file
bun test --coverage # with coverage report
# Lint & Format (Biome)
bun run lint # check only
bun run lint:fix # auto-fix
bun run format # format all src/
# Health check
bun run health
# Check unused exports
bun run check:unused
# Remote Control Server
bun run rcs
# Docs dev server (Mintlify)
bun run docs:dev
```
详细的测试规范、覆盖状态和改进计划见 `docs/testing-spec.md`
## Architecture
### Runtime & Build
- **Runtime**: Bun (not Node.js). All imports, builds, and execution use Bun APIs.
- **Build**: `build.ts` 执行 `Bun.build()` with `splitting: true`,入口 `src/entrypoints/cli.tsx`,输出 `dist/cli.js` + chunk files。Build 默认启用 19 个 feature见下方 Feature Flag 段)。构建后自动替换 `import.meta.require` 为 Node.js 兼容版本(产物 bun/node 都可运行)。
- **Dev mode**: `scripts/dev.ts` 通过 Bun `-d` flag 注入 `MACRO.*` defines运行 `src/entrypoints/cli.tsx`。默认启用全部 feature。
- **Module system**: ESM (`"type": "module"`), TSX with `react-jsx` transform.
- **Monorepo**: Bun workspaces — 14 个 internal packages in `packages/` resolved via `workspace:*`
- **Lint/Format**: Biome (`biome.json`)。`bun run lint` / `bun run lint:fix` / `bun run format`
- **Defines**: 集中管理在 `scripts/defines.ts`。当前版本 `2.1.888`
- **CI**: GitHub Actions — `ci.yml`(构建+测试)、`release-rcs.yml`RCS 发布)、`update-contributors.yml`(自动更新贡献者)。
### Entry & Bootstrap
1. **`src/entrypoints/cli.tsx`** (323 行) — True entrypoint。`main()` 函数按优先级处理多条快速路径:
- `--version` / `-v` — 零模块加载
- `--dump-system-prompt` — feature-gated (DUMP_SYSTEM_PROMPT)
- `--Codex-in-chrome-mcp` / `--chrome-native-host`
- `--computer-use-mcp` — 独立 MCP server 模式
- `--daemon-worker=<kind>` — feature-gated (DAEMON)
- `remote-control` / `rc` / `remote` / `sync` / `bridge` — feature-gated (BRIDGE_MODE)
- `daemon` [subcommand] — feature-gated (DAEMON)
- `ps` / `logs` / `attach` / `kill` / `--bg` — feature-gated (BG_SESSIONS)
- `new` / `list` / `reply` — Template job commands
- `environment-runner` / `self-hosted-runner` — BYOC runner
- `--tmux` + `--worktree` 组合
- 默认路径:加载 `main.tsx` 启动完整 CLI
2. **`src/main.tsx`** (~6970 行) — Commander.js CLI definition。注册大量 subcommands`mcp` (serve/add/remove/list...)、`server``ssh``open``auth``plugin``agents``auto-mode``doctor``update` 等。主 `.action()` 处理器负责权限、MCP、会话恢复、REPL/Headless 模式分发。
3. **`src/entrypoints/init.ts`** — One-time initialization (telemetry, config, trust dialog)。
### Core Loop
- **`src/query.ts`** — The main API query function. Sends messages to Codex API, handles streaming responses, processes tool calls, and manages the conversation turn loop.
- **`src/QueryEngine.ts`** — Higher-level orchestrator wrapping `query()`. Manages conversation state, compaction, file history snapshots, attribution, and turn-level bookkeeping. Used by the REPL screen.
- **`src/screens/REPL.tsx`** — The interactive REPL screen (React/Ink component). Handles user input, message display, tool permission prompts, and keyboard shortcuts.
### API Layer
- **`src/services/api/Codex.ts`** — Core API client. Builds request params (system prompt, messages, tools, betas), calls the Anthropic SDK streaming endpoint, and processes `BetaRawMessageStreamEvent` events.
- **7 providers**: `firstParty` (Anthropic direct), `bedrock` (AWS), `vertex` (Google Cloud), `foundry`, `openai`, `gemini`, `grok` (xAI)。
- Provider selection in `src/utils/model/providers.ts`。优先级modelType 参数 > 环境变量 > 默认 firstParty。
### Tool System
- **`src/Tool.ts`** — Tool interface definition (`Tool` type) and utilities (`findToolByName`, `toolMatchesName`).
- **`src/tools.ts`** (387 行) — Tool registry. Assembles the tool list; some tools are conditionally loaded via `feature()` flags or `process.env.USER_TYPE`.
- **`src/tools/<ToolName>/`** — 55 个 tool 目录。主要分类:
- **文件操作**: FileEditTool, FileReadTool, FileWriteTool, GlobTool, GrepTool
- **Shell/执行**: BashTool, PowerShellTool, REPLTool
- **Agent 系统**: AgentTool, TaskCreateTool, TaskUpdateTool, TaskListTool, TaskGetTool
- **规划**: EnterPlanModeTool, ExitPlanModeV2Tool, VerifyPlanExecutionTool
- **Web/MCP**: WebFetchTool, WebSearchTool, MCPTool, McpAuthTool
- **调度**: CronCreateTool, CronDeleteTool, CronListTool
- **其他**: LSPTool, ConfigTool, SkillTool, EnterWorktreeTool, ExitWorktreeTool 等
- **`src/tools/shared/`** — Tool 共享工具函数。
### UI Layer (Ink)
- **`src/ink.ts`** — Ink render wrapper with ThemeProvider injection.
- **`packages/@ant/ink/`** — Custom Ink frameworkforked/internal包含 components、core、hooks、keybindings、theme、utils。注意不是 `src/ink/`
- **`src/components/`** — 149 个组件目录/文件,渲染于终端 Ink 环境中。关键组件:
- `App.tsx` — Root provider (AppState, Stats, FpsMetrics)
- `Messages.tsx` / `MessageRow.tsx` — Conversation message rendering
- `PromptInput/` — User input handling
- `permissions/` — Tool permission approval UI
- `design-system/` — 复用 UI 组件Dialog, FuzzyPicker, ProgressBar, ThemeProvider 等)
- Components use React Compiler runtime (`react/compiler-runtime`) — decompiled output has `_c()` memoization calls throughout.
### State Management
- **`src/state/AppState.tsx`** — Central app state type and context provider. Contains messages, tools, permissions, MCP connections, etc.
- **`src/state/AppStateStore.ts`** — Default state and store factory.
- **`src/state/store.ts`** — Zustand-style store for AppState (`createStore`).
- **`src/state/selectors.ts`** — State selectors.
- **`src/bootstrap/state.ts`** — Module-level singletons for session-global state (session ID, CWD, project root, token counts, model overrides, client type, permission mode).
### Workspace Packages
| Package | 说明 |
|---------|------|
| `packages/@ant/ink/` | Forked Ink 框架components、hooks、keybindings、theme |
| `packages/@ant/computer-use-mcp/` | Computer Use MCP server截图/键鼠/剪贴板/应用管理) |
| `packages/@ant/computer-use-input/` | 键鼠模拟dispatcher + darwin/win32/linux backend |
| `packages/@ant/computer-use-swift/` | 截图 + 应用管理dispatcher + per-platform backend |
| `packages/@ant/Codex-for-chrome-mcp/` | Chrome 浏览器控制(通过 `--chrome` 启用) |
| `packages/remote-control-server/` | 自托管 Remote Control ServerDocker 部署,含 Web UI |
| `packages/swarm/` | Swarm 解耦模块 |
| `packages/shell/` | Shell 抽象 |
| `packages/audio-capture-napi/` | 原生音频捕获(已恢复) |
| `packages/color-diff-napi/` | 颜色差异计算完整实现11 tests |
| `packages/image-processor-napi/` | 图像处理(已恢复) |
| `packages/modifiers-napi/` | 键盘修饰键检测stub |
| `packages/url-handler-napi/` | URL scheme 处理stub |
### Bridge / Remote Control
- **`src/bridge/`** (~37 files) — Remote Control / Bridge 模式。feature-gated by `BRIDGE_MODE`。包含 bridge API、会话管理、JWT 认证、消息传输、权限回调等。Entry: `bridgeMain.ts`
- **`packages/remote-control-server/`** — 自托管 RCS支持 Docker 部署,含 Web UI 控制面板。通过 `bun run rcs` 启动。
- CLI 快速路径: `Codex remote-control` / `Codex rc` / `Codex bridge`
- 详见 `docs/features/remote-control-self-hosting.md`
### Daemon Mode
- **`src/daemon/`** — Daemon 模式(长驻 supervisor。feature-gated by `DAEMON`。包含 `main.ts`entry`workerRegistry.ts`worker 管理)。
### Context & System Prompt
- **`src/context.ts`** — Builds system/user context for the API call (git status, date, AGENTS.md contents, memory files).
- **`src/utils/claudemd.ts`** — Discovers and loads AGENTS.md files from project hierarchy.
### Feature Flag System
Feature flags control which functionality is enabled at runtime. 代码中统一通过 `import { feature } from 'bun:bundle'` 导入,调用 `feature('FLAG_NAME')` 返回 `boolean`
**启用方式**: 环境变量 `FEATURE_<FLAG_NAME>=1`。例如 `FEATURE_BUDDY=1 bun run dev`
**Build 默认 features**19 个,见 `build.ts`:
- 基础: `BUDDY`, `TRANSCRIPT_CLASSIFIER`, `BRIDGE_MODE`, `AGENT_TRIGGERS_REMOTE`, `CHICAGO_MCP`, `VOICE_MODE`
- 统计/缓存: `SHOT_STATS`, `PROMPT_CACHE_BREAK_DETECTION`, `TOKEN_BUDGET`
- P0 本地: `AGENT_TRIGGERS`, `ULTRATHINK`, `BUILTIN_EXPLORE_PLAN_AGENTS`, `LODESTONE`
- P1 API 依赖: `EXTRACT_MEMORIES`, `VERIFICATION_AGENT`, `KAIROS_BRIEF`, `AWAY_SUMMARY`, `ULTRAPLAN`
- P2: `DAEMON`
**Dev mode 默认**: 全部启用(见 `scripts/dev.ts`)。
**类型声明**: `src/types/internal-modules.d.ts` 中声明了 `bun:bundle` 模块的 `feature` 函数签名。
**新增功能的正确做法**: 保留 `import { feature } from 'bun:bundle'` + `feature('FLAG_NAME')` 的标准模式,在运行时通过环境变量或配置控制,不要绕过 feature flag 直接 import。
### Multi-API 兼容层
所有兼容层均采用流适配器模式:将第三方 API 格式转为 Anthropic 内部格式,下游代码完全不改。
#### OpenAI 兼容层
通过 `CLAUDE_CODE_USE_OPENAI=1` 启用,支持 Ollama/DeepSeek/vLLM 等任意 OpenAI Chat Completions 协议端点。含 DeepSeek thinking mode 支持。
- **`src/services/api/openai/`** — client、消息/工具转换、流适配、模型映射
- 关键环境变量:`CLAUDE_CODE_USE_OPENAI``OPENAI_API_KEY``OPENAI_BASE_URL``OPENAI_MODEL`
#### Gemini 兼容层
通过 `CLAUDE_CODE_USE_GEMINI=1` 启用。独立环境变量体系。
- **`src/services/api/gemini/`** — client、模型映射、类型定义
- 关键环境变量:`GEMINI_API_KEY`(必填)、`GEMINI_MODEL`(直接指定)、`GEMINI_DEFAULT_SONNET_MODEL`/`GEMINI_DEFAULT_OPUS_MODEL`(按能力映射)
- 模型映射优先级:`GEMINI_MODEL` > `GEMINI_DEFAULT_*_MODEL` > `ANTHROPIC_DEFAULT_*_MODEL`(已废弃) > 原样返回
#### Grok 兼容层
通过 `CLAUDE_CODE_USE_GROK=1` 启用。自定义模型映射支持 xAI Grok API。
- **`src/services/api/grok/`** — client、模型映射
详见各兼容层的 docs 文档。
### Stubbed/Deleted Modules
| Module | Status |
|--------|--------|
| Computer Use (`@ant/*`) | Restored — macOS + Windows + Linux后端完整度不一 |
| `*-napi` packages | `audio-capture-napi``image-processor-napi` 已恢复;`color-diff-napi` 完整;`modifiers-napi``url-handler-napi` 仍为 stub |
| Voice Mode | Restored — Push-to-Talk 语音输入(需 Anthropic OAuth |
| OpenAI/Gemini/Grok 兼容层 | Restored |
| Remote Control Server | Restored — 自托管 RCS + Web UI |
| Analytics / GrowthBook / Sentry | Empty implementations |
| Magic Docs / LSP Server | Removed |
| Plugins / Marketplace | Removed |
| MCP OAuth | Simplified |
### Key Type Files
- **`src/types/global.d.ts`** — Declares `MACRO`, `BUILD_TARGET`, `BUILD_ENV` and internal Anthropic-only identifiers.
- **`src/types/internal-modules.d.ts`** — Type declarations for `bun:bundle`, `bun:ffi`, `@anthropic-ai/mcpb`.
- **`src/types/message.ts`** — Message type hierarchy (UserMessage, AssistantMessage, SystemMessage, etc.).
- **`src/types/permissions.ts`** — Permission mode and result types.
## Testing
- **框架**: `bun:test`(内置断言 + mock
- **当前状态**: 2472 tests / 138 files / 0 fail
- **单元测试**: 就近放置于 `src/**/__tests__/`,文件名 `<module>.test.ts`
- **集成测试**: `tests/integration/` — 4 个文件cli-arguments, context-build, message-pipeline, tool-chain
- **共享 mock/fixture**: `tests/mocks/`api-responses, file-system, fixtures/
- **命名**: `describe("functionName")` + `test("behavior description")`,英文
- **Mock 模式**: 对重依赖模块使用 `mock.module()` + `await import()` 解锁(必须内联在测试文件中,不能从共享 helper 导入)
- **包测试**: `packages/` 下各包也有独立测试(如 `color-diff-napi` 11 tests
### 类型检查
项目使用 TypeScript strict 模式,**tsc 必须零错误**。每次修改后运行:
```bash
bunx tsc --noEmit
```
**类型规范**
- 生产代码禁止 `as any`;测试文件中 mock 数据可用 `as any`
- 类型不匹配优先用 `as unknown as SpecificType` 双重断言,或补充 interface
- 未知结构对象用 `Record<string, unknown>` 替代 `any`
- 联合类型用类型守卫type guard收窄不要强转
- `msg.request` 属性访问:`const req = msg.request as Record<string, unknown>`
- Ink `color` prop`as keyof Theme` 而非 `as any`
## Working with This Codebase
- **tsc must pass** — `bunx tsc --noEmit` 必须零错误,任何修改都不能引入新的类型错误。
- **Feature flags** — 默认全部关闭(`feature()` 返回 `false`。Dev/build 各有自己的默认启用列表。不要在 `cli.tsx` 中重定义 `feature` 函数。
- **React Compiler output** — Components have decompiled memoization boilerplate (`const $ = _c(N)`). This is normal.
- **`bun:bundle` import** — `import { feature } from 'bun:bundle'` 是 Bun 内置模块,由运行时/构建器解析。不要用自定义函数替代它。**`feature()` 只能直接用在 `if` 语句或三元表达式的条件位置**Bun 编译器限制),不能赋值给变量、不能放在箭头函数体里、不能作为 `&&` 链的一部分。正确:`if (feature('X')) {}``feature('X') ? a : b`
- **`src/` path alias** — tsconfig maps `src/*` to `./src/*`. Imports like `import { ... } from 'src/utils/...'` are valid.
- **MACRO defines** — 集中管理在 `scripts/defines.ts`。Dev mode 通过 `bun -d` 注入build 通过 `Bun.build({ define })` 注入。修改版本号等常量只改这个文件。
- **构建产物兼容 Node.js** — `build.ts` 会自动后处理 `import.meta.require`,产物可直接用 `node dist/cli.js` 运行。
- **Biome 配置** — 大量 lint 规则被关闭decompiled 代码不适合严格 lint`.tsx` 文件用 120 行宽 + 强制分号;其他文件 80 行宽 + 按需分号。
- **Ink 框架在 `packages/@ant/ink/`** — 不是 `src/ink/`该目录不存在。Ink 相关的组件、hooks、keybindings 都在 packages 中。
- **Provider 优先级** — `modelType` 参数 > 环境变量 > 默认 `firstParty`。新增 provider 需在 `src/utils/model/providers.ts` 注册。

View File

@@ -40,6 +40,8 @@ const DEFAULT_BUILD_FEATURES = [
'KAIROS',
'COORDINATOR_MODE',
'LAN_PIPES',
'BG_SESSIONS',
'TEMPLATES',
// 'REVIEW_ARTIFACT', // API 请求无响应,需进一步排查 schema 兼容性
// P3: poor mode (disable extract_memories + prompt_suggestion)
'POOR',

View File

@@ -0,0 +1,318 @@
# Daemon 重构设计方案
> 分支: `feat/integrate-5-branches`
> 基于: `f41745cb` (= main `11bb3f62` 内容)
> 日期: 2026-04-13
## 一、问题概述
### 1.1 命令结构散乱
当前后台进程相关的命令分布在三个不同的位置,没有统一的命名空间:
| 命令 | 注册位置 | 入口 |
|------|---------|------|
| `claude daemon start/status/stop` | `cli.tsx` 快速路径 L203 | `daemon/main.ts` |
| `claude ps` | `cli.tsx` 快速路径 L220 | `cli/bg.ts` |
| `claude logs <x>` | `cli.tsx` 快速路径 L232 | `cli/bg.ts` |
| `claude attach <x>` | `cli.tsx` 快速路径 L236 | `cli/bg.ts` |
| `claude kill <x>` | `cli.tsx` 快速路径 L238 | `cli/bg.ts` |
| `claude --bg` | `cli.tsx` 快速路径 L244 | `cli/bg.ts` |
| `claude new/list/reply` | `cli.tsx` 快速路径 L250 | `cli/handlers/templateJobs.ts` |
| `claude rollback` | `main.tsx` Commander.js L6525 | `cli/rollback.ts` |
| `claude up` | `main.tsx` Commander.js L6511 | `cli/up.ts` |
**问题**:
- `ps/logs/attach/kill``daemon` 逻辑上都是后台进程管理,但互不关联
- 这些命令都**只有 CLI 入口**REPL 里输入 `/daemon``/ps` 不存在
- `new/list/reply` 是模板任务系统的顶级命令,容易与其他命令冲突(特别是 `list`
### 1.2 Windows 不支持
`--bg``attach` 硬依赖 tmux
- `bg.ts:handleBgFlag()` 第一步就检查 tmux不可用直接报错退出
- `bg.ts:attachHandler()``tmux attach-session`,无 tmux 替代方案
- Windows (包括 VS Code 终端) 完全无法使用后台会话功能
### 1.3 无 REPL 入口
对比 `/mcp` 的双注册模式:
- **CLI**: `claude mcp serve/add/remove/list` (Commander.js, `main.tsx:5760`)
- **REPL**: `/mcp enable/disable/reconnect` (slash command, `commands/mcp/index.ts`)
`daemon`/`bg`/`job` 系列只有 CLI 快速路径REPL 中完全不可用。
## 二、目标
1. **层级化命令结构**: 参照 `/mcp` 模式,将后台管理收归 `/daemon`,模板任务收归 `/job`
2. **跨平台后台会话**: Windows / macOS / Linux 都能启动、附着、终止后台会话
3. **双注册**: CLI (`claude daemon ...`) + REPL (`/daemon ...`) 同时可用
4. **向后兼容**: 旧命令保留但输出 deprecation 提示
## 三、命令结构设计
### 3.1 `/daemon` — 后台进程管理
合并 daemon supervisor + bg sessions 为统一命名空间:
```
claude daemon <subcommand> ← CLI 入口 (cli.tsx 快速路径)
/daemon <subcommand> ← REPL 入口 (slash command, local-jsx)
子命令:
status 综合状态面板 (daemon + 所有会话)
start [--dir <path>] 启动 daemon supervisor
stop 停止 daemon
bg [args...] 启动后台会话
attach [target] 附着到后台会话
logs [target] 查看会话日志
kill [target] 终止会话
(无参数) 等同于 status
```
**CLI 快速路径路由** (`cli.tsx`):
```typescript
// 新: 统一入口
if (feature('DAEMON') && args[0] === 'daemon') {
const sub = args[1] || 'status'
switch (sub) {
case 'start': case 'stop': case 'status':
await daemonMain([sub, ...args.slice(2)])
break
case 'bg':
await bg.handleBgStart(args.slice(2))
break
case 'attach': case 'logs': case 'kill':
await bg[`${sub}Handler`](args[2])
break
}
}
// 向后兼容 (deprecated)
if (feature('BG_SESSIONS') && ['ps','logs','attach','kill'].includes(args[0])) {
console.warn(`[deprecated] Use: claude daemon ${args[0] === 'ps' ? 'status' : args[0]}`)
// ... delegate to daemon subcommand
}
```
**REPL 斜杠命令** (`commands/daemon/index.ts`):
```typescript
const daemon = {
type: 'local-jsx',
name: 'daemon',
description: 'Manage background sessions and daemon',
argumentHint: '[status|start|stop|bg|attach|logs|kill]',
isEnabled: () => feature('DAEMON') || feature('BG_SESSIONS'),
load: () => import('./daemon.js'),
} satisfies Command
```
### 3.2 `/job` — 模板任务管理
```
claude job <subcommand> ← CLI 入口
/job <subcommand> ← REPL 入口
子命令:
list 列出模板和活跃任务
new <template> [args] 从模板创建任务
reply <id> <text> 回复任务
status <id> 查看任务状态
(无参数) 等同于 list
```
### 3.3 独立命令 (不变)
```
claude up 保持顶级 (简短的 bootstrap 命令)
claude rollback [target] 保持顶级 (低频运维命令)
```
## 四、跨平台后台引擎
### 4.1 引擎抽象
```typescript
// src/cli/bg/engine.ts
export interface BgEngine {
readonly name: string
/** 当前平台是否可用 */
available(): Promise<boolean>
/** 启动后台会话 */
start(opts: BgStartOptions): Promise<BgStartResult>
/** 附着到后台会话blocking */
attach(session: SessionEntry): Promise<void>
}
export interface BgStartOptions {
sessionName: string
args: string[]
env: Record<string, string | undefined>
logPath: string
cwd: string
}
export interface BgStartResult {
pid: number
sessionName: string
logPath: string
engineUsed: string
}
```
### 4.2 三种引擎实现
| 引擎 | 平台 | 启动方式 | attach 方式 |
|------|------|---------|------------|
| TmuxEngine | macOS/Linux (有 tmux) | `tmux new-session -d` | `tmux attach-session` |
| DetachedEngine | Windows / 无 tmux 的 macOS/Linux | `spawn({ detached, stdio→logFile })` | `tail -f` 日志文件 |
#### DetachedEngine 详细设计
**启动 (`start`)**:
```typescript
// 1. 打开日志文件 fd
const logFd = fs.openSync(logPath, 'a')
// 2. detached spawn, stdout/stderr 重定向到日志
const child = spawn(process.execPath, execArgs, {
detached: true,
stdio: ['ignore', logFd, logFd],
env,
cwd,
})
child.unref()
fs.closeSync(logFd)
// 3. 写 sessions/<PID>.json
```
**附着 (`attach`)**:
```typescript
// 跨平台 tail -f 实现
// 1. 读取已有日志内容输出到 stdout
// 2. fs.watch(logPath) 监听变化
// 3. 每次变化读取新增内容
// 4. Ctrl+C 退出 tail不杀后台进程
```
#### 引擎选择逻辑
```typescript
// src/cli/bg/engines/index.ts
export async function selectEngine(): Promise<BgEngine> {
if (process.platform === 'win32') {
return new DetachedEngine()
}
const tmux = new TmuxEngine()
if (await tmux.available()) {
return tmux
}
return new DetachedEngine()
}
```
### 4.3 SessionEntry 扩展
```typescript
interface SessionEntry {
// ... 现有字段
engine: 'tmux' | 'detached' // 新增: 记录使用的引擎
tmuxSessionName?: string // tmux 引擎才有
logPath?: string // 两种引擎都有
}
```
`attach` 时根据 `session.engine` 选择对应的 attach 策略。
## 五、文件变更清单
### 新增文件 (10 个)
```
src/cli/bg/engine.ts BgEngine 接口定义
src/cli/bg/engines/tmux.ts TmuxEngine (从 bg.ts 提取)
src/cli/bg/engines/detached.ts DetachedEngine (新实现)
src/cli/bg/engines/index.ts 引擎选择 + re-export
src/cli/bg/tail.ts 跨平台日志 tail (用于 detached attach)
src/commands/daemon/index.ts /daemon REPL 斜杠命令注册
src/commands/daemon/daemon.tsx /daemon 子命令路由 + status UI
src/commands/job/index.ts /job REPL 斜杠命令注册
src/commands/job/job.tsx /job 子命令路由 + UI
docs/features/daemon-restructure-design.md 本设计文档
```
### 修改文件 (6 个)
```
src/cli/bg.ts 重构: handler 函数改为调用 BgEngine
src/entrypoints/cli.tsx 快速路径: daemon 统一入口 + 向后兼容
src/commands.ts 注册 /daemon 和 /job 斜杠命令
src/daemon/main.ts daemonMain() 增加 bg/ps/logs 子命令分发
src/main.tsx Commander.js: 可选注册 daemon/job 子命令
src/cli/handlers/templateJobs.ts 适配 /job 入口 (可能不需改)
```
### 不动的文件
```
src/daemon/state.ts daemon PID 状态管理 (无需改)
src/jobs/state.ts job 状态管理 (无需改)
src/jobs/templates.ts 模板发现 (无需改)
src/jobs/classifier.ts 任务分类器 (无需改)
src/cli/rollback.ts 保持顶级命令 (无需改)
src/cli/up.ts 保持顶级命令 (无需改)
```
## 六、可行性分析
### 6.1 风险评估
| 风险 | 级别 | 缓解措施 |
|------|------|---------|
| cli.tsx 快速路径修改影响启动性能 | 低 | 仅改路由逻辑import 仍然 lazy |
| DetachedEngine 的 attach 在 Windows 上 fs.watch 不可靠 | 中 | 使用轮询 fallback (setInterval + fs.stat) |
| 向后兼容的 deprecation 可能破坏脚本 | 低 | 旧命令保持可用,仅输出 stderr 警告 |
| REPL 中 /daemon bg 需要 spawn 子进程 | 中 | 参考 /assistant 的 NewInstallWizard (已有 spawn 先例) |
| tsc 类型兼容 | 低 | 接口定义清晰,不引入 any |
### 6.2 工作量估计
| Task | 文件数 | 复杂度 |
|------|--------|--------|
| Task 013: BgEngine 抽象 + 引擎实现 | 5 新增 + 1 修改 | 中 |
| Task 014: /daemon 命令层级化 | 3 新增 + 3 修改 | 中 |
| Task 015: /job 命令层级化 | 2 新增 + 2 修改 | 低 |
| Task 016: 向后兼容 + 测试 | 0 新增 + 2 修改 | 低 |
### 6.3 依赖关系
```
Task 013 (BgEngine) ← 无依赖,可独立开发
Task 014 (/daemon) ← 依赖 Task 013 (引擎选择)
Task 015 (/job) ← 无依赖,可与 013 并行
Task 016 (兼容) ← 依赖 Task 014 + 015
```
## 七、设计决策记录
### D1: 为什么 daemon + bg sessions 合为一个命名空间?
用户视角:都是"后台运行的东西"。分开会导致 `claude daemon status` 看 supervisor + `claude ps` 看会话,割裂感强。合并后 `claude daemon status` 一次性展示 supervisor 状态 + 所有会话列表。
### D2: 为什么 rollback/up 不收入 daemon
它们本质是**版本管理/环境初始化**,不是后台进程管理。`claude up` 是同步阻塞的 setup 脚本,不涉及 daemon 或后台会话。保持顶级更直观。
### D3: 为什么 DetachedEngine 的 attach 用 tail 而不是 IPC
1. 日志文件是最简单的跨平台方案,无需额外依赖
2. UDS Pipe IPC 系统 (usePipeIpc) 设计用于实例间通信,不是终端附着
3. tmux attach 的体验(完整 PTY无法在纯 detached 模式下复制tail 是最诚实的替代
### D4: 为什么不用 Windows Terminal 的 tab/pane API
Windows Terminal 的 `wt.exe` 新窗口/标签功能不够通用——用户可能在 VS Code、ConEmu、cmder 等终端中。detached + log 是唯一跨终端方案。

View File

@@ -0,0 +1,310 @@
# Stub 恢复设计 1-4
> 日期2026-04-12
> 目标:基于当前代码边界,为下一阶段 4 个 stub/半 stub 命令面给出可实施的设计方案。
> 排序原则:按建议实施顺序排序,不按问题严重性排序。
## 设计原则
- 先做能独立闭环、收益明确、改动边界清晰的项。
- 大项拆成 `MVP``Phase 2+`,避免一次性掉进大范围恢复。
- 优先复用已有状态、传输层、日志与配置能力,不重造协议。
- 设计以当前仓库实际代码为准,不以旧文档的理想状态为准。
## 1. `claude daemon status` / `claude daemon stop`
### 现状
- `start` 路径已有完整 supervisor + worker 生命周期:
[src/daemon/main.ts](</e:/Source_code/Claude-code-bast/src/daemon/main.ts:1>)
[src/daemon/workerRegistry.ts](</e:/Source_code/Claude-code-bast/src/daemon/workerRegistry.ts:1>)
- `status` / `stop` 目前只是占位输出:
[src/daemon/main.ts](</e:/Source_code/Claude-code-bast/src/daemon/main.ts:49>)
- `/remote-control-server` 有自己的命令内 UI 状态,但只维护当前进程内的 `daemonProcess`,并不适合作为跨进程 CLI 管理基础:
[src/commands/remoteControlServer/remoteControlServer.tsx](</e:/Source_code/Claude-code-bast/src/commands/remoteControlServer/remoteControlServer.tsx:32>)
### 目标
-`claude daemon status``claude daemon stop` 在另一个 CLI 进程中也能正确工作。
- 不依赖 TUI 内存态,不要求当前命令进程就是启动 daemon 的那个进程。
### MVP 方案
- 新增 daemon 状态文件,例如:
`~/.claude/daemon/remote-control.json`
- `start` 时写入:
- supervisor pid
- cwd
- startedAt
- worker kinds
- 最近状态
- `status`
- 读取状态文件
- 用现有进程探测能力验证 pid 是否存活
- 输出 `running / stopped / stale`
- stale 时自动清理状态文件
- `stop`
- 读取 pid
- 发送 `SIGTERM`
- 等待退出
- 超时后 `SIGKILL`
- 清理状态文件
### 代码范围
- 新增 `src/daemon/state.ts`
- 修改 [src/daemon/main.ts](</e:/Source_code/Claude-code-bast/src/daemon/main.ts:1>)
- 轻量修改 [src/commands/remoteControlServer/remoteControlServer.tsx](</e:/Source_code/Claude-code-bast/src/commands/remoteControlServer/remoteControlServer.tsx:32>),让 UI 尽量读取同一份状态文件
### 验证
1. `claude daemon start`
2. 新开终端执行 `claude daemon status`
3. 执行 `claude daemon stop`
4. 再次执行 `claude daemon status`,确认返回 `stopped` 或清晰的 `stale cleaned`
### 风险
- Windows 信号模型和 Unix 不同,`stop` 需要超时兜底。
- 当前设计默认单 supervisor不处理多实例并发。
### 工作量判断
-
- 适合作为下一步的首选实现项
## 2. `BG_SESSIONS`
### 现状
- fast-path 已接好:
[src/entrypoints/cli.tsx](</e:/Source_code/Claude-code-bast/src/entrypoints/cli.tsx:218>)
- session registry 已有真实实现:
[src/utils/concurrentSessions.ts](</e:/Source_code/Claude-code-bast/src/utils/concurrentSessions.ts:1>)
- `exit` 在 bg session 内已会 `tmux detach-client`
[src/commands/exit/exit.tsx](</e:/Source_code/Claude-code-bast/src/commands/exit/exit.tsx:20>)
- 但 CLI handler 仍全空:
[src/cli/bg.ts](</e:/Source_code/Claude-code-bast/src/cli/bg.ts:1>)
- task summary 仍然是 stub
[src/utils/taskSummary.ts](</e:/Source_code/Claude-code-bast/src/utils/taskSummary.ts:1>)
### 目标
- 先把 `ps` / `logs` / `kill` 做成真正有用的 session 管理命令。
- 不在第一阶段就强行补完 `attach` / `--bg`
### Phase 2AMVP
- 实现 `ps`
- 从 registry 读取 live sessions
- 展示 pid、kind、sessionId、cwd、name、startedAt、bridgeSessionId
- 如果有 activity/status则一并展示
- 实现 `logs`
- 支持按 `sessionId / pid / name` 查找
- 优先复用本地 transcript/log 读取能力
- 如果 registry 里存在 `logPath`,支持 tail 文件
- 实现 `kill`
- 解析目标 session
- 发退出信号
- 清理 stale registry
### Phase 2B后续
- 实现 `attach`
- 实现 `--bg`
- 实现 `taskSummary` 的中途状态更新
### 为什么要拆
- 现有 registry 记录了 `pid / sessionId / name / logPath`
- 但没有可靠的 tmux attach target
- 所以 `attach``--bg` 不是简单补 handler而是需要补启动/附着元数据设计
### 代码范围
- 修改 [src/cli/bg.ts](</e:/Source_code/Claude-code-bast/src/cli/bg.ts:1>)
- 修改 [src/utils/concurrentSessions.ts](</e:/Source_code/Claude-code-bast/src/utils/concurrentSessions.ts:1>) 以便后续 attach/--bg 扩展
- 修改 [src/utils/taskSummary.ts](</e:/Source_code/Claude-code-bast/src/utils/taskSummary.ts:1>)
- 复用:
[src/utils/sessionStorage.ts](</e:/Source_code/Claude-code-bast/src/utils/sessionStorage.ts:3870>)
[src/utils/udsClient.ts](</e:/Source_code/Claude-code-bast/src/utils/udsClient.ts:1>)
### 验证
1. `ps` 能列出 live sessions
2. `logs <sessionId|pid|name>` 能输出对应日志
3. `kill <sessionId|pid|name>` 能结束目标 session
### 风险
- `attach` / `--bg` 第二阶段需要 tmux 元数据设计
- Windows 下 tmux 路径需要明确降级策略
### 工作量判断
- `ps/logs/kill` 中等
- `attach/--bg` 明显更大,应分阶段
## 3. `TEMPLATES`
### 现状
- 命令入口只有 fast-path
[src/entrypoints/cli.tsx](</e:/Source_code/Claude-code-bast/src/entrypoints/cli.tsx:249>)
- handler 是空的:
[src/cli/handlers/templateJobs.ts](</e:/Source_code/Claude-code-bast/src/cli/handlers/templateJobs.ts:1>)
- `markdownConfigLoader` 已把 `templates` 纳入配置目录:
[src/utils/markdownConfigLoader.ts](</e:/Source_code/Claude-code-bast/src/utils/markdownConfigLoader.ts:29>)
- `query / stopHooks` 已预留 job classifier 链路:
[src/query/stopHooks.ts](</e:/Source_code/Claude-code-bast/src/query/stopHooks.ts:103>)
- `jobs/classifier.ts` 仍是 stub
[src/jobs/classifier.ts](</e:/Source_code/Claude-code-bast/src/jobs/classifier.ts:1>)
### 目标
-`new / list / reply` 做成可用的模板任务系统。
- 第一阶段不碰复杂的自动分类与自动执行。
### MVP 方案
- 模板来源:
`.claude/templates/*.md`
- 模板格式:
复用现有 markdown + frontmatter 解析,不另外设计 DSL
- `list`
- 列出所有模板
- 显示模板名、description、路径
- `new <template> [args...]`
- 解析模板
-`~/.claude/jobs/<job-id>/` 下创建 job 目录
- 写入 `template.md``input.txt``state.json`
- 返回 job id 与目录
- `reply <job-id> <text>`
- 将回复写入 `replies.jsonl``input.txt`
- 更新 `state.json`
### Phase 2
- 恢复 [src/jobs/classifier.ts](</e:/Source_code/Claude-code-bast/src/jobs/classifier.ts:1>)
- 让带 `CLAUDE_JOB_DIR` 的 job session 在 turn 完成后自动更新 `state.json`
- 再决定是否补自动 job runner
### 为什么要拆
- 当前证据表明这是“template job commands”不是单纯模板列表
- 但自动 job 运行链路没有足够现成实现,先做文件系统 job lifecycle 更稳
### 代码范围
- 修改 [src/cli/handlers/templateJobs.ts](</e:/Source_code/Claude-code-bast/src/cli/handlers/templateJobs.ts:1>)
- 新增 `src/jobs/state.ts`
- 新增 `src/jobs/templates.ts`
- Phase 2 再改 [src/jobs/classifier.ts](</e:/Source_code/Claude-code-bast/src/jobs/classifier.ts:1>)
### 验证
1. `list` 能列出 `.claude/templates`
2. `new` 能创建 job 目录和状态文件
3. `reply` 能更新 job 内容和状态
4. Phase 2 再验证 classifier 写状态
### 风险
- frontmatter schema 需要先定义最小字段集
- 一旦扩展到“自动运行 job”范围会明显膨胀
### 工作量判断
- MVP 中等
- 完整 job 系统偏大
## 4. `assistant [sessionId]`
### 现状
- attach 主流程其实已经存在:
[src/main.tsx](</e:/Source_code/Claude-code-bast/src/main.tsx:4708>)
- 远端 viewer 所需基础模块已存在:
[src/remote/RemoteSessionManager.ts](</e:/Source_code/Claude-code-bast/src/remote/RemoteSessionManager.ts:1>)
[src/hooks/useAssistantHistory.ts](</e:/Source_code/Claude-code-bast/src/hooks/useAssistantHistory.ts:1>)
[src/assistant/sessionHistory.ts](</e:/Source_code/Claude-code-bast/src/assistant/sessionHistory.ts:1>)
- 真正 stub 的主要是:
[src/assistant/sessionDiscovery.ts](</e:/Source_code/Claude-code-bast/src/assistant/sessionDiscovery.ts:1>)
[src/assistant/AssistantSessionChooser.ts](</e:/Source_code/Claude-code-bast/src/assistant/AssistantSessionChooser.ts:1>)
[src/commands/assistant/assistant.ts](</e:/Source_code/Claude-code-bast/src/commands/assistant/assistant.ts:7>)
[src/assistant/index.ts](</e:/Source_code/Claude-code-bast/src/assistant/index.ts:1>)
### 目标
- 不一次性恢复整个 KAIROS 助手系统。
- 先做“明确 sessionId 的 viewer attach 可用”,再逐步补 discovery / chooser / install。
### Phase 4AMVP
- 只支持 `claude assistant <sessionId>`
-`claude assistant` 无参数模式,先返回明确提示:
- 当前版本需要显式 `sessionId`
- discovery 尚未启用
- 这样可以直接复用现有 attach 分支,不必先恢复 chooser/install wizard
### Phase 4B
- 恢复 `discoverAssistantSessions()`
- 数据来源优先复用现有 sessions / bridge / teleport API而不是新协议
-`claude assistant` 无参数时能拿到候选 session 列表
### Phase 4C
- 恢复 `AssistantSessionChooser`
- 多 session 时可交互选择
### Phase 4D
- 最后考虑 install wizard 辅助函数
- 这部分属于“没有 session 时如何引导”,不是 attach 核心路径
### 为什么要拆
- attach 渲染层与远端消息通道大部分已经在
- 真正缺的是“如何发现目标 session”和“如何交互选择”
- 如果把 `src/assistant/index.ts` 的整套 KAIROS 正常模式也一起拉进来,范围会失控
### 代码范围
- Phase 4A
- [src/main.tsx](</e:/Source_code/Claude-code-bast/src/main.tsx:4708>)
- [src/commands/assistant/index.ts](</e:/Source_code/Claude-code-bast/src/commands/assistant/index.ts:1>)
- Phase 4B
- [src/assistant/sessionDiscovery.ts](</e:/Source_code/Claude-code-bast/src/assistant/sessionDiscovery.ts:1>)
- Phase 4C
- [src/assistant/AssistantSessionChooser.ts](</e:/Source_code/Claude-code-bast/src/assistant/AssistantSessionChooser.ts:1>)
- Phase 4D
- [src/commands/assistant/assistant.ts](</e:/Source_code/Claude-code-bast/src/commands/assistant/assistant.ts:7>)
### 验证
1. `claude assistant <sessionId>` 能进入 remote viewer
2. 历史懒加载工作正常
3. 无参数模式先给出明确提示
4. 后续阶段再分别验证 discovery / chooser / install
### 风险
- 这是四项里范围最大的
- 一旦把 KAIROS 正常模式整体拉入会从“viewer attach”膨胀成“完整 assistant mode 恢复”
### 工作量判断
- Phase 4A 中等
- 4A-4D 全做完很大
## 建议执行顺序
1. `claude daemon status` / `claude daemon stop`
2. `BG_SESSIONS` 先做 `ps/logs/kill`
3. `TEMPLATES` 先做 job 文件系统 MVP
4. `assistant [sessionId]` 先做显式 sessionId attach再补 discovery/chooser/install
## 简短结论
这四项里,最适合立刻实现的是 `daemon status/stop``BG_SESSIONS``TEMPLATES` 适合按 MVP 先补 handler 与文件系统闭环。`assistant [sessionId]` 不能整块硬上应该按“attach → discovery → chooser → install”拆开恢复。

View File

@@ -0,0 +1,77 @@
# Task 001: daemon status / stop
> 来源: [stub-recovery-design-1-4.md](../features/stub-recovery-design-1-4.md) 第 1 项
> 优先级: P0 (首选实现项)
> 工作量: 小
> 状态: DONE
## 目标
`claude daemon status``claude daemon stop` 在任意 CLI 进程中都能正确工作,不依赖 TUI 内存态。
## 背景
- `start` 路径已有完整 supervisor + worker 生命周期 (`src/daemon/main.ts`, `src/daemon/workerRegistry.ts`)
- `status` / `stop` 目前只是占位输出 (`src/daemon/main.ts:49`)
- `/remote-control-server` 有自己的命令内 UI 状态,但只维护当前进程内的 `daemonProcess`,不适合跨进程管理
## 实现方案
### 新增文件
| 文件 | 说明 |
|------|------|
| `src/daemon/state.ts` | daemon 状态文件读写模块 |
### 修改文件
| 文件 | 改动 |
|------|------|
| `src/daemon/main.ts` | `start` 写入状态文件;`status`/`stop` 调用 state 模块 |
| `src/commands/remoteControlServer/remoteControlServer.tsx` | 读取同一份状态文件(轻量改动) |
### 状态文件
路径: `~/.claude/daemon/remote-control.json`
```json
{
"pid": 12345,
"cwd": "/path/to/project",
"startedAt": "2026-04-12T10:00:00Z",
"workerKinds": ["bridge", "rcs"],
"lastStatus": "running"
}
```
### status 逻辑
1. 读取状态文件
2. 用进程探测验证 pid 是否存活
3. 输出 `running` / `stopped` / `stale`
4. stale 时自动清理状态文件
### stop 逻辑
1. 读取 pid
2. 发送 `SIGTERM`
3. 等待退出(超时兜底)
4. 超时后 `SIGKILL`
5. 清理状态文件
## 验证步骤
- [ ] `claude daemon start` 正常启动并写入状态文件
- [ ] 新开终端执行 `claude daemon status`,显示 `running`
- [ ] 执行 `claude daemon stop`daemon 正常退出
- [ ] 再次执行 `claude daemon status`,返回 `stopped``stale cleaned`
- [ ] Windows 下 stop 超时兜底正常工作
## 风险
- Windows 信号模型和 Unix 不同,`stop` 需要超时兜底
- 当前设计默认单 supervisor不处理多实例并发
## 依赖
无外部依赖,可独立实施。

View File

@@ -0,0 +1,80 @@
# Task 002: BG_SESSIONS — ps / logs / kill
> 来源: [stub-recovery-design-1-4.md](../features/stub-recovery-design-1-4.md) 第 2 项
> 优先级: P1
> 工作量: 中等
> 状态: DONE
> 阶段: Phase 2A (MVP)
## 目标
`ps` / `logs` / `kill` 做成真正有用的 session 管理命令。不在第一阶段补完 `attach` / `--bg`
## 背景
- fast-path 已接好 (`src/entrypoints/cli.tsx:218`)
- session registry 已有真实实现 (`src/utils/concurrentSessions.ts`)
- `exit` 在 bg session 内已会 `tmux detach-client` (`src/commands/exit/exit.tsx:20`)
- CLI handler 仍全空 (`src/cli/bg.ts`)
- task summary 仍然是 stub (`src/utils/taskSummary.ts`)
## 实现方案
### 修改文件
| 文件 | 改动 |
|------|------|
| `src/cli/bg.ts` | 实现 `ps` / `logs` / `kill` handler |
| `src/utils/concurrentSessions.ts` | 扩展以便后续 attach/--bg 使用 |
| `src/utils/taskSummary.ts` | 补充基础实现 |
### 复用模块
- `src/utils/sessionStorage.ts` — session 存储
- `src/utils/udsClient.ts` — UDS 通信
### ps 命令
- 从 registry 读取 live sessions
- 展示: pid, kind, sessionId, cwd, name, startedAt, bridgeSessionId
- 如果有 activity/status一并展示
### logs 命令
- 支持按 `sessionId` / `pid` / `name` 查找
- 优先复用本地 transcript/log 读取能力
- 如果 registry 里存在 `logPath`,支持 tail 文件
### kill 命令
- 解析目标 session
- 发退出信号
- 清理 stale registry
## 验证步骤
- [ ] `ps` 能列出当前 live sessions
- [ ] `logs <sessionId|pid|name>` 能输出对应日志
- [ ] `kill <sessionId|pid|name>` 能结束目标 session 并清理 registry
- [ ] 无 live session 时各命令有明确提示
## Phase 2B (后续)
- [ ] 实现 `attach`
- [ ] 实现 `--bg`
- [ ] 实现 `taskSummary` 的中途状态更新
### 为什么拆分
- 现有 registry 记录了 `pid / sessionId / name / logPath`
- 但没有可靠的 tmux attach target
- `attach``--bg` 需要补启动/附着元数据设计,不是简单补 handler
## 风险
- `attach` / `--bg` 第二阶段需要 tmux 元数据设计
- Windows 下 tmux 路径需要明确降级策略
## 依赖
- Task 001 (daemon 状态管理可复用模式,但非硬性依赖)

View File

@@ -0,0 +1,87 @@
# Task 003: TEMPLATES — job 文件系统 MVP
> 来源: [stub-recovery-design-1-4.md](../features/stub-recovery-design-1-4.md) 第 3 项
> 优先级: P2
> 工作量: 中等
> 状态: DONE
> 阶段: MVP
## 目标
`new` / `list` / `reply` 做成可用的模板任务系统。第一阶段不碰复杂的自动分类与自动执行。
## 背景
- 命令入口只有 fast-path (`src/entrypoints/cli.tsx:249`)
- handler 是空的 (`src/cli/handlers/templateJobs.ts`)
- `markdownConfigLoader` 已把 `templates` 纳入配置目录 (`src/utils/markdownConfigLoader.ts:29`)
- `query/stopHooks` 已预留 job classifier 链路 (`src/query/stopHooks.ts:103`)
- `jobs/classifier.ts` 仍是 stub (`src/jobs/classifier.ts`)
## 实现方案
### 新增文件
| 文件 | 说明 |
|------|------|
| `src/jobs/state.ts` | job 状态管理 |
| `src/jobs/templates.ts` | 模板解析与列表 |
### 修改文件
| 文件 | 改动 |
|------|------|
| `src/cli/handlers/templateJobs.ts` | 实现 `new` / `list` / `reply` handler |
### 模板来源
`.claude/templates/*.md`
### 模板格式
复用现有 markdown + frontmatter 解析,不另外设计 DSL。
### list 命令
- 列出所有模板
- 显示: 模板名, description, 路径
### new 命令
- 解析模板
-`~/.claude/jobs/<job-id>/` 下创建 job 目录
- 写入 `template.md`, `input.txt`, `state.json`
- 返回 job id 与目录路径
### reply 命令
- 将回复写入 `replies.jsonl``input.txt`
- 更新 `state.json`
## 验证步骤
- [ ] `list` 能列出 `.claude/templates` 下的所有模板
- [ ] `new <template> [args...]` 能创建 job 目录和状态文件
- [ ] `reply <job-id> <text>` 能更新 job 内容和状态
- [ ] frontmatter schema 最小字段集已定义
## Phase 2 (后续)
- [ ] 恢复 `src/jobs/classifier.ts`
- [ ] 让带 `CLAUDE_JOB_DIR` 的 job session 在 turn 完成后自动更新 `state.json`
- [ ] 再决定是否补自动 job runner
### 为什么拆分
- 当前是 "template job commands",不是单纯模板列表
- 自动 job 运行链路没有足够现成实现
- 先做文件系统 job lifecycle 更稳
## 风险
- frontmatter schema 需要先定义最小字段集
- 一旦扩展到"自动运行 job",范围会明显膨胀
## 依赖
无硬性依赖,可独立实施。

View File

@@ -0,0 +1,103 @@
# Task 004: assistant [sessionId] — 分阶段恢复
> 来源: [stub-recovery-design-1-4.md](../features/stub-recovery-design-1-4.md) 第 4 项
> 优先级: P3
> 工作量: Phase 4A 中等4A-4D 全做完很大
> 状态: Phase 4A DONE, 4B-4D TODO
## 目标
不一次性恢复整个 KAIROS 助手系统。先做"明确 sessionId 的 viewer attach 可用",再逐步补 discovery / chooser / install。
## 背景
- attach 主流程已存在 (`src/main.tsx:4708`)
- 远端 viewer 所需基础模块已存在:
- `src/remote/RemoteSessionManager.ts`
- `src/hooks/useAssistantHistory.ts`
- `src/assistant/sessionHistory.ts`
- 真正 stub 的主要是:
- `src/assistant/sessionDiscovery.ts`
- `src/assistant/AssistantSessionChooser.ts`
- `src/commands/assistant/assistant.ts:7`
- `src/assistant/index.ts`
## 分阶段实现
### Phase 4A: MVP — 显式 sessionId attach
**修改文件:**
| 文件 | 改动 |
|------|------|
| `src/main.tsx` | 确保 attach 分支可用 |
| `src/commands/assistant/index.ts` | 实现显式 sessionId 参数入口 |
**行为:**
- `claude assistant <sessionId>` — 进入 remote viewer
- `claude assistant` (无参数) — 返回明确提示: 当前版本需要显式 sessionIddiscovery 尚未启用
**验证:**
- [ ] `claude assistant <sessionId>` 能进入 remote viewer
- [ ] 历史懒加载工作正常
- [ ] 无参数模式给出明确提示
### Phase 4B: session discovery
**修改文件:**
| 文件 | 改动 |
|------|------|
| `src/assistant/sessionDiscovery.ts` | 恢复 `discoverAssistantSessions()` |
**行为:**
- 数据来源优先复用现有 sessions / bridge / teleport API不新增协议
- `claude assistant` 无参数时能拿到候选 session 列表
**验证:**
- [ ] 无参数调用能列出可用 sessions
- [ ] 数据来源复用现有通道
### Phase 4C: session chooser
**修改文件:**
| 文件 | 改动 |
|------|------|
| `src/assistant/AssistantSessionChooser.ts` | 恢复交互式选择器 |
**行为:**
- 多 session 时可交互选择
**验证:**
- [ ] 多个 session 时弹出选择器
- [ ] 选择后正确 attach
### Phase 4D: install wizard
**修改文件:**
| 文件 | 改动 |
|------|------|
| `src/commands/assistant/assistant.ts` | 恢复 install wizard 辅助函数 |
**行为:**
- 没有 session 时如何引导用户
**验证:**
- [ ] 无可用 session 时引导用户创建/连接
## 为什么拆分
- attach 渲染层与远端消息通道大部分已在
- 真正缺的是"如何发现目标 session"和"如何交互选择"
- 如果把 `src/assistant/index.ts` 的整套 KAIROS 正常模式也一起拉进来,范围会失控
## 风险
- 这是四项里范围最大的
- 一旦把 KAIROS 正常模式整体拉入,会从"viewer attach"膨胀成"完整 assistant mode 恢复"
## 依赖
- Task 002 的 session registry 模式可复用

View File

@@ -0,0 +1,196 @@
# Task 013: BgEngine 跨平台后台引擎抽象
> 设计文档: [daemon-restructure-design.md](../features/daemon-restructure-design.md) § 四
> 依赖: 无
> 分支: `feat/integrate-5-branches`
## 目标
`src/cli/bg.ts` 中硬编码的 tmux 逻辑提取为引擎抽象层,实现 TmuxEngine + DetachedEngine使后台会话功能在 Windows / macOS / Linux 上都能工作。
## 背景
当前 `bg.ts``handleBgFlag()``attachHandler()` 直接调用 tmux 命令。Windows 上 `--bg` 直接报错退出。需要一个引擎抽象层,根据平台和可用工具自动选择最佳方案。
## 文件清单
### 新增
| 文件 | 说明 |
|------|------|
| `src/cli/bg/engine.ts` | BgEngine 接口 + BgStartOptions/BgStartResult 类型 |
| `src/cli/bg/engines/tmux.ts` | TmuxEngine: 从 `bg.ts` 提取 tmux 相关逻辑 |
| `src/cli/bg/engines/detached.ts` | DetachedEngine: spawn({ detached }) + logFile 重定向 |
| `src/cli/bg/engines/index.ts` | selectEngine() 自动选择 + re-export |
| `src/cli/bg/tail.ts` | 跨平台日志 tail: fs.watch + 轮询 fallback |
### 修改
| 文件 | 变更 |
|------|------|
| `src/cli/bg.ts` | `handleBgFlag()` 改为调用 `selectEngine().start()``attachHandler()` 改为调用 `engine.attach()` |
## 实现方案
### 1. BgEngine 接口 (`src/cli/bg/engine.ts`)
```typescript
export interface BgEngine {
readonly name: string
available(): Promise<boolean>
start(opts: BgStartOptions): Promise<BgStartResult>
attach(session: SessionEntry): Promise<void>
}
export interface BgStartOptions {
sessionName: string
args: string[] // CLI args (去除 --bg)
env: Record<string, string | undefined>
logPath: string
cwd: string
}
export interface BgStartResult {
pid: number
sessionName: string
logPath: string
engineUsed: 'tmux' | 'detached'
}
```
### 2. TmuxEngine (`src/cli/bg/engines/tmux.ts`)
`bg.ts:handleBgFlag()``bg.ts:attachHandler()` 提取:
- `available()`: `execFileNoThrow('tmux', ['-V'])` 返回 code === 0
- `start()`: `tmux new-session -d -s <name> <cmd>`
- `attach()`: `tmux attach-session -t <session.tmuxSessionName>`
### 3. DetachedEngine (`src/cli/bg/engines/detached.ts`)
```typescript
export class DetachedEngine implements BgEngine {
readonly name = 'detached'
async available(): Promise<boolean> {
return true // 总是可用
}
async start(opts: BgStartOptions): Promise<BgStartResult> {
const logFd = openSync(opts.logPath, 'a')
const child = spawn(process.execPath, [process.argv[1]!, ...opts.args], {
detached: true,
stdio: ['ignore', logFd, logFd],
env: opts.env,
cwd: opts.cwd,
})
child.unref()
closeSync(logFd)
return {
pid: child.pid!,
sessionName: opts.sessionName,
logPath: opts.logPath,
engineUsed: 'detached',
}
}
async attach(session: SessionEntry): Promise<void> {
// 委托给 tail.ts
await tailLog(session.logPath!)
}
}
```
### 4. 日志 Tail (`src/cli/bg/tail.ts`)
```typescript
/**
* 跨平台实时日志输出。Ctrl+C 退出,不杀后台进程。
*
* 策略:
* 1. 读取已有内容输出
* 2. fs.watch() 监听文件变化 (主方案)
* 3. 如果 fs.watch 不可靠 (某些 Windows 网络驱动器)fallback 到 500ms 轮询
*/
export async function tailLog(logPath: string): Promise<void>
```
### 5. 引擎选择 (`src/cli/bg/engines/index.ts`)
```typescript
export async function selectEngine(): Promise<BgEngine> {
if (process.platform === 'win32') {
return new DetachedEngine()
}
const tmux = new TmuxEngine()
if (await tmux.available()) {
return tmux
}
return new DetachedEngine()
}
```
### 6. bg.ts 重构
`handleBgFlag()` 改名为 `handleBgStart()`,内部逻辑:
```typescript
export async function handleBgStart(args: string[]): Promise<void> {
const engine = await selectEngine()
const sessionName = `claude-bg-${randomUUID().slice(0, 8)}`
const logPath = join(getClaudeConfigHomeDir(), 'sessions', 'logs', `${sessionName}.log`)
const result = await engine.start({
sessionName,
args: filteredArgs,
env: { ...process.env, CLAUDE_CODE_SESSION_KIND: 'bg', ... },
logPath,
cwd: process.cwd(),
})
console.log(`Background session started: ${result.sessionName}`)
console.log(` Engine: ${result.engineUsed}`)
console.log(` Log: ${result.logPath}`)
console.log(` Use \`claude daemon attach ${result.sessionName}\` to reconnect.`)
}
```
`attachHandler()` 根据 `session.engine` 字段选择引擎:
```typescript
export async function attachHandler(target: string | undefined): Promise<void> {
// ... 找到 session
if (session.engine === 'tmux' && session.tmuxSessionName) {
const tmux = new TmuxEngine()
await tmux.attach(session)
} else {
const detached = new DetachedEngine()
await detached.attach(session)
}
}
```
## SessionEntry 扩展
`sessions/<PID>.json` 新增 `engine` 字段:
```json
{
"pid": 12345,
"engine": "detached",
"logPath": "~/.claude/sessions/logs/claude-bg-a1b2c3d4.log",
"sessionId": "...",
"cwd": "..."
}
```
兼容旧格式: 如果 `engine` 字段缺失,检查 `tmuxSessionName` 存在则为 `tmux`,否则为 `detached`
## 验证清单
- [ ] Windows: `claude daemon bg` 启动后台会话,无 tmux 依赖
- [ ] Windows: `claude daemon attach <name>` 以 tail 模式附着Ctrl+C 退出不杀进程
- [ ] macOS/Linux (有 tmux): 行为与当前一致
- [ ] macOS/Linux (无 tmux): 自动 fallback 到 detached 引擎
- [ ] `claude daemon status` 正确显示 engine 类型
- [ ] 旧格式 session JSON (无 engine 字段) 兼容
- [ ] tsc --noEmit 零错误
- [ ] bun test 通过

View File

@@ -0,0 +1,275 @@
# Task 014: /daemon 命令层级化
> 设计文档: [daemon-restructure-design.md](../features/daemon-restructure-design.md) § 三.1
> 依赖: Task 013 (BgEngine 抽象)
> 分支: `feat/integrate-5-branches`
## 目标
将散落的 `daemon start/stop/status` + `ps/logs/attach/kill` + `--bg` 统一收归 `/daemon` 命名空间,实现 CLI + REPL 双注册。
## 背景
当前这些命令注册在两个互不关联的位置:
- `cli.tsx:203-212`: `daemon [start|status|stop]``daemon/main.ts`
- `cli.tsx:217-246`: `ps|logs|attach|kill|--bg``cli/bg.ts`
需要合并为统一的 `claude daemon <subcommand>` 入口,并新增 REPL `/daemon` 斜杠命令。
## 文件清单
### 新增
| 文件 | 说明 |
|------|------|
| `src/commands/daemon/index.ts` | `/daemon` REPL 斜杠命令注册 (type: local-jsx) |
| `src/commands/daemon/daemon.tsx` | `/daemon` 子命令路由 + status UI 组件 |
### 修改
| 文件 | 变更 |
|------|------|
| `src/entrypoints/cli.tsx` | 统一 daemon 快速路径: `daemon <sub>` 路由到对应 handler。旧命令 `ps/logs/attach/kill` 保留但输出 deprecation 警告后代理 |
| `src/commands.ts` | 注册 `/daemon` 斜杠命令 (feature-gated: DAEMON \|\| BG_SESSIONS) |
| `src/daemon/main.ts` | `daemonMain()` 扩展: 支持 `bg/attach/logs/kill/ps` 子命令 (委托给 bg.ts handlers) |
## 实现方案
### 1. CLI 快速路径统一 (`cli.tsx`)
**改前** (两段独立路由):
```typescript
// 段 1: daemon
if (feature('DAEMON') && args[0] === 'daemon') {
await daemonMain(args.slice(1))
}
// 段 2: bg sessions
if (feature('BG_SESSIONS') && ['ps','logs','attach','kill'].includes(args[0])) {
// ...switch/case
}
```
**改后** (统一入口):
```typescript
// 统一 daemon 入口 — 合并 daemon supervisor + bg sessions
if (
(feature('DAEMON') || feature('BG_SESSIONS')) &&
args[0] === 'daemon'
) {
profileCheckpoint('cli_daemon_path')
const { enableConfigs } = await import('../utils/config.js')
enableConfigs()
const { initSinks } = await import('../utils/sinks.js')
initSinks()
const { daemonMain } = await import('../daemon/main.js')
await daemonMain(args.slice(1))
return
}
// --bg 快捷方式 → daemon bg
if (
feature('BG_SESSIONS') &&
(args.includes('--bg') || args.includes('--background'))
) {
profileCheckpoint('cli_daemon_path')
const { enableConfigs } = await import('../utils/config.js')
enableConfigs()
const bg = await import('../cli/bg.js')
await bg.handleBgStart(args.filter(a => a !== '--bg' && a !== '--background'))
return
}
// 向后兼容: ps/logs/attach/kill → daemon <sub> (deprecated)
if (
feature('BG_SESSIONS') &&
['ps', 'logs', 'attach', 'kill'].includes(args[0] ?? '')
) {
const mapped = args[0] === 'ps' ? 'status' : args[0]
console.error(`[deprecated] Use: claude daemon ${mapped} ${args.slice(1).join(' ')}`.trim())
const { enableConfigs } = await import('../utils/config.js')
enableConfigs()
const { daemonMain } = await import('../daemon/main.js')
await daemonMain([args[0]!, ...args.slice(1)])
return
}
```
### 2. daemonMain 扩展 (`daemon/main.ts`)
```typescript
export async function daemonMain(args: string[]): Promise<void> {
const subcommand = args[0] || 'status'
switch (subcommand) {
// --- Supervisor 管理 ---
case 'start':
await runSupervisor(args.slice(1))
break
case 'stop':
await handleDaemonStop()
break
// --- 会话管理 (委托给 bg.ts) ---
case 'status':
case 'ps':
await showUnifiedStatus() // 新: daemon 状态 + 会话列表
break
case 'bg':
const bg = await import('../cli/bg.js')
await bg.handleBgStart(args.slice(1))
break
case 'attach':
const bg2 = await import('../cli/bg.js')
await bg2.attachHandler(args[1])
break
case 'logs':
const bg3 = await import('../cli/bg.js')
await bg3.logsHandler(args[1])
break
case 'kill':
const bg4 = await import('../cli/bg.js')
await bg4.killHandler(args[1])
break
case '--help': case '-h': case 'help':
printHelp()
break
default:
console.error(`Unknown daemon subcommand: ${subcommand}`)
printHelp()
process.exitCode = 1
}
}
```
### 3. 统一状态面板 (`showUnifiedStatus`)
```typescript
async function showUnifiedStatus(): Promise<void> {
// 1. Daemon supervisor 状态
const daemonResult = queryDaemonStatus()
console.log('=== Daemon Supervisor ===')
switch (daemonResult.status) {
case 'running':
console.log(` Status: running (PID: ${daemonResult.state!.pid})`)
console.log(` Workers: ${daemonResult.state!.workerKinds.join(', ')}`)
break
case 'stopped':
console.log(' Status: stopped')
break
case 'stale':
console.log(' Status: stale (cleaned up)')
break
}
// 2. 后台会话列表
console.log('\n=== Background Sessions ===')
const bg = await import('../cli/bg.js')
await bg.psHandler([])
}
```
### 4. REPL 斜杠命令注册
**`src/commands/daemon/index.ts`**:
```typescript
import type { Command } from '../../commands.js'
import { feature } from 'bun:bundle'
const daemon = {
type: 'local-jsx',
name: 'daemon',
description: 'Manage background sessions and daemon',
argumentHint: '[status|start|stop|bg|attach|logs|kill]',
isEnabled: () => {
if (feature('DAEMON')) return true
if (feature('BG_SESSIONS')) return true
return false
},
load: () => import('./daemon.js'),
} satisfies Command
export default daemon
```
**`src/commands/daemon/daemon.tsx`**:
```typescript
export async function call(
onDone: LocalJSXCommandOnDone,
context: LocalJSXCommandContext,
args: string,
): Promise<React.ReactNode> {
const parts = args.trim().split(/\s+/)
const sub = parts[0] || 'status'
switch (sub) {
case 'status':
case 'ps':
// 调用 showUnifiedStatus捕获输出
// 返回文本结果
break
case 'bg':
// REPL 中启动后台会话
break
case 'start':
case 'stop':
case 'attach':
case 'logs':
case 'kill':
// 委托给对应 handler
break
default:
onDone(`Unknown: ${sub}. Use: status|start|stop|bg|attach|logs|kill`)
return null
}
}
```
**`src/commands.ts`** 添加:
```typescript
// 条件导入
const daemonCmd =
feature('DAEMON') || feature('BG_SESSIONS')
? require('./commands/daemon/index.js').default
: null
// COMMANDS 数组中添加
...(daemonCmd ? [daemonCmd] : []),
```
### 5. 更新 help 文本 (`daemon/main.ts`)
```
Claude Code Daemon — background process management
USAGE
claude daemon [subcommand]
SUBCOMMANDS
status Show daemon and session status (default)
start Start the daemon supervisor
stop Stop the daemon
bg Start a background session
attach Attach to a background session
logs Show session logs
kill Kill a session
help Show this help
REPL
/daemon [subcommand] Same commands available in interactive mode
```
## 验证清单
- [ ] `claude daemon` (无参数) 显示统一状态面板
- [ ] `claude daemon status` 显示 supervisor + 会话列表
- [ ] `claude daemon start/stop` 与当前行为一致
- [ ] `claude daemon bg` 启动后台会话 (调用 BgEngine)
- [ ] `claude daemon attach/logs/kill <target>` 功能正常
- [ ] `claude ps` 输出 deprecation 警告 + 正常工作
- [ ] `claude logs/attach/kill` 同上
- [ ] `claude --bg` 快捷方式正常
- [ ] REPL 中 `/daemon` 可用tab 补全显示
- [ ] REPL 中 `/daemon status` 显示状态信息
- [ ] tsc --noEmit 零错误
- [ ] bun test 通过

View File

@@ -0,0 +1,177 @@
# Task 015: /job 命令层级化
> 设计文档: [daemon-restructure-design.md](../features/daemon-restructure-design.md) § 三.2
> 依赖: 无 (可与 Task 013 并行)
> 分支: `feat/integrate-5-branches`
## 目标
`claude new/list/reply` 收归 `/job` 命名空间,实现 CLI + REPL 双注册。
## 背景
当前 `new`, `list`, `reply` 是顶级 CLI 命令 (`cli.tsx:250-261`),容易与其他命令冲突(特别是 `list` 这种通用词)。需要收归 `claude job <subcommand>` 并新增 REPL `/job` 入口。
## 文件清单
### 新增
| 文件 | 说明 |
|------|------|
| `src/commands/job/index.ts` | `/job` REPL 斜杠命令注册 |
| `src/commands/job/job.tsx` | `/job` 子命令路由 |
### 修改
| 文件 | 变更 |
|------|------|
| `src/entrypoints/cli.tsx` | 新增 `job` 快速路径 + 旧 `new/list/reply` deprecation 代理 |
| `src/commands.ts` | 注册 `/job` 斜杠命令 |
### 不动
| 文件 | 说明 |
|------|------|
| `src/cli/handlers/templateJobs.ts` | 内部 handler 不变,只是被调用方式变了 |
| `src/jobs/state.ts` | job 状态管理不变 |
| `src/jobs/templates.ts` | 模板发现不变 |
| `src/jobs/classifier.ts` | 任务分类器不变 |
## 实现方案
### 1. CLI 快速路径 (`cli.tsx`)
**改后**:
```typescript
// 新: claude job <subcommand>
if (
feature('TEMPLATES') &&
args[0] === 'job'
) {
profileCheckpoint('cli_templates_path')
const { templatesMain } = await import('../cli/handlers/templateJobs.js')
await templatesMain(args.slice(1))
process.exit(0)
}
// 向后兼容 (deprecated)
if (
feature('TEMPLATES') &&
(args[0] === 'new' || args[0] === 'list' || args[0] === 'reply')
) {
console.error(`[deprecated] Use: claude job ${args[0]} ${args.slice(1).join(' ')}`.trim())
profileCheckpoint('cli_templates_path')
const { templatesMain } = await import('../cli/handlers/templateJobs.js')
await templatesMain(args)
process.exit(0)
}
```
### 2. templateJobs.ts 新增 status 子命令
在现有 `switch` 中增加:
```typescript
case 'status':
handleStatus(args.slice(1))
break
```
```typescript
function handleStatus(args: string[]): void {
const jobId = args[0]
if (!jobId) {
console.error('Usage: claude job status <job-id>')
process.exitCode = 1
return
}
const state = readJobState(jobId)
if (!state) {
console.error(`Job not found: ${jobId}`)
process.exitCode = 1
return
}
console.log(`Job: ${state.jobId}`)
console.log(` Template: ${state.templateName}`)
console.log(` Status: ${state.status}`)
console.log(` Created: ${state.createdAt}`)
console.log(` Updated: ${state.updatedAt}`)
}
```
### 3. REPL 斜杠命令
**`src/commands/job/index.ts`**:
```typescript
import type { Command } from '../../commands.js'
import { feature } from 'bun:bundle'
const job = {
type: 'local-jsx',
name: 'job',
description: 'Manage template jobs',
argumentHint: '[list|new|reply|status]',
isEnabled: () => {
if (feature('TEMPLATES')) return true
return false
},
load: () => import('./job.js'),
} satisfies Command
export default job
```
**`src/commands/job/job.tsx`**:
```typescript
export async function call(
onDone: LocalJSXCommandOnDone,
_context: LocalJSXCommandContext,
args: string,
): Promise<React.ReactNode> {
const parts = args.trim().split(/\s+/)
const sub = parts[0] || 'list'
// 委托给 templatesMain
const { templatesMain } = await import('../../cli/handlers/templateJobs.js')
// 捕获 console.log 输出作为结果返回给 REPL
const lines: string[] = []
const origLog = console.log
const origError = console.error
console.log = (...a: unknown[]) => lines.push(a.join(' '))
console.error = (...a: unknown[]) => lines.push(a.join(' '))
try {
await templatesMain([sub, ...parts.slice(1)])
} finally {
console.log = origLog
console.error = origError
}
onDone(lines.join('\n') || 'Done.', { display: 'system' })
return null
}
```
### 4. commands.ts 注册
```typescript
const jobCmd = feature('TEMPLATES')
? require('./commands/job/index.js').default
: null
// COMMANDS 数组:
...(jobCmd ? [jobCmd] : []),
```
## 验证清单
- [ ] `claude job list` 列出模板
- [ ] `claude job new <template>` 创建任务
- [ ] `claude job reply <id> <text>` 回复任务
- [ ] `claude job status <id>` 显示任务状态
- [ ] `claude job` (无参数) 等同于 `claude job list`
- [ ] `claude new/list/reply` 输出 deprecation 警告 + 正常工作
- [ ] REPL 中 `/job` 可用
- [ ] REPL 中 `/job list` 显示模板列表
- [ ] tsc --noEmit 零错误
- [ ] bun test 通过

View File

@@ -0,0 +1,123 @@
# Task 016: 向后兼容 + 测试
> 设计文档: [daemon-restructure-design.md](../features/daemon-restructure-design.md) § 五
> 依赖: Task 014, Task 015
> 分支: `feat/integrate-5-branches`
## 目标
确保旧命令向后兼容 (deprecation 警告 + 正常代理),并为重构后的命令结构编写测试。
## 文件清单
### 新增
| 文件 | 说明 |
|------|------|
| `src/daemon/__tests__/daemonMain.test.ts` | daemonMain 子命令路由测试 |
| `src/cli/bg/__tests__/engine.test.ts` | BgEngine 选择逻辑测试 |
| `src/cli/bg/__tests__/detached.test.ts` | DetachedEngine 启动/停止测试 |
| `src/cli/bg/__tests__/tail.test.ts` | 日志 tail 功能测试 |
### 修改
| 文件 | 变更 |
|------|------|
| `src/entrypoints/cli.tsx` | 确认 deprecation 路径正确代理 |
## 实现方案
### 1. 向后兼容矩阵
| 旧命令 | 新命令 | 处理方式 |
|--------|--------|---------|
| `claude ps` | `claude daemon status` | stderr 输出 `[deprecated] Use: claude daemon status`,然后执行 |
| `claude logs <x>` | `claude daemon logs <x>` | 同上 |
| `claude attach <x>` | `claude daemon attach <x>` | 同上 |
| `claude kill <x>` | `claude daemon kill <x>` | 同上 |
| `claude --bg` | `claude daemon bg` | 保留为快捷方式,**不** deprecate (太常用) |
| `claude new <t>` | `claude job new <t>` | stderr deprecation + 执行 |
| `claude list` | `claude job list` | stderr deprecation + 执行 |
| `claude reply <id>` | `claude job reply <id>` | stderr deprecation + 执行 |
**关键**: deprecation 输出到 stderr 而非 stdout不影响脚本管道。
### 2. 测试计划
#### 2.1 daemonMain 路由测试
```typescript
describe('daemonMain', () => {
test('无参数默认 status', async () => { ... })
test('start 调用 runSupervisor', async () => { ... })
test('stop 调用 handleDaemonStop', async () => { ... })
test('bg 委托给 bg.handleBgStart', async () => { ... })
test('attach 委托给 bg.attachHandler', async () => { ... })
test('logs 委托给 bg.logsHandler', async () => { ... })
test('kill 委托给 bg.killHandler', async () => { ... })
test('未知子命令设置 exitCode=1', async () => { ... })
})
```
#### 2.2 引擎选择测试
```typescript
describe('selectEngine', () => {
test('win32 返回 DetachedEngine', async () => { ... })
test('darwin + tmux 可用返回 TmuxEngine', async () => { ... })
test('darwin + tmux 不可用返回 DetachedEngine', async () => { ... })
test('linux + tmux 可用返回 TmuxEngine', async () => { ... })
})
```
#### 2.3 DetachedEngine 测试
```typescript
describe('DetachedEngine', () => {
test('available 始终返回 true', async () => { ... })
test('start 创建 detached 子进程并写入日志', async () => { ... })
test('start 返回的 PID 文件存在', async () => { ... })
})
```
#### 2.4 Tail 测试
```typescript
describe('tailLog', () => {
test('输出已有日志内容', async () => { ... })
test('追加内容时实时输出', async () => { ... })
test('SIGINT 退出 tail', async () => { ... })
})
```
### 3. 集成验证脚本
可选: 在 `scripts/` 下添加一个手动验证脚本:
```bash
#!/bin/bash
# scripts/verify-daemon-restructure.sh
echo "=== 1. claude daemon status ==="
bun run dev -- daemon status
echo "=== 2. claude daemon bg (should start) ==="
bun run dev -- daemon bg --help
echo "=== 3. claude ps (deprecated) ==="
bun run dev -- ps 2>&1 | head -1
echo "=== 4. claude job list ==="
bun run dev -- job list
echo "=== 5. claude list (deprecated) ==="
bun run dev -- list 2>&1 | head -1
```
## 验证清单
- [ ] 旧命令全部正常工作 (仅多一行 stderr 警告)
- [ ] `--bg` 保持无警告
- [ ] 所有新增测试通过
- [ ] 现有 2695 个测试无回归
- [ ] tsc --noEmit 零错误
- [ ] 手动在 Windows + macOS/Linux 上验证关键路径

View File

@@ -0,0 +1,88 @@
# OpenClaw Autonomy Baseline Test Spec
## Purpose
This test spec locks the current behavior of the existing trigger and context layers before any formal autonomy-subsystem implementation begins.
At this stage, production code is read-only. Only test files, fixtures, and planning documents may change.
## Goal
Establish a stable baseline around the parts of `Claude-code-bast` that later autonomy work is most likely to touch:
- proactive state handling
- cron task storage semantics
- cron scheduler helper semantics
- user-context cache and `CLAUDE.md` injection behavior
## Out of Scope for This Baseline Round
- New authority behavior (`AGENTS.md` / `HEARTBEAT.md`)
- New detached-run ledger behavior
- New flow behavior
- UI redesign
## Files Under Baseline Protection
- `src/proactive/index.ts`
- `src/utils/cronTasks.ts`
- `src/utils/cronScheduler.ts`
- `src/context.ts`
## Test Files Added In This Round
- `src/proactive/__tests__/state.baseline.test.ts`
- `src/commands/__tests__/proactive.baseline.test.ts`
- `src/utils/__tests__/cronTasks.baseline.test.ts`
- `src/utils/__tests__/cronScheduler.baseline.test.ts`
- `src/__tests__/context.baseline.test.ts`
## Baseline Assertions
### Proactive state
1. Activating proactive mode sets active state and activation source.
2. Pausing proactive mode suppresses `shouldTick()` and clears `nextTickAt`.
3. Blocking context suppresses `shouldTick()` and clears `nextTickAt`.
4. Subscribers are notified on state transitions.
5. The `/proactive` command enables proactive mode and emits the expected hidden reminder.
6. The `/proactive` command disables proactive mode on the second invocation.
### Cron task storage
1. Session-only cron tasks remain in memory only.
2. Durable cron tasks are persisted to `.claude/scheduled_tasks.json`.
3. Daemon-style `dir`-scoped reads exclude session-only cron tasks.
4. `removeCronTasks()` without `dir` can remove session-only tasks.
5. `removeCronTasks()` with `dir` does not mutate session-only task storage.
### Cron scheduler helpers
1. `isRecurringTaskAged()` preserves current aging semantics.
2. `buildMissedTaskNotification()` preserves the current AskUserQuestion safety wording.
3. `buildMissedTaskNotification()` preserves code-fence hardening for prompt bodies that contain backticks.
### User context caching
1. `getUserContext()` includes `currentDate`.
2. `getUserContext()` includes mocked `claudeMd` content when memory loading is enabled.
3. `CLAUDE_CODE_DISABLE_CLAUDE_MDS` suppresses `claudeMd`.
4. `setSystemPromptInjection()` clears the memoized user-context cache.
5. `getSystemContext()` reflects the injection after cache invalidation.
## Remaining Baseline Gaps
The following areas are intentionally deferred because they require higher-cost harnessing and should still avoid production-code changes:
1. `useScheduledTasks.ts` hook-level runtime behavior
2. `src/cli/print.ts` full headless scheduler loop behavior
3. `useProactive.ts` hook timer behavior
4. end-to-end queue interaction between proactive ticks and `SleepTool`
## Acceptance
This baseline round is complete when:
1. The four new test files pass.
2. No production source files are modified.
3. The tests are stable enough to serve as a pre-implementation guardrail.

View File

@@ -1,7 +1,9 @@
import { feature } from 'bun:bundle'
import { z } from 'zod/v4'
import type { ToolResultBlockParam } from 'src/Tool.js'
import { buildTool } from 'src/Tool.js'
import { lazySchema } from 'src/utils/lazySchema.js'
import { logForDebugging } from 'src/utils/debug.js'
const PUSH_NOTIFICATION_TOOL_NAME = 'PushNotification'
@@ -74,14 +76,58 @@ Requires Remote Control to be configured. Respects user notification settings (t
}
},
async call(_input: PushInput) {
// Push delivery is handled by the Remote Control / KAIROS transport layer.
// Without the KAIROS runtime, this tool is not available.
return {
data: {
sent: false,
error: 'PushNotification requires the KAIROS transport layer.',
},
async call(input: PushInput, context) {
const appState = context.getAppState()
// Try bridge delivery first (for remote/mobile viewers)
if (appState.replBridgeEnabled) {
if (feature('BRIDGE_MODE')) {
try {
const { getBridgeAccessToken, getBridgeBaseUrl } = await import(
'src/bridge/bridgeConfig.js'
)
const { getSessionId } = await import('src/bootstrap/state.js')
const token = getBridgeAccessToken()
const sessionId = getSessionId()
if (token && sessionId) {
const baseUrl = getBridgeBaseUrl()
const axios = (await import('axios')).default
const response = await axios.post(
`${baseUrl}/v1/sessions/${sessionId}/events`,
{
events: [
{
type: 'push_notification',
title: input.title,
body: input.body,
priority: input.priority ?? 'normal',
},
],
},
{
headers: {
Authorization: `Bearer ${token}`,
'Content-Type': 'application/json',
'anthropic-version': '2023-06-01',
},
timeout: 10_000,
validateStatus: (s: number) => s < 500,
},
)
if (response.status >= 200 && response.status < 300) {
logForDebugging(`[PushNotification] delivered via bridge session=${sessionId}`)
return { data: { sent: true } }
}
logForDebugging(`[PushNotification] bridge delivery failed: status=${response.status}`)
}
} catch (e) {
logForDebugging(`[PushNotification] bridge delivery error: ${e}`)
}
}
}
// Fallback: no bridge available, push was not delivered to a remote device.
logForDebugging(`[PushNotification] no bridge available, not delivered: ${input.title}`)
return { data: { sent: false, error: 'No Remote Control bridge configured. Notification not delivered.' } }
},
})

View File

@@ -70,14 +70,51 @@ Guidelines:
}
},
async call(_input: SendUserFileInput) {
// File transfer is handled by the KAIROS assistant transport layer.
// Without the KAIROS runtime, this tool is not available.
async call(input: SendUserFileInput, context) {
const { file_path } = input
const { stat } = await import('fs/promises')
// Verify file exists and is readable
let fileSize: number
try {
const fileStat = await stat(file_path)
if (!fileStat.isFile()) {
return {
data: { sent: false, file_path, error: 'Path is not a file.' },
}
}
fileSize = fileStat.size
} catch {
return {
data: { sent: false, file_path, error: 'File does not exist or is not readable.' },
}
}
// Attempt bridge upload if available (so web viewers can download)
const appState = context.getAppState()
let fileUuid: string | undefined
if (appState.replBridgeEnabled) {
try {
const { uploadBriefAttachment } = await import(
'@claude-code-best/builtin-tools/tools/BriefTool/upload.js'
)
fileUuid = await uploadBriefAttachment(file_path, fileSize, {
replBridgeEnabled: true,
signal: context.abortController.signal,
})
} catch {
// Best-effort upload — local path is always available
}
}
const delivered = !appState.replBridgeEnabled || Boolean(fileUuid)
return {
data: {
sent: false,
file_path: _input.file_path,
error: 'SendUserFile requires the KAIROS assistant transport layer.',
sent: delivered,
file_path,
size: fileSize,
...(fileUuid ? { file_uuid: fileUuid } : {}),
...(!delivered ? { error: 'Bridge upload failed. File available at local path.' } : {}),
},
}
},

View File

@@ -0,0 +1,10 @@
/** Thin logging wrapper — silent in test environment, uses console in production. */
const isTest = process.env.NODE_ENV === "test" || (typeof Bun !== "undefined" && !!Bun.env.BUN_TEST);
export function log(...args: unknown[]): void {
if (!isTest) console.log(...args);
}
export function error(...args: unknown[]): void {
if (!isTest) console.error(...args);
}

View File

@@ -1,3 +1,4 @@
import { log, error as logError } from "../../logger";
import { Hono } from "hono";
import { createBunWebSocket } from "hono/bun";
import { validateApiKey } from "../../auth/api-key";
@@ -30,14 +31,14 @@ function authenticateRequest(c: any, label: string, expectedSessionId?: string):
const payload = verifyWorkerJwt(token);
if (payload) {
if (expectedSessionId && payload.session_id !== expectedSessionId) {
console.log(`[Auth] ${label}: FAILED — JWT session_id mismatch`);
log(`[Auth] ${label}: FAILED — JWT session_id mismatch`);
return false;
}
return true;
}
}
console.log(`[Auth] ${label}: FAILED — no valid API key or JWT`);
log(`[Auth] ${label}: FAILED — no valid API key or JWT`);
return false;
}
@@ -83,7 +84,7 @@ app.get(
const session = getSession(sessionId);
if (!session) {
console.log(`[WS] Upgrade rejected: session ${sessionId} not found`);
log(`[WS] Upgrade rejected: session ${sessionId} not found`);
return {
onOpen(_evt, ws) {
ws.close(4001, "session not found");
@@ -91,7 +92,7 @@ app.get(
};
}
console.log(`[WS] Upgrade accepted: session=${sessionId}`);
log(`[WS] Upgrade accepted: session=${sessionId}`);
return {
onOpen(_evt, ws) {
handleWebSocketOpen(ws as any, sessionId);
@@ -108,7 +109,7 @@ app.get(
handleWebSocketClose(ws as any, sessionId, closeEvt?.code, closeEvt?.reason);
},
onError(evt, ws) {
console.error(`[WS] Error on session=${sessionId}:`, evt);
logError(`[WS] Error on session=${sessionId}:`, evt);
handleWebSocketClose(ws as any, sessionId, 1006, "websocket error");
},
};

View File

@@ -1,3 +1,4 @@
import { log, error as logError } from "../../logger";
import { Hono } from "hono";
import {
createSession,
@@ -22,7 +23,7 @@ app.post("/", acceptCliHeaders, apiKeyAuth, async (c) => {
try {
await createWorkItem(body.environment_id, session.id);
} catch (err) {
console.error(`[RCS] Failed to create work item: ${(err as Error).message}`);
logError(`[RCS] Failed to create work item: ${(err as Error).message}`);
}
}

View File

@@ -1,3 +1,4 @@
import { log, error as logError } from "../../logger";
import { Hono } from "hono";
import { uuidAuth } from "../../auth/middleware";
import { getSession, updateSessionStatus } from "../../services/session";
@@ -29,9 +30,9 @@ app.post("/sessions/:id/events", uuidAuth, async (c) => {
const body = await c.req.json();
const eventType = body.type || "user";
console.log(`[RC-DEBUG] web -> server: POST /web/sessions/${sessionId}/events type=${eventType} content=${JSON.stringify(body).slice(0, 200)}`);
log(`[RC-DEBUG] web -> server: POST /web/sessions/${sessionId}/events type=${eventType} content=${JSON.stringify(body).slice(0, 200)}`);
const event = publishSessionEvent(sessionId, eventType, body, "outbound");
console.log(`[RC-DEBUG] web -> server: published outbound event id=${event.id} type=${event.type} direction=${event.direction} subscribers=${getEventBus(sessionId).subscriberCount()}`);
log(`[RC-DEBUG] web -> server: published outbound event id=${event.id} type=${event.type} direction=${event.direction} subscribers=${getEventBus(sessionId).subscriberCount()}`);
return c.json({ status: "ok", event }, 200);
});

View File

@@ -1,3 +1,4 @@
import { log, error as logError } from "../../logger";
import { Hono } from "hono";
import { uuidAuth } from "../../auth/middleware";
import { getSession, createSession } from "../../services/session";
@@ -28,7 +29,7 @@ app.post("/sessions", uuidAuth, async (c) => {
try {
await createWorkItem(body.environment_id, session.id);
} catch (err) {
console.error(`[RCS] Failed to create work item: ${(err as Error).message}`);
logError(`[RCS] Failed to create work item: ${(err as Error).message}`);
}
}

View File

@@ -1,3 +1,4 @@
import { log, error as logError } from "../logger";
import { storeListActiveEnvironments, storeUpdateEnvironment } from "../store";
import { storeListSessions, storeUpdateSession } from "../store";
import { config } from "../config";
@@ -12,7 +13,7 @@ export function startDisconnectMonitor() {
const envs = storeListActiveEnvironments();
for (const env of envs) {
if (env.lastPollAt && now - env.lastPollAt.getTime() > timeoutMs) {
console.log(`[RCS] Environment ${env.id} timed out (no poll for ${Math.round((now - env.lastPollAt.getTime()) / 1000)}s)`);
log(`[RCS] Environment ${env.id} timed out (no poll for ${Math.round((now - env.lastPollAt.getTime()) / 1000)}s)`);
storeUpdateEnvironment(env.id, { status: "disconnected" });
}
}
@@ -23,7 +24,7 @@ export function startDisconnectMonitor() {
if (session.status === "running" || session.status === "idle") {
const elapsed = now - session.updatedAt.getTime();
if (elapsed > timeoutMs * 2) {
console.log(`[RCS] Session ${session.id} marked inactive (no update for ${Math.round(elapsed / 1000)}s)`);
log(`[RCS] Session ${session.id} marked inactive (no update for ${Math.round(elapsed / 1000)}s)`);
storeUpdateSession(session.id, { status: "inactive" });
}
}

View File

@@ -1,3 +1,4 @@
import { log, error as logError } from "../logger";
import {
storeCreateWorkItem,
storeGetWorkItem,
@@ -35,7 +36,7 @@ export async function createWorkItem(environmentId: string, sessionId: string):
const secret = encodeWorkSecret();
const record = storeCreateWorkItem({ environmentId, sessionId, secret });
console.log(`[RCS] Work item created: ${record.id} for env=${environmentId} session=${sessionId}`);
log(`[RCS] Work item created: ${record.id} for env=${environmentId} session=${sessionId}`);
return record.id;
}

View File

@@ -1,3 +1,5 @@
import { log, error as logError } from "../logger";
export interface SessionEvent {
id: string;
sessionId: string;
@@ -33,12 +35,12 @@ export class EventBus {
createdAt: Date.now(),
};
this.events.push(full);
console.log(`[RC-DEBUG] bus publish: sessionId=${event.sessionId} type=${event.type} dir=${event.direction} seq=${full.seqNum} subscribers=${this.subscribers.size}`);
log(`[RC-DEBUG] bus publish: sessionId=${event.sessionId} type=${event.type} dir=${event.direction} seq=${full.seqNum} subscribers=${this.subscribers.size}`);
for (const cb of this.subscribers) {
try {
cb(full);
} catch (err) {
console.error(`[RC-DEBUG] bus subscriber error:`, err);
logError(`[RC-DEBUG] bus subscriber error:`, err);
}
}
return full;

View File

@@ -1,3 +1,4 @@
import { log, error as logError } from "../logger";
import type { Context } from "hono";
import type { SessionEvent } from "./event-bus";
import { getEventBus } from "./event-bus";
@@ -76,7 +77,7 @@ export function createSSEStream(c: Context, sessionId: string, fromSeqNum = 0) {
seqNum: event.seqNum,
});
try {
console.log(`[RC-DEBUG] SSE -> web: sessionId=${sessionId} type=${event.type} dir=${event.direction} seq=${event.seqNum}`);
log(`[RC-DEBUG] SSE -> web: sessionId=${sessionId} type=${event.type} dir=${event.direction} seq=${event.seqNum}`);
controller.enqueue(encoder.encode(`id: ${event.seqNum}\nevent: message\ndata: ${data}\n\n`));
} catch {
unsub();

View File

@@ -2,6 +2,7 @@ import type { WSContext } from "hono/ws";
import { getEventBus } from "./event-bus";
import type { SessionEvent } from "./event-bus";
import { publishSessionEvent } from "../services/transport";
import { log, error as logError } from "../logger";
// Per-connection cleanup, keyed by sessionId (only one WS per session)
interface CleanupEntry {
@@ -96,13 +97,13 @@ function toSDKMessage(event: SessionEvent): string {
/** Called from onOpen — subscribes to event bus, forwards outbound events to bridge WS */
export function handleWebSocketOpen(ws: WSContext, sessionId: string) {
const openTime = Date.now();
console.log(`[RC-DEBUG] [WS] Open session=${sessionId}`);
log(`[RC-DEBUG] [WS] Open session=${sessionId}`);
activeConnections.add(ws);
// If there's an existing connection for this session, clean it up first
const existing = cleanupBySession.get(sessionId);
if (existing) {
console.log(`[WS] Replacing existing connection for session=${sessionId}`);
log(`[WS] Replacing existing connection for session=${sessionId}`);
existing.unsub();
clearInterval(existing.keepalive);
activeConnections.delete(existing.ws);
@@ -114,7 +115,7 @@ export function handleWebSocketOpen(ws: WSContext, sessionId: string) {
// the full conversation history — assistant replies are inbound events.
const missed = bus.getEventsSince(0);
if (missed.length > 0) {
console.log(`[WS] Replaying ${missed.length} missed event(s)`);
log(`[WS] Replaying ${missed.length} missed event(s)`);
for (const event of missed) {
if (ws.readyState !== 1) break;
try {
@@ -130,10 +131,10 @@ export function handleWebSocketOpen(ws: WSContext, sessionId: string) {
if (event.direction !== "outbound") return;
try {
const sdkMsg = toSDKMessage(event);
console.log(`[RC-DEBUG] [WS] -> bridge (outbound): type=${event.type} len=${sdkMsg.length} msg=${sdkMsg.slice(0, 300)}`);
log(`[RC-DEBUG] [WS] -> bridge (outbound): type=${event.type} len=${sdkMsg.length} msg=${sdkMsg.slice(0, 300)}`);
ws.send(sdkMsg);
} catch (err) {
console.error("[RC-DEBUG] [WS] send error:", err);
logError("[RC-DEBUG] [WS] send error:", err);
}
});
@@ -161,7 +162,7 @@ export function handleWebSocketMessage(ws: WSContext, sessionId: string, data: s
try {
ingestBridgeMessage(sessionId, JSON.parse(line));
} catch (err) {
console.error("[WS] parse error:", err);
logError("[WS] parse error:", err);
}
}
}
@@ -173,7 +174,7 @@ export function handleWebSocketClose(ws: WSContext, sessionId: string, code?: nu
const entry = cleanupBySession.get(sessionId);
const duration = entry ? Math.round((Date.now() - entry.openTime) / 1000) : -1;
console.log(`[WS] Close session=${sessionId} code=${code ?? "none"} reason=${reason || "(none)"} duration=${duration}s`);
log(`[WS] Close session=${sessionId} code=${code ?? "none"} reason=${reason || "(none)"} duration=${duration}s`);
if (entry) {
entry.unsub();
@@ -215,7 +216,7 @@ export function ingestBridgeMessage(sessionId: string, msg: Record<string, unkno
const eventType = deriveEventType(msg);
console.log(`[RC-DEBUG] [WS] <- bridge (inbound): sessionId=${sessionId} type=${eventType}${msg.uuid ? ` uuid=${msg.uuid}` : ""} msg=${JSON.stringify(msg).slice(0, 300)}`);
log(`[RC-DEBUG] [WS] <- bridge (inbound): sessionId=${sessionId} type=${eventType}${msg.uuid ? ` uuid=${msg.uuid}` : ""} msg=${JSON.stringify(msg).slice(0, 300)}`);
let payload: unknown;
@@ -255,7 +256,7 @@ export function closeAllConnections(): void {
const count = activeConnections.size;
if (count === 0) return;
console.log(`[WS] Gracefully closing ${count} active connection(s)...`);
log(`[WS] Gracefully closing ${count} active connection(s)...`);
for (const [sessionId, entry] of cleanupBySession) {
try {
entry.unsub();
@@ -269,5 +270,5 @@ export function closeAllConnections(): void {
}
cleanupBySession.clear();
activeConnections.clear();
console.log("[WS] All connections closed");
log("[WS] All connections closed");
}

View File

@@ -47,6 +47,8 @@ const DEFAULT_FEATURES = [
"KAIROS",
"COORDINATOR_MODE",
"LAN_PIPES",
"BG_SESSIONS",
"TEMPLATES",
// "REVIEW_ARTIFACT", // API 请求无响应,需进一步排查 schema 兼容性
// P3: poor mode (disable extract_memories + prompt_suggestion)
"POOR",

View File

@@ -0,0 +1,91 @@
import { afterEach, beforeEach, describe, expect, test } from 'bun:test'
import {
resetStateForTests,
setOriginalCwd,
setProjectRoot,
} from '../bootstrap/state'
import {
getSystemContext,
getUserContext,
setSystemPromptInjection,
} from '../context'
import { clearMemoryFileCaches } from '../utils/claudemd'
import { cleanupTempDir, createTempDir, writeTempFile } from '../../tests/mocks/file-system'
let tempDir = ''
let projectClaudeMdContent = ''
beforeEach(async () => {
tempDir = await createTempDir('context-baseline-')
projectClaudeMdContent = `baseline-${Date.now()}`
resetStateForTests()
setOriginalCwd(tempDir)
setProjectRoot(tempDir)
await writeTempFile(tempDir, 'CLAUDE.md', projectClaudeMdContent)
clearMemoryFileCaches()
getUserContext.cache.clear?.()
getSystemContext.cache.clear?.()
setSystemPromptInjection(null)
delete process.env.CLAUDE_CODE_DISABLE_CLAUDE_MDS
})
afterEach(async () => {
clearMemoryFileCaches()
getUserContext.cache.clear?.()
getSystemContext.cache.clear?.()
setSystemPromptInjection(null)
delete process.env.CLAUDE_CODE_DISABLE_CLAUDE_MDS
resetStateForTests()
if (tempDir) {
await cleanupTempDir(tempDir)
}
})
describe('context baseline', () => {
test('getUserContext includes currentDate and project CLAUDE.md content', async () => {
const ctx = await getUserContext()
expect(ctx.currentDate).toContain("Today's date is")
expect(ctx.claudeMd).toContain(projectClaudeMdContent)
})
test('CLAUDE_CODE_DISABLE_CLAUDE_MDS suppresses claudeMd loading', async () => {
process.env.CLAUDE_CODE_DISABLE_CLAUDE_MDS = '1'
const ctx = await getUserContext()
expect(ctx.currentDate).toContain("Today's date is")
expect(ctx.claudeMd).toBeUndefined()
})
test('setSystemPromptInjection clears the memoized user-context cache', async () => {
const first = await getUserContext()
process.env.CLAUDE_CODE_DISABLE_CLAUDE_MDS = '1'
const second = await getUserContext()
expect(first.claudeMd).toContain(projectClaudeMdContent)
expect(second.claudeMd).toContain(projectClaudeMdContent)
setSystemPromptInjection('cache-break')
const third = await getUserContext()
expect(third.claudeMd).toBeUndefined()
})
test('getSystemContext reflects system prompt injection after cache invalidation', async () => {
const first = await getSystemContext()
expect(first.gitStatus).toBeUndefined()
expect(first.cacheBreaker).toBeUndefined()
setSystemPromptInjection('baseline-cache-break')
const second = await getSystemContext()
if ('cacheBreaker' in second) {
expect(second.cacheBreaker).toContain('baseline-cache-break')
} else {
expect(second.gitStatus).toBeUndefined()
}
})
})

View File

@@ -1,3 +0,0 @@
// Auto-generated stub — replace with real implementation
export {};
export const AssistantSessionChooser: (props: Record<string, unknown>) => null = () => null;

View File

@@ -0,0 +1,54 @@
import * as React from 'react';
import { useState } from 'react';
import { Box, Text } from '@anthropic/ink';
import { Dialog } from '../components/design-system/Dialog.js';
import { ListItem } from '../components/design-system/ListItem.js';
import { useRegisterOverlay } from '../context/overlayContext.js';
import { useKeybindings } from '../keybindings/useKeybinding.js';
import type { AssistantSession } from './sessionDiscovery.js';
interface Props {
sessions: AssistantSession[];
onSelect: (id: string) => void;
onCancel: () => void;
}
/**
* Interactive session chooser for `claude assistant` when multiple
* CCR sessions are discovered. Renders a Dialog with up/down navigation.
*
* Session IDs are in `session_*` compat format — passed directly to
* createRemoteSessionConfig() for viewer attach.
*/
export function AssistantSessionChooser({ sessions, onSelect, onCancel }: Props): React.ReactNode {
useRegisterOverlay('assistant-session-chooser');
const [focusIndex, setFocusIndex] = useState(0);
useKeybindings(
{
'select:next': () => setFocusIndex(i => (i + 1) % sessions.length),
'select:previous': () => setFocusIndex(i => (i - 1 + sessions.length) % sessions.length),
'select:accept': () => onSelect(sessions[focusIndex]!.id),
},
{ context: 'Select' },
);
return (
<Dialog title="Select Assistant Session" onCancel={onCancel} hideInputGuide>
<Box flexDirection="column" gap={1}>
<Text>Multiple sessions found. Select one to attach:</Text>
<Box flexDirection="column">
{sessions.map((s, i) => (
<ListItem key={s.id} isFocused={focusIndex === i}>
<Box>
<Text>{s.title || s.id.slice(0, 20)}</Text>
<Text dimColor> [{s.status}]</Text>
</Box>
</ListItem>
))}
</Box>
<Text dimColor> navigate · Enter select · Esc cancel</Text>
</Box>
</Dialog>
);
}

View File

@@ -5,21 +5,20 @@ import { getFeatureValue_CACHED_MAY_BE_STALE } from '../services/analytics/growt
/**
* Runtime gate for KAIROS features.
*
* Build-time: feature('KAIROS') must be on (checked by caller before
* this module is required).
* Two-layer gate:
* 1. Build-time: feature('KAIROS') must be on
* 2. Runtime: tengu_kairos_assistant GrowthBook flag (remote kill switch)
*
* Runtime: tengu_kairos_assistant GrowthBook flag acts as a remote kill
* switch, and kairosActive state must be true (set during bootstrap when
* the session qualifies for KAIROS features).
* Called by main.tsx BEFORE setKairosActive(true) — must NOT check
* kairosActive (that would deadlock: gate needs active, active needs gate).
* The caller (main.tsx L1826-1832) sets kairosActive after this returns true.
*/
export async function isKairosEnabled(): Promise<boolean> {
if (!feature('KAIROS')) {
return false
}
if (
!getFeatureValue_CACHED_MAY_BE_STALE('tengu_kairos_assistant', false)
) {
if (!getFeatureValue_CACHED_MAY_BE_STALE('tengu_kairos_assistant', false)) {
return false
}
return getKairosActive()
return true
}

View File

@@ -1,9 +1,64 @@
// Auto-generated stub — replace with real implementation
export {}
export const isAssistantMode: () => boolean = () => false
export const initializeAssistantTeam: () => Promise<void> = async () => {}
export const markAssistantForced: () => void = () => {}
export const isAssistantForced: () => boolean = () => false
export const getAssistantSystemPromptAddendum: () => string = () => ''
export const getAssistantActivationPath: () => string | undefined = () =>
undefined
import { readFileSync } from 'fs'
import { join } from 'path'
import { getKairosActive } from '../bootstrap/state.js'
import { getClaudeConfigHomeDir } from '../utils/envUtils.js'
let _assistantForced = false
/**
* Whether the current session is in assistant (KAIROS) daemon mode.
* Wraps the bootstrap kairosActive state set by main.tsx after gate check.
*/
export function isAssistantMode(): boolean {
return getKairosActive()
}
/**
* Mark this session as forced assistant mode (--assistant flag).
* Skips the GrowthBook gate check — daemon is pre-entitled.
*/
export function markAssistantForced(): void {
_assistantForced = true
}
export function isAssistantForced(): boolean {
return _assistantForced
}
/**
* Pre-create an in-process team so Agent(name) can spawn teammates
* without TeamCreate.
*
* Phase 1: returns undefined so main.tsx's `assistantTeamContext ?? computeInitialTeamContext()`
* correctly falls back. Returning {} would bypass the ?? operator since {} is truthy.
*
* Phase 2: should return a full team context object matching AppState.teamContext shape.
*/
export async function initializeAssistantTeam(): Promise<undefined> {
return undefined
}
/**
* Assistant-specific system prompt addendum loaded from ~/.claude/agents/assistant.md.
* Returns empty string if the file doesn't exist.
*/
export function getAssistantSystemPromptAddendum(): string {
try {
return readFileSync(
join(getClaudeConfigHomeDir(), 'agents', 'assistant.md'),
'utf-8',
)
} catch {
return ''
}
}
/**
* How assistant mode was activated. Used for diagnostics/analytics.
* - 'daemon': via --assistant flag (Agent SDK daemon)
* - 'gate': via GrowthBook gate check
*/
export function getAssistantActivationPath(): string | undefined {
if (!isAssistantMode()) return undefined
return _assistantForced ? 'daemon' : 'gate'
}

View File

@@ -1,3 +1,51 @@
// Auto-generated stub — replace with real implementation
export type AssistantSession = { id: string; [key: string]: unknown };
export const discoverAssistantSessions: () => Promise<AssistantSession[]> = () => Promise.resolve([]);
import { logForDebugging } from '../utils/debug.js'
/**
* Minimal session type for assistant discovery.
* Only `id` is consumed by main.tsx (L4757); other fields are for chooser display.
* ID format is `session_*` (compat prefix) — viewer endpoints use /v1/sessions/*.
*/
export type AssistantSession = {
id: string
title: string
status: string
created_at: string
}
/**
* Discover assistant sessions on Anthropic CCR.
*
* Reuses the existing fetchCodeSessionsFromSessionsAPI() which calls
* GET /v1/sessions with proper OAuth + anthropic-beta headers.
*
* Throws on failure — main.tsx L4720-4725 catch displays the error.
* Does NOT return [] on error (that would silently redirect to install wizard).
*/
export async function discoverAssistantSessions(): Promise<AssistantSession[]> {
const { fetchCodeSessionsFromSessionsAPI } = await import(
'../utils/teleport/api.js'
)
let allSessions
try {
allSessions = await fetchCodeSessionsFromSessionsAPI()
} catch (err) {
logForDebugging(
`[assistant:discovery] fetchCodeSessionsFromSessionsAPI failed: ${err}`,
)
throw err
}
// Filter to active/working sessions only — completed/archived are not attachable
return allSessions
.filter(
s =>
s.status === 'idle' || s.status === 'working' || s.status === 'waiting',
)
.map(s => ({
id: s.id,
title: s.title || 'Untitled',
status: s.status,
created_at: s.created_at ?? '',
}))
}

View File

@@ -1,7 +1,314 @@
// Auto-generated stub — replace with real implementation
export {};
export const psHandler: (args: string[]) => Promise<void> = (async () => {}) as (args: string[]) => Promise<void>;
export const logsHandler: (sessionId: string | undefined) => Promise<void> = (async () => {}) as (sessionId: string | undefined) => Promise<void>;
export const attachHandler: (sessionId: string | undefined) => Promise<void> = (async () => {}) as (sessionId: string | undefined) => Promise<void>;
export const killHandler: (sessionId: string | undefined) => Promise<void> = (async () => {}) as (sessionId: string | undefined) => Promise<void>;
export const handleBgFlag: (args: string[]) => Promise<void> = (async () => {}) as (args: string[]) => Promise<void>;
import { readdir, readFile, unlink } from 'fs/promises'
import { join } from 'path'
import { randomUUID } from 'crypto'
import { getClaudeConfigHomeDir } from '../utils/envUtils.js'
import { isProcessRunning } from '../utils/genericProcessUtils.js'
import { jsonParse } from '../utils/slowOperations.js'
import { selectEngine } from './bg/engines/index.js'
import type { SessionEntry } from './bg/engine.js'
export type { SessionEntry } from './bg/engine.js'
function getSessionsDir(): string {
return join(getClaudeConfigHomeDir(), 'sessions')
}
export async function listLiveSessions(): Promise<SessionEntry[]> {
const dir = getSessionsDir()
let files: string[]
try {
files = await readdir(dir)
} catch {
return []
}
const sessions: SessionEntry[] = []
for (const file of files) {
if (!/^\d+\.json$/.test(file)) continue
const pid = parseInt(file.slice(0, -5), 10)
if (!isProcessRunning(pid)) {
void unlink(join(dir, file)).catch(() => {})
continue
}
try {
const raw = await readFile(join(dir, file), 'utf-8')
const entry = jsonParse(raw) as SessionEntry
sessions.push(entry)
} catch {
// Corrupt file — skip
}
}
return sessions
}
export function findSession(
sessions: SessionEntry[],
target: string,
): SessionEntry | undefined {
const asNum = parseInt(target, 10)
return sessions.find(
s =>
s.sessionId === target ||
s.pid === asNum ||
(s.name && s.name === target),
)
}
function formatTime(ts: number): string {
return new Date(ts).toLocaleString()
}
/**
* Resolve the engine type for an existing session.
* Backward-compatible: sessions without an `engine` field are inferred
* from the presence of `tmuxSessionName`.
*/
function resolveSessionEngine(session: SessionEntry): 'tmux' | 'detached' {
if (session.engine) return session.engine
return session.tmuxSessionName ? 'tmux' : 'detached'
}
/**
* `claude daemon status` / `claude ps` — list live sessions.
*/
export async function psHandler(_args: string[]): Promise<void> {
const sessions = await listLiveSessions()
if (sessions.length === 0) {
console.log('No active sessions.')
return
}
console.log(
`${sessions.length} active session${sessions.length > 1 ? 's' : ''}:\n`,
)
for (const s of sessions) {
const engineType = resolveSessionEngine(s)
const parts: string[] = [
` PID: ${s.pid}`,
` Kind: ${s.kind}`,
` Engine: ${engineType}`,
` Session: ${s.sessionId}`,
` CWD: ${s.cwd}`,
]
if (s.name) parts.push(` Name: ${s.name}`)
if (s.startedAt) parts.push(` Started: ${formatTime(s.startedAt)}`)
if (s.status) parts.push(` Status: ${s.status}`)
if (s.waitingFor) parts.push(` Waiting for: ${s.waitingFor}`)
if (s.bridgeSessionId) parts.push(` Bridge: ${s.bridgeSessionId}`)
if (s.tmuxSessionName) parts.push(` Tmux: ${s.tmuxSessionName}`)
if (s.logPath) parts.push(` Log: ${s.logPath}`)
console.log(parts.join('\n'))
console.log()
}
}
/**
* `claude daemon logs <target>` — show logs for a session.
*/
export async function logsHandler(target: string | undefined): Promise<void> {
const sessions = await listLiveSessions()
if (!target) {
if (sessions.length === 0) {
console.log('No active sessions.')
return
}
if (sessions.length === 1) {
target = sessions[0]!.sessionId
} else {
console.log('Multiple sessions active. Specify one:')
for (const s of sessions) {
const label = s.name ? `${s.name} (${s.sessionId})` : s.sessionId
console.log(` ${label} PID=${s.pid}`)
}
return
}
}
const session = findSession(sessions, target)
if (!session) {
console.error(`Session not found: ${target}`)
process.exitCode = 1
return
}
if (!session.logPath) {
console.log(`No log path recorded for session ${session.sessionId}`)
return
}
try {
const content = await readFile(session.logPath, 'utf-8')
process.stdout.write(content)
} catch (e) {
console.error(`Failed to read log file: ${session.logPath}`)
console.error(e instanceof Error ? e.message : String(e))
process.exitCode = 1
}
}
/**
* `claude daemon attach <target>` — attach to a background session.
*
* Engine-aware: tmux sessions use tmux attach, detached sessions use log tail.
*/
export async function attachHandler(target: string | undefined): Promise<void> {
const sessions = await listLiveSessions()
if (!target) {
// Find bg sessions (tmux or detached)
const bgSessions = sessions.filter(
s => s.tmuxSessionName || s.engine === 'detached',
)
if (bgSessions.length === 0) {
console.log(
'No background sessions to attach to. Start one with `claude daemon bg`.',
)
return
}
if (bgSessions.length === 1) {
target = bgSessions[0]!.sessionId
} else {
console.log('Multiple background sessions. Specify one:')
for (const s of bgSessions) {
const label = s.name ? `${s.name} (${s.sessionId})` : s.sessionId
const engineType = resolveSessionEngine(s)
console.log(` ${label} PID=${s.pid} engine=${engineType}`)
}
return
}
}
const session = findSession(sessions, target)
if (!session) {
console.error(`Session not found: ${target}`)
process.exitCode = 1
return
}
const engineType = resolveSessionEngine(session)
try {
if (engineType === 'tmux') {
const { TmuxEngine } = await import('./bg/engines/tmux.js')
const tmux = new TmuxEngine()
if (!(await tmux.available())) {
console.error('tmux is no longer available. Cannot attach to tmux session.')
process.exitCode = 1
return
}
await tmux.attach(session)
} else {
const { DetachedEngine } = await import('./bg/engines/detached.js')
const detached = new DetachedEngine()
await detached.attach(session)
}
} catch (e) {
console.error(e instanceof Error ? e.message : String(e))
process.exitCode = 1
}
}
/**
* `claude daemon kill <target>` — kill a session.
*/
export async function killHandler(target: string | undefined): Promise<void> {
const sessions = await listLiveSessions()
if (!target) {
if (sessions.length === 0) {
console.log('No active sessions to kill.')
return
}
console.log('Specify a session to kill:')
for (const s of sessions) {
const label = s.name ? `${s.name} (${s.sessionId})` : s.sessionId
console.log(` ${label} PID=${s.pid}`)
}
return
}
const session = findSession(sessions, target)
if (!session) {
console.error(`Session not found: ${target}`)
process.exitCode = 1
return
}
console.log(`Killing session ${session.sessionId} (PID: ${session.pid})...`)
try {
process.kill(session.pid, 'SIGTERM')
} catch {
console.log('Session already exited.')
return
}
await new Promise(resolve => setTimeout(resolve, 2000))
if (isProcessRunning(session.pid)) {
try {
process.kill(session.pid, 'SIGKILL')
console.log('Session force-killed.')
} catch {
console.log('Session exited during grace period.')
}
} else {
console.log('Session stopped.')
}
const pidFile = join(getSessionsDir(), `${session.pid}.json`)
void unlink(pidFile).catch(() => {})
}
/**
* `claude daemon bg [args]` — start a background session.
*
* Cross-platform: uses TmuxEngine on macOS/Linux when tmux is available,
* falls back to DetachedEngine on Windows or when tmux is absent.
*/
export async function handleBgStart(args: string[]): Promise<void> {
const engine = await selectEngine()
const sessionName = `claude-bg-${randomUUID().slice(0, 8)}`
const logPath = join(
getClaudeConfigHomeDir(),
'sessions',
'logs',
`${sessionName}.log`,
)
// Strip --bg/--background from args (for backward-compat shortcut)
const filteredArgs = args.filter(a => a !== '--bg' && a !== '--background')
try {
const result = await engine.start({
sessionName,
args: filteredArgs,
env: { ...process.env },
logPath,
cwd: process.cwd(),
})
console.log(`Background session started: ${result.sessionName}`)
console.log(` Engine: ${result.engineUsed}`)
console.log(` Log: ${result.logPath}`)
console.log()
console.log(`Use \`claude daemon attach ${result.sessionName}\` to reconnect.`)
console.log(`Use \`claude daemon status\` to check status.`)
console.log(`Use \`claude daemon kill ${result.sessionName}\` to stop.`)
} catch (e) {
console.error(e instanceof Error ? e.message : String(e))
process.exitCode = 1
}
}
// Legacy export alias — kept for backward compatibility with cli.tsx
export const handleBgFlag = handleBgStart

View File

@@ -0,0 +1,15 @@
import { describe, test, expect } from 'bun:test'
import { DetachedEngine } from '../engines/detached.js'
describe('DetachedEngine', () => {
test('name is "detached"', () => {
const engine = new DetachedEngine()
expect(engine.name).toBe('detached')
})
test('available always returns true', async () => {
const engine = new DetachedEngine()
const result = await engine.available()
expect(result).toBe(true)
})
})

View File

@@ -0,0 +1,37 @@
import { describe, test, expect } from 'bun:test'
describe('selectEngine', () => {
test('returns engine with valid BgEngine interface', async () => {
const { selectEngine } = await import('../engines/index.js')
const engine = await selectEngine()
expect(engine.name).toBeDefined()
expect(['tmux', 'detached']).toContain(engine.name)
expect(typeof engine.available).toBe('function')
expect(typeof engine.start).toBe('function')
expect(typeof engine.attach).toBe('function')
})
test('engine.available() returns a boolean', async () => {
const { selectEngine } = await import('../engines/index.js')
const engine = await selectEngine()
const result = await engine.available()
expect(typeof result).toBe('boolean')
})
})
describe('SessionEntry type', () => {
test('engine field accepts tmux or detached', async () => {
// Verify the module loads and exports the expected interface shape
const mod = await import('../engine.js')
expect(mod).toBeDefined()
const entry = {
pid: 123,
sessionId: 'test',
cwd: '/tmp',
startedAt: Date.now(),
kind: 'bg',
engine: 'detached' as const,
}
expect(entry.engine).toBe('detached')
})
})

View File

@@ -0,0 +1,8 @@
import { describe, test, expect } from 'bun:test'
describe('tailLog', () => {
test('module exports tailLog function', async () => {
const mod = await import('../tail.js')
expect(typeof mod.tailLog).toBe('function')
})
})

47
src/cli/bg/engine.ts Normal file
View File

@@ -0,0 +1,47 @@
/**
* BgEngine — cross-platform background session engine abstraction.
*
* Implementations:
* TmuxEngine — macOS/Linux with tmux installed
* DetachedEngine — Windows, or macOS/Linux without tmux (fallback)
*/
export interface SessionEntry {
pid: number
sessionId: string
cwd: string
startedAt: number
kind: string
name?: string
logPath?: string
entrypoint?: string
status?: string
waitingFor?: string
updatedAt?: number
bridgeSessionId?: string
agent?: string
tmuxSessionName?: string
engine?: 'tmux' | 'detached'
}
export interface BgStartOptions {
sessionName: string
args: string[]
env: Record<string, string | undefined>
logPath: string
cwd: string
}
export interface BgStartResult {
pid: number
sessionName: string
logPath: string
engineUsed: 'tmux' | 'detached'
}
export interface BgEngine {
readonly name: 'tmux' | 'detached'
available(): Promise<boolean>
start(opts: BgStartOptions): Promise<BgStartResult>
attach(session: SessionEntry): Promise<void>
}

View File

@@ -0,0 +1,51 @@
import { spawn } from 'child_process'
import { openSync, closeSync, mkdirSync } from 'fs'
import { dirname } from 'path'
import type { BgEngine, BgStartOptions, BgStartResult, SessionEntry } from '../engine.js'
import { tailLog } from '../tail.js'
export class DetachedEngine implements BgEngine {
readonly name = 'detached' as const
async available(): Promise<boolean> {
return true
}
async start(opts: BgStartOptions): Promise<BgStartResult> {
mkdirSync(dirname(opts.logPath), { recursive: true })
const logFd = openSync(opts.logPath, 'a')
const entrypoint = process.argv[1]!
const child = spawn(process.execPath, [entrypoint, ...opts.args], {
detached: true,
stdio: ['ignore', logFd, logFd],
env: {
...opts.env,
CLAUDE_CODE_SESSION_KIND: 'bg',
CLAUDE_CODE_SESSION_NAME: opts.sessionName,
CLAUDE_CODE_SESSION_LOG: opts.logPath,
} as Record<string, string>,
cwd: opts.cwd,
})
child.unref()
closeSync(logFd)
const pid = child.pid ?? 0
return {
pid,
sessionName: opts.sessionName,
logPath: opts.logPath,
engineUsed: 'detached',
}
}
async attach(session: SessionEntry): Promise<void> {
if (!session.logPath) {
throw new Error(`Session ${session.sessionId} has no log path.`)
}
await tailLog(session.logPath)
}
}

View File

@@ -0,0 +1,17 @@
export type { BgEngine, BgStartOptions, BgStartResult, SessionEntry } from '../engine.js'
export async function selectEngine(): Promise<import('../engine.js').BgEngine> {
if (process.platform === 'win32') {
const { DetachedEngine } = await import('./detached.js')
return new DetachedEngine()
}
const { TmuxEngine } = await import('./tmux.js')
const tmux = new TmuxEngine()
if (await tmux.available()) {
return tmux
}
const { DetachedEngine } = await import('./detached.js')
return new DetachedEngine()
}

View File

@@ -0,0 +1,73 @@
import { spawnSync } from 'child_process'
import { execFileNoThrow } from '../../../utils/execFileNoThrow.js'
import { quote } from '../../../utils/bash/shellQuote.js'
import type { BgEngine, BgStartOptions, BgStartResult, SessionEntry } from '../engine.js'
export class TmuxEngine implements BgEngine {
readonly name = 'tmux' as const
async available(): Promise<boolean> {
const { code } = await execFileNoThrow('tmux', ['-V'], { useCwd: false })
return code === 0
}
async start(opts: BgStartOptions): Promise<BgStartResult> {
const entrypoint = process.argv[1]!
const cmd = quote([process.execPath, entrypoint, ...opts.args])
const tmuxEnv: Record<string, string | undefined> = {
...opts.env,
CLAUDE_CODE_SESSION_KIND: 'bg',
CLAUDE_CODE_SESSION_NAME: opts.sessionName,
CLAUDE_CODE_SESSION_LOG: opts.logPath,
CLAUDE_CODE_TMUX_SESSION: opts.sessionName,
}
const result = spawnSync(
'tmux',
['new-session', '-d', '-s', opts.sessionName, cmd],
{ stdio: 'inherit', env: tmuxEnv },
)
if (result.status !== 0) {
throw new Error('Failed to create tmux session.')
}
// tmux doesn't directly report the child PID; we return 0.
// The actual session process writes its own PID file.
return {
pid: 0,
sessionName: opts.sessionName,
logPath: opts.logPath,
engineUsed: 'tmux',
}
}
async attach(session: SessionEntry): Promise<void> {
if (!session.tmuxSessionName) {
throw new Error(`Session ${session.sessionId} has no tmux session name.`)
}
const result = spawnSync(
'tmux',
['attach-session', '-t', session.tmuxSessionName],
{ stdio: 'inherit' },
)
if (result.status !== 0) {
throw new Error(
`Failed to attach to tmux session '${session.tmuxSessionName}'.`,
)
}
}
}
export function getTmuxInstallHint(): string {
if (process.platform === 'darwin') {
return 'Install with: brew install tmux'
}
if (process.platform === 'win32') {
return 'tmux is not natively available on Windows. Consider using WSL.'
}
return 'Install with: sudo apt install tmux (or your package manager)'
}

70
src/cli/bg/tail.ts Normal file
View File

@@ -0,0 +1,70 @@
import {
openSync,
readSync,
closeSync,
statSync,
watchFile,
unwatchFile,
createReadStream,
} from 'fs'
import { createInterface } from 'readline'
/**
* Cross-platform real-time log output. Ctrl+C exits tail without killing
* the background process.
*
* Strategy:
* 1. Read existing content and output to stdout
* 2. Use fs.watchFile() (polling-based — works everywhere including Windows)
* 3. On change, read new bytes from the last known position
* 4. SIGINT exits cleanly
*/
export async function tailLog(logPath: string): Promise<void> {
let position = 0
// Output existing content
try {
const stat = statSync(logPath)
position = stat.size
if (position > 0) {
const stream = createReadStream(logPath, { start: 0, end: position - 1 })
const rl = createInterface({ input: stream })
for await (const line of rl) {
process.stdout.write(line + '\n')
}
}
} catch {
// File may not exist yet — that's fine
}
console.log('\n[tail] Watching for new output... (Ctrl+C to detach)\n')
return new Promise<void>(resolve => {
const onSignal = (): void => {
unwatchFile(logPath)
process.removeListener('SIGINT', onSignal)
console.log('\n[tail] Detached from session.')
resolve()
}
process.on('SIGINT', onSignal)
watchFile(logPath, { interval: 300 }, () => {
try {
const stat = statSync(logPath)
if (stat.size <= position) return
const fd = openSync(logPath, 'r')
try {
const buf = Buffer.alloc(stat.size - position)
readSync(fd, buf, 0, buf.length, position)
process.stdout.write(buf)
position = stat.size
} finally {
closeSync(fd)
}
} catch {
// File may have been deleted or truncated
}
})
})
}

View File

@@ -1,13 +1,216 @@
// Auto-generated stub — replace with real implementation
import type { Command } from '@commander-js/extra-typings';
import type { Command } from '@commander-js/extra-typings'
import {
createTask,
getTask,
updateTask,
listTasks,
getTasksDir,
} from '../../utils/tasks.js'
import { getRecentActivity } from '../../utils/logoV2Utils.js'
import type { LogOption } from '../../types/logs.js'
export {};
export const logHandler: (logId: string | number | undefined) => Promise<void> = (async () => {}) as (logId: string | number | undefined) => Promise<void>;
export const errorHandler: (num: number | undefined) => Promise<void> = (async () => {}) as (num: number | undefined) => Promise<void>;
export const exportHandler: (source: string, outputFile: string) => Promise<void> = (async () => {}) as (source: string, outputFile: string) => Promise<void>;
export const taskCreateHandler: (subject: string, opts: { description?: string; list?: string }) => Promise<void> = (async () => {}) as (subject: string, opts: { description?: string; list?: string }) => Promise<void>;
export const taskListHandler: (opts: { list?: string; pending?: boolean; json?: boolean }) => Promise<void> = (async () => {}) as (opts: { list?: string; pending?: boolean; json?: boolean }) => Promise<void>;
export const taskGetHandler: (id: string, opts: { list?: string }) => Promise<void> = (async () => {}) as (id: string, opts: { list?: string }) => Promise<void>;
export const taskUpdateHandler: (id: string, opts: { list?: string; status?: string; subject?: string; description?: string; owner?: string; clearOwner?: boolean }) => Promise<void> = (async () => {}) as (id: string, opts: { list?: string; status?: string; subject?: string; description?: string; owner?: string; clearOwner?: boolean }) => Promise<void>;
export const taskDirHandler: (opts: { list?: string }) => Promise<void> = (async () => {}) as (opts: { list?: string }) => Promise<void>;
export const completionHandler: (shell: string, opts: { output?: string }, program: Command) => Promise<void> = (async () => {}) as (shell: string, opts: { output?: string }, program: Command) => Promise<void>;
const DEFAULT_LIST = 'default'
// ─── Group C: Task CRUD ──────────────────────────────────────────────────────
export async function taskCreateHandler(
subject: string,
opts: { description?: string; list?: string },
): Promise<void> {
const listId = opts.list || DEFAULT_LIST
const id = await createTask(listId, {
subject,
description: opts.description || '',
status: 'pending',
blocks: [],
blockedBy: [],
})
console.log(`Created task ${id}: ${subject}`)
}
export async function taskListHandler(opts: {
list?: string
pending?: boolean
json?: boolean
}): Promise<void> {
const listId = opts.list || DEFAULT_LIST
let tasks = await listTasks(listId)
if (opts.pending) {
tasks = tasks.filter(t => t.status === 'pending')
}
if (opts.json) {
console.log(JSON.stringify(tasks, null, 2))
return
}
if (tasks.length === 0) {
console.log('No tasks found.')
return
}
for (const t of tasks) {
console.log(` [${t.status}] ${t.id}: ${t.subject}`)
if (t.description) console.log(` ${t.description}`)
if (t.owner) console.log(` owner: ${t.owner}`)
}
}
export async function taskGetHandler(
id: string,
opts: { list?: string },
): Promise<void> {
const listId = opts.list || DEFAULT_LIST
const task = await getTask(listId, id)
if (!task) {
console.error(`Task not found: ${id}`)
process.exitCode = 1
return
}
console.log(JSON.stringify(task, null, 2))
}
export async function taskUpdateHandler(
id: string,
opts: {
list?: string
status?: string
subject?: string
description?: string
owner?: string
clearOwner?: boolean
},
): Promise<void> {
const listId = opts.list || DEFAULT_LIST
const updates: Record<string, unknown> = {}
if (opts.status) updates.status = opts.status
if (opts.subject) updates.subject = opts.subject
if (opts.description) updates.description = opts.description
if (opts.owner) updates.owner = opts.owner
if (opts.clearOwner) updates.owner = undefined
const task = await updateTask(listId, id, updates)
if (!task) {
console.error(`Task not found: ${id}`)
process.exitCode = 1
return
}
console.log(`Updated task ${id}: [${task.status}] ${task.subject}`)
}
export async function taskDirHandler(opts: { list?: string }): Promise<void> {
const listId = opts.list || DEFAULT_LIST
console.log(getTasksDir(listId))
}
// ─── Group B: Log / Error / Export ───────────────────────────────────────────
export async function logHandler(
logId: string | number | undefined,
): Promise<void> {
const logs = await getRecentActivity()
if (logId === undefined) {
if (logs.length === 0) {
console.log('No recent sessions.')
return
}
for (let i = 0; i < Math.min(logs.length, 20); i++) {
const log = logs[i]!
const date = log.modified
? new Date(log.modified).toLocaleString()
: 'unknown'
const title =
(log as Record<string, unknown>).title || log.sessionId || 'untitled'
console.log(` ${i}: ${title} (${date})`)
}
return
}
const idx = typeof logId === 'string' ? parseInt(logId, 10) : logId
const log =
Number.isFinite(idx) && idx >= 0 && idx < logs.length
? logs[idx]
: logs.find(l => l.sessionId === String(logId))
if (!log) {
console.error(`Session not found: ${logId}`)
process.exitCode = 1
return
}
console.log(JSON.stringify(log, null, 2))
}
export async function errorHandler(num: number | undefined): Promise<void> {
// Error log viewing — shows recent session errors
const logs = await getRecentActivity()
const count = num ?? 5
console.log(`Last ${count} sessions:`)
for (let i = 0; i < Math.min(count, logs.length); i++) {
const log = logs[i]!
const date = log.modified
? new Date(log.modified).toLocaleString()
: 'unknown'
console.log(` ${i}: ${log.sessionId} (${date})`)
}
}
export async function exportHandler(
source: string,
outputFile: string,
): Promise<void> {
const { writeFile, readFile } = await import('fs/promises')
const logs = await getRecentActivity()
// Try as index first
const idx = parseInt(source, 10)
let log: LogOption | undefined
if (Number.isFinite(idx) && idx >= 0 && idx < logs.length) {
log = logs[idx]
} else {
log = logs.find(l => l.sessionId === source)
}
if (!log) {
// Try as file path
try {
const content = await readFile(source, 'utf-8')
await writeFile(outputFile, content, 'utf-8')
console.log(`Exported ${source}${outputFile}`)
return
} catch {
console.error(`Source not found: ${source}`)
process.exitCode = 1
return
}
}
await writeFile(outputFile, JSON.stringify(log, null, 2), 'utf-8')
console.log(`Exported session ${log.sessionId}${outputFile}`)
}
// ─── Group D: Completion ─────────────────────────────────────────────────────
export async function completionHandler(
shell: string,
opts: { output?: string },
_program: Command,
): Promise<void> {
const { regenerateCompletionCache } = await import(
'../../utils/completionCache.js'
)
if (opts.output) {
// Generate and write to file
await regenerateCompletionCache()
console.log(`Completion cache regenerated for ${shell}.`)
} else {
// Regenerate and output to stdout
await regenerateCompletionCache()
console.log(`Completion cache regenerated for ${shell}.`)
}
}

View File

@@ -1,3 +1,158 @@
// Auto-generated stub — replace with real implementation
export {};
export const templatesMain: (args: string[]) => Promise<void> = () => Promise.resolve();
import { randomUUID } from 'crypto'
import { listTemplates, loadTemplate } from '../../jobs/templates.js'
import {
createJob,
readJobState,
appendJobReply,
getJobDir,
} from '../../jobs/state.js'
/**
* Entry point for template job commands: `new`, `list`, `reply`.
* Called from cli.tsx fast-path.
*/
export async function templatesMain(args: string[]): Promise<void> {
const subcommand = args[0]
switch (subcommand) {
case 'list':
handleList()
break
case 'new':
handleNew(args.slice(1))
break
case 'reply':
handleReply(args.slice(1))
break
case 'status':
handleStatus(args.slice(1))
break
default:
console.error(`Unknown template command: ${subcommand}`)
printUsage()
process.exitCode = 1
}
}
function printUsage(): void {
console.log(`
Template Job Commands:
claude job list List available templates
claude job new <template> [args] Create a new job from a template
claude job reply <job-id> <text> Reply to an existing job
claude job status <job-id> Show job status
`)
}
function handleStatus(args: string[]): void {
const jobId = args[0]
if (!jobId) {
console.error('Usage: claude job status <job-id>')
process.exitCode = 1
return
}
const state = readJobState(jobId)
if (!state) {
console.error(`Job not found: ${jobId}`)
process.exitCode = 1
return
}
console.log(`Job: ${state.jobId}`)
console.log(` Template: ${state.templateName}`)
console.log(` Status: ${state.status}`)
console.log(` Created: ${state.createdAt}`)
console.log(` Updated: ${state.updatedAt}`)
console.log(` Args: ${state.args.join(' ') || '(none)'}`)
}
function handleList(): void {
const templates = listTemplates()
if (templates.length === 0) {
console.log('No templates found.')
console.log('Place .md files in .claude/templates/ or ~/.claude/templates/')
return
}
console.log(
`${templates.length} template${templates.length > 1 ? 's' : ''} found:\n`,
)
for (const t of templates) {
console.log(` ${t.name}`)
console.log(` ${t.description}`)
console.log(` Path: ${t.filePath}`)
console.log()
}
}
function handleNew(args: string[]): void {
const templateName = args[0]
if (!templateName) {
console.error('Usage: claude job new <template> [args...]')
process.exitCode = 1
return
}
const template = loadTemplate(templateName)
if (!template) {
console.error(`Template not found: ${templateName}`)
console.log('\nAvailable templates:')
for (const t of listTemplates()) {
console.log(` ${t.name}`)
}
process.exitCode = 1
return
}
const jobId = randomUUID().slice(0, 8)
const inputText = args.slice(1).join(' ')
const rawContent = `---\n${Object.entries(template.frontmatter)
.map(([k, v]) => `${k}: ${v}`)
.join('\n')}\n---\n${template.content}`
const dir = createJob(
jobId,
templateName,
rawContent,
inputText,
args.slice(1),
)
console.log(`Job created: ${jobId}`)
console.log(` Template: ${templateName}`)
console.log(` Directory: ${dir}`)
if (inputText) {
console.log(` Input: ${inputText}`)
}
}
function handleReply(args: string[]): void {
const jobId = args[0]
const text = args.slice(1).join(' ')
if (!jobId || !text) {
console.error('Usage: claude job reply <job-id> <text>')
process.exitCode = 1
return
}
const state = readJobState(jobId)
if (!state) {
console.error(`Job not found: ${jobId}`)
process.exitCode = 1
return
}
const ok = appendJobReply(jobId, text)
if (ok) {
console.log(`Reply added to job ${jobId}`)
console.log(` Directory: ${getJobDir(jobId)}`)
} else {
console.error(`Failed to append reply to job ${jobId}`)
process.exitCode = 1
}
}

View File

@@ -320,6 +320,17 @@ import {
logQueryProfileReport,
} from 'src/utils/queryProfiler.js'
import { asSessionId } from 'src/types/ids.js'
import {
commitAutonomyQueuedPrompt,
createAutonomyQueuedPrompt,
createProactiveAutonomyCommands,
finalizeAutonomyRunCompleted,
finalizeAutonomyRunFailed,
markAutonomyRunCompleted,
markAutonomyRunFailed,
markAutonomyRunRunning,
} from 'src/utils/autonomyRuns.js'
import { prepareAutonomyTurnPrompt } from 'src/utils/autonomyAuthority.js'
import { jsonStringify } from '../utils/slowOperations.js'
import { skillChangeDetector } from '../utils/skills/skillChangeDetector.js'
import { getCommands, clearCommandsCache } from '../commands.js'
@@ -362,9 +373,12 @@ const proactiveModule =
feature('PROACTIVE') || feature('KAIROS')
? (require('../proactive/index.js') as typeof import('../proactive/index.js'))
: null
const cronSchedulerModule = require('../utils/cronScheduler.js') as typeof import('../utils/cronScheduler.js')
const cronJitterConfigModule = require('../utils/cronJitterConfig.js') as typeof import('../utils/cronJitterConfig.js')
const cronGate = require('@claude-code-best/builtin-tools/tools/ScheduleCronTool/prompt.js') as typeof import('@claude-code-best/builtin-tools/tools/ScheduleCronTool/prompt.js')
const cronSchedulerModule =
require('../utils/cronScheduler.js') as typeof import('../utils/cronScheduler.js')
const cronJitterConfigModule =
require('../utils/cronJitterConfig.js') as typeof import('../utils/cronJitterConfig.js')
const cronGate =
require('@claude-code-best/builtin-tools/tools/ScheduleCronTool/prompt.js') as typeof import('@claude-code-best/builtin-tools/tools/ScheduleCronTool/prompt.js')
const extractMemoriesModule = feature('EXTRACT_MEMORIES')
? (require('../services/extractMemories/extractMemories.js') as typeof import('../services/extractMemories/extractMemories.js'))
: null
@@ -1180,7 +1194,9 @@ function runHeadlessStreaming(
removeInterruptedMessage(mutableMessages, turnInterruptionState.message)
enqueue({
mode: 'prompt',
value: turnInterruptionState.message.message!.content as string | ContentBlockParam[],
value: turnInterruptionState.message.message!.content as
| string
| ContentBlockParam[],
uuid: randomUUID(),
})
}
@@ -1642,7 +1658,10 @@ function runHeadlessStreaming(
connection.config.type === 'stdio' ||
connection.config.type === undefined
) {
const stdioConfig = connection.config as { command: string; args: string[] }
const stdioConfig = connection.config as {
command: string
args: string[]
}
config = {
type: 'stdio' as const,
command: stdioConfig.command,
@@ -1804,7 +1823,8 @@ function runHeadlessStreaming(
}
for (const [name, config] of Object.entries(sdkMcpConfigs)) {
if (config.type === 'sdk' && !(name in supportedConfigs)) {
supportedConfigs[name] = config as unknown as McpServerConfigForProcessTransport
supportedConfigs[name] =
config as unknown as McpServerConfigForProcessTransport
}
}
const { response, sdkServersChanged } =
@@ -1839,15 +1859,23 @@ function runHeadlessStreaming(
) {
return
}
const tickContent = `<${TICK_TAG}>${new Date().toLocaleTimeString()}</${TICK_TAG}>`
enqueue({
mode: 'prompt' as const,
value: tickContent,
uuid: randomUUID(),
priority: 'later',
isMeta: true,
})
void run()
void (async () => {
const commands = await createProactiveAutonomyCommands({
basePrompt: `<${TICK_TAG}>${new Date().toLocaleTimeString()}</${TICK_TAG}>`,
currentDir: cwd(),
shouldCreate: () => !inputClosed,
})
for (const command of commands) {
if (inputClosed) {
return
}
enqueue({
...command,
uuid: randomUUID(),
})
}
void run()
})()
}, 0)
}
: undefined
@@ -2092,6 +2120,9 @@ function runHeadlessStreaming(
}
const input = command.value
const autonomyRunIds = batch
.map(item => item.autonomy?.runId)
.filter((runId): runId is string => Boolean(runId))
if (structuredIO instanceof RemoteIO && command.mode === 'prompt') {
logEvent('tengu_bridge_message_received', {
@@ -2141,107 +2172,151 @@ function runHeadlessStreaming(
// const-capture: TS loses `while ((command = dequeue()))` narrowing
// inside the closure.
const cmd = command
await runWithWorkload(cmd.workload ?? options.workload, async () => {
for await (const message of ask({
commands: uniqBy(
[...currentCommands, ...appState.mcp.commands],
'name',
),
prompt: input,
promptUuid: cmd.uuid,
isMeta: cmd.isMeta,
cwd: cwd(),
tools: allTools,
verbose: options.verbose,
mcpClients: allMcpClients,
thinkingConfig: options.thinkingConfig,
maxTurns: options.maxTurns,
maxBudgetUsd: options.maxBudgetUsd,
taskBudget: options.taskBudget,
canUseTool,
userSpecifiedModel: activeUserSpecifiedModel,
fallbackModel: options.fallbackModel,
jsonSchema: getInitJsonSchema() ?? options.jsonSchema,
mutableMessages,
getReadFileCache: () =>
pendingSeeds.size === 0
? readFileState
: mergeFileStateCaches(readFileState, pendingSeeds),
setReadFileCache: cache => {
readFileState = cache
for (const [path, seed] of pendingSeeds.entries()) {
const existing = readFileState.get(path)
if (!existing || seed.timestamp > existing.timestamp) {
readFileState.set(path, seed)
for (const runId of autonomyRunIds) {
await markAutonomyRunRunning(runId)
}
let lastResultIsError = false
try {
await runWithWorkload(
cmd.workload ?? options.workload,
async () => {
for await (const message of ask({
commands: uniqBy(
[...currentCommands, ...appState.mcp.commands],
'name',
),
prompt: input,
promptUuid: cmd.uuid,
isMeta: cmd.isMeta,
cwd: cwd(),
tools: allTools,
verbose: options.verbose,
mcpClients: allMcpClients,
thinkingConfig: options.thinkingConfig,
maxTurns: options.maxTurns,
maxBudgetUsd: options.maxBudgetUsd,
taskBudget: options.taskBudget,
canUseTool,
userSpecifiedModel: activeUserSpecifiedModel,
fallbackModel: options.fallbackModel,
jsonSchema: getInitJsonSchema() ?? options.jsonSchema,
mutableMessages,
getReadFileCache: () =>
pendingSeeds.size === 0
? readFileState
: mergeFileStateCaches(readFileState, pendingSeeds),
setReadFileCache: cache => {
readFileState = cache
for (const [path, seed] of pendingSeeds.entries()) {
const existing = readFileState.get(path)
if (!existing || seed.timestamp > existing.timestamp) {
readFileState.set(path, seed)
}
}
pendingSeeds.clear()
},
customSystemPrompt: options.systemPrompt,
appendSystemPrompt: options.appendSystemPrompt,
getAppState,
setAppState,
abortController,
replayUserMessages: options.replayUserMessages,
includePartialMessages: options.includePartialMessages,
handleElicitation: (serverName, params, elicitSignal) =>
structuredIO.handleElicitation(
serverName,
params.message,
undefined,
elicitSignal,
params.mode,
params.url,
'elicitationId' in params
? params.elicitationId
: undefined,
),
agents: currentAgents,
orphanedPermission: cmd.orphanedPermission,
setSDKStatus: status => {
output.enqueue({
type: 'system',
subtype: 'status',
status: status as 'compacting' | null,
session_id: getSessionId(),
uuid: randomUUID(),
})
},
})) {
// Forward messages to bridge incrementally (mid-turn) so
// claude.ai sees progress and the connection stays alive
// while blocked on permission requests.
forwardMessagesToBridge()
if (message.type === 'result') {
lastResultIsError = !!(message as Record<string, unknown>)
.is_error
// Flush pending SDK events so they appear before result on the stream.
for (const event of drainSdkEvents()) {
output.enqueue(event)
}
// Hold-back: don't emit result while background agents are running
const currentState = getAppState()
if (
getRunningTasks(currentState).some(
t =>
(t.type === 'local_agent' ||
t.type === 'local_workflow') &&
isBackgroundTask(t),
)
) {
heldBackResult = message as StdoutMessage
} else {
heldBackResult = null
output.enqueue(message as StdoutMessage)
}
} else {
// Flush SDK events (task_started, task_progress) so background
// agent progress is streamed in real-time, not batched until result.
for (const event of drainSdkEvents()) {
output.enqueue(event)
}
output.enqueue(message as StdoutMessage)
}
}
pendingSeeds.clear()
},
customSystemPrompt: options.systemPrompt,
appendSystemPrompt: options.appendSystemPrompt,
getAppState,
setAppState,
abortController,
replayUserMessages: options.replayUserMessages,
includePartialMessages: options.includePartialMessages,
handleElicitation: (serverName, params, elicitSignal) =>
structuredIO.handleElicitation(
serverName,
params.message,
undefined,
elicitSignal,
params.mode,
params.url,
'elicitationId' in params ? params.elicitationId : undefined,
),
agents: currentAgents,
orphanedPermission: cmd.orphanedPermission,
setSDKStatus: status => {
output.enqueue({
type: 'system',
subtype: 'status',
status: status as 'compacting' | null,
session_id: getSessionId(),
uuid: randomUUID(),
) // end runWithWorkload
if (lastResultIsError) {
for (const runId of autonomyRunIds) {
await finalizeAutonomyRunFailed({
runId,
error: 'ask() returned an error result',
})
},
})) {
// Forward messages to bridge incrementally (mid-turn) so
// claude.ai sees progress and the connection stays alive
// while blocked on permission requests.
forwardMessagesToBridge()
if (message.type === 'result') {
// Flush pending SDK events so they appear before result on the stream.
for (const event of drainSdkEvents()) {
output.enqueue(event)
}
} else {
for (const runId of autonomyRunIds) {
const nextCommands = await finalizeAutonomyRunCompleted({
runId,
currentDir: cwd(),
priority: 'later',
workload: cmd.workload ?? options.workload,
})
for (const nextCommand of nextCommands) {
enqueue({
...nextCommand,
uuid: randomUUID(),
})
}
// Hold-back: don't emit result while background agents are running
const currentState = getAppState()
if (
getRunningTasks(currentState).some(
t =>
(t.type === 'local_agent' ||
t.type === 'local_workflow') &&
isBackgroundTask(t),
)
) {
heldBackResult = message as StdoutMessage
} else {
heldBackResult = null
output.enqueue(message as StdoutMessage)
}
} else {
// Flush SDK events (task_started, task_progress) so background
// agent progress is streamed in real-time, not batched until result.
for (const event of drainSdkEvents()) {
output.enqueue(event)
}
output.enqueue(message as StdoutMessage)
}
}
}) // end runWithWorkload
} catch (error) {
for (const runId of autonomyRunIds) {
await finalizeAutonomyRunFailed({
runId,
error: String(error),
})
}
throw error
}
for (const uuid of batchUuids) {
notifyCommandLifecycle(uuid, 'completed')
@@ -2253,10 +2328,15 @@ function runHeadlessStreaming(
if (feature('FILE_PERSISTENCE') && turnStartTime !== undefined) {
void executeFilePersistence(
{ turnStartTime } as import('src/utils/filePersistence/types.js').TurnStartTime,
{
turnStartTime,
} as import('src/utils/filePersistence/types.js').TurnStartTime,
abortController.signal,
result => {
const filesResult = result as unknown as { persistedFiles: { filename: string; file_id: string }[]; failedFiles: { filename: string; error: string }[] }
const filesResult = result as unknown as {
persistedFiles: { filename: string; file_id: string }[]
failedFiles: { filename: string; error: string }[]
}
output.enqueue({
type: 'system' as const,
subtype: 'files_persisted' as const,
@@ -2700,28 +2780,73 @@ function runHeadlessStreaming(
// the end of run() picks up the queued command.
let cronScheduler: import('../utils/cronScheduler.js').CronScheduler | null =
null
if (
cronGate.isKairosCronEnabled()
) {
if (cronGate.isKairosCronEnabled()) {
cronScheduler = cronSchedulerModule.createCronScheduler({
onFire: prompt => {
if (inputClosed) return
enqueue({
mode: 'prompt',
value: prompt,
uuid: randomUUID(),
priority: 'later',
// System-generated — matches useScheduledTasks.ts REPL equivalent.
// Without this, messages.ts metaProp eval is {} → prompt leaks
// into visible transcript when cron fires mid-turn in -p mode.
isMeta: true,
// Threaded to cc_workload= in the billing-header attribution block
// so the API can serve cron requests at lower QoS. drainCommandQueue
// reads this per-iteration and hoists it into bootstrap state for
// the ask() call.
workload: WORKLOAD_CRON,
})
void run()
void (async () => {
const prepared = await prepareAutonomyTurnPrompt({
basePrompt: prompt,
trigger: 'scheduled-task',
currentDir: cwd(),
})
if (inputClosed) return
const command = await commitAutonomyQueuedPrompt({
prepared,
currentDir: cwd(),
workload: WORKLOAD_CRON,
})
if (inputClosed) return
enqueue({
...command,
uuid: randomUUID(),
})
void run()
})()
},
onFireTask: task => {
if (inputClosed) return
void (async () => {
if (task.agentId) {
const prepared = await prepareAutonomyTurnPrompt({
basePrompt: task.prompt,
trigger: 'scheduled-task',
currentDir: cwd(),
})
if (inputClosed) return
const command = await commitAutonomyQueuedPrompt({
prepared,
currentDir: cwd(),
sourceId: task.id,
sourceLabel: task.prompt,
workload: WORKLOAD_CRON,
})
await markAutonomyRunFailed(
command.autonomy!.runId,
`No teammate runtime available for scheduled task owner ${task.agentId} in headless mode.`,
)
return
}
const prepared = await prepareAutonomyTurnPrompt({
basePrompt: task.prompt,
trigger: 'scheduled-task',
currentDir: cwd(),
})
if (inputClosed) return
const command = await commitAutonomyQueuedPrompt({
prepared,
currentDir: cwd(),
sourceId: task.id,
sourceLabel: task.prompt,
workload: WORKLOAD_CRON,
})
if (inputClosed) return
enqueue({
...command,
uuid: randomUUID(),
})
void run()
})()
},
isLoading: () => running || inputClosed,
getJitterConfig: cronJitterConfigModule?.getCronJitterConfig,
@@ -2996,7 +3121,9 @@ function runHeadlessStreaming(
sdkClient.type === 'connected' &&
sdkClient.client?.transport?.onmessage
) {
sdkClient.client.transport.onmessage(mcpRequest.message as import('@modelcontextprotocol/sdk/types.js').JSONRPCMessage)
sdkClient.client.transport.onmessage(
mcpRequest.message as import('@modelcontextprotocol/sdk/types.js').JSONRPCMessage,
)
}
sendControlResponseSuccess(msg)
} else if (msg.request.subtype === 'rewind_files') {
@@ -3061,7 +3188,10 @@ function runHeadlessStreaming(
sendControlResponseSuccess(msg)
} else if (msg.request.subtype === 'mcp_set_servers') {
const { response, sdkServersChanged } = await applyMcpServerChanges(
msg.request.servers as Record<string, McpServerConfigForProcessTransport>,
msg.request.servers as Record<
string,
McpServerConfigForProcessTransport
>,
)
sendControlResponseSuccess(msg, response)
@@ -3131,7 +3261,8 @@ function runHeadlessStreaming(
model: a.model === 'inherit' ? undefined : a.model,
})),
plugins,
mcpServers: buildMcpServerStatuses() as SDKControlReloadPluginsResponse['mcpServers'],
mcpServers:
buildMcpServerStatuses() as SDKControlReloadPluginsResponse['mcpServers'],
error_count: r.error_count,
} satisfies SDKControlReloadPluginsResponse)
} catch (error) {
@@ -3406,7 +3537,7 @@ function runHeadlessStreaming(
mcp: {
...prev.mcp,
clients: prev.mcp.clients.map(c =>
c.name === serverName as string ? result.client : c,
c.name === (serverName as string) ? result.client : c,
),
tools: [
...reject(prev.mcp.tools, t =>
@@ -3455,7 +3586,9 @@ function runHeadlessStreaming(
})
.finally(() => {
// Clean up only if this is still the active flow
if (activeOAuthFlows.get(serverName as string) === controller) {
if (
activeOAuthFlows.get(serverName as string) === controller
) {
activeOAuthFlows.delete(serverName as string)
oauthCallbackSubmitters.delete(serverName as string)
oauthManualCallbackUsed.delete(serverName as string)
@@ -3570,7 +3703,9 @@ function runHeadlessStreaming(
// next API call re-reads keychain/file and works. No respawn.
await installOAuthTokens(tokens)
logEvent('tengu_oauth_success', {
loginWithClaudeAi: (loginWithClaudeAi ?? true) as boolean | number,
loginWithClaudeAi: (loginWithClaudeAi ?? true) as
| boolean
| number,
})
})
.finally(() => {
@@ -3618,10 +3753,7 @@ function runHeadlessStreaming(
req.subtype === 'claude_oauth_wait_for_completion'
) {
if (!claudeOAuth) {
sendControlResponseError(
msg,
'No active claude_authenticate flow',
)
sendControlResponseError(msg, 'No active claude_authenticate flow')
} else {
// Inject the manual code synchronously — must happen in stdin
// message order so a subsequent claude_authenticate doesn't
@@ -3681,7 +3813,7 @@ function runHeadlessStreaming(
mcp: {
...prev.mcp,
clients: prev.mcp.clients.map(c =>
c.name === serverName as string ? result.client : c,
c.name === (serverName as string) ? result.client : c,
),
tools: [
...reject(prev.mcp.tools, t => t.name?.startsWith(prefix)),
@@ -4116,9 +4248,13 @@ function runHeadlessStreaming(
mode: 'prompt' as const,
// file_attachments rides the protobuf catchall from the web composer.
// Same-ref no-op when absent (no 'file_attachments' key).
value: await resolveAndPrepend(userMsg, (userMsg.message as { content: ContentBlockParam[] }).content),
value: await resolveAndPrepend(
userMsg,
(userMsg.message as { content: ContentBlockParam[] }).content,
),
uuid: userMsg.uuid as `${string}-${string}-${string}-${string}-${string}`,
priority: (userMsg as { priority?: string }).priority as import('src/types/textInputTypes.js').QueuePriority,
priority: (userMsg as { priority?: string })
.priority as import('src/types/textInputTypes.js').QueuePriority,
})
// Increment prompt count for attribution tracking and save snapshot
// The snapshot persists promptCount so it survives compaction
@@ -4447,7 +4583,10 @@ async function handleInitializeRequest(
const accountInfo = getAccountInformation()
if (request.hooks) {
const hooks: Partial<Record<HookEvent, HookCallbackMatcher[]>> = {}
for (const [event, matchers] of Object.entries(request.hooks) as [string, Array<{ hookCallbackIds: string[]; timeout?: number; matcher?: string }>][]) {
for (const [event, matchers] of Object.entries(request.hooks) as [
string,
Array<{ hookCallbackIds: string[]; timeout?: number; matcher?: string }>,
][]) {
hooks[event as HookEvent] = matchers.map(matcher => {
const callbacks = matcher.hookCallbackIds.map(callbackId => {
return structuredIO.createHookCallback(callbackId, matcher.timeout)
@@ -4489,7 +4628,11 @@ async function handleInitializeRequest(
// getAccountInformation() returns undefined under 3P providers, so the
// other fields are all absent. apiProvider disambiguates "not logged
// in" (firstParty + tokenSource:none) from "3P, login not applicable".
apiProvider: getAPIProvider() as 'firstParty' | 'bedrock' | 'vertex' | 'foundry',
apiProvider: getAPIProvider() as
| 'firstParty'
| 'bedrock'
| 'vertex'
| 'foundry',
},
pid: process.pid,
}
@@ -4537,7 +4680,11 @@ async function handleRewindFiles(
dryRun: boolean,
): Promise<RewindFilesResult> {
if (!fileHistoryEnabled()) {
return { canRewind: false, error: 'File rewinding is not enabled.', filesChanged: [] }
return {
canRewind: false,
error: 'File rewinding is not enabled.',
filesChanged: [],
}
}
if (!fileHistoryCanRestore(appState.fileHistory, userMessageId)) {
return {
@@ -4842,7 +4989,10 @@ function reregisterChannelHandlerAfterReconnect(
value: wrapChannelMessage(connection.name, content, meta),
priority: 'next',
isMeta: true,
origin: { kind: 'channel', server: connection.name } as unknown as string,
origin: {
kind: 'channel',
server: connection.name,
} as unknown as string,
skipSlashCommands: true,
})
},
@@ -5266,13 +5416,21 @@ export async function handleOrphanedPermissionResponse({
onEnqueued?: () => void
handledToolUseIds: Set<string>
}): Promise<boolean> {
const responseInner = message.response as { subtype?: string; response?: Record<string, unknown>; request_id?: string } | undefined
const responseInner = message.response as
| {
subtype?: string
response?: Record<string, unknown>
request_id?: string
}
| undefined
if (
responseInner?.subtype === 'success' &&
responseInner.response?.toolUseID &&
typeof responseInner.response.toolUseID === 'string'
) {
const permissionResult = responseInner.response as PermissionResult & { toolUseID?: string }
const permissionResult = responseInner.response as PermissionResult & {
toolUseID?: string
}
const toolUseID = permissionResult.toolUseID
if (!toolUseID) {
return false

View File

@@ -1,2 +1,70 @@
// Auto-generated stub
export async function rollback(target?: string, options?: { list?: boolean; dryRun?: boolean; safe?: boolean }): Promise<void> {}
/**
* `claude rollback [target]` — roll back to a previous Claude Code version.
*
* ANT-only command (USER_TYPE === "ant").
*
* Options:
* --list List recent published versions
* --dry-run Show what would be installed without installing
* --safe Roll back to the server-pinned safe version
*/
export async function rollback(
target?: string,
options?: { list?: boolean; dryRun?: boolean; safe?: boolean },
): Promise<void> {
if (options?.list) {
console.log('Recent versions:')
console.log(' (version listing requires access to the release registry)')
console.log(' Use `claude update --list` for available versions.')
return
}
if (options?.safe) {
console.log('Safe rollback: would install the server-pinned safe version.')
if (options.dryRun) {
console.log(' (dry run — no changes made)')
return
}
console.log(' Safe version pinning requires access to the release API.')
console.log(' Contact oncall for the current safe version.')
return
}
if (!target) {
console.error(
'Usage: claude rollback [target]\n\n' +
'Options:\n' +
' -l, --list List recent published versions\n' +
' --dry-run Show what would be installed\n' +
' --safe Roll back to server-pinned safe version\n\n' +
'Examples:\n' +
' claude rollback 2.1.880\n' +
' claude rollback --list\n' +
' claude rollback --safe',
)
process.exitCode = 1
return
}
console.log(`Rolling back to version ${target}...`)
if (options?.dryRun) {
console.log(` (dry run — would install ${target})`)
return
}
// Version rollback via npm/bun
const { spawnSync } = await import('child_process')
const result = spawnSync(
'npm',
['install', '-g', `@anthropic-ai/claude-code@${target}`],
{ stdio: 'inherit' },
)
if (result.status !== 0) {
console.error(`Rollback failed with exit code ${result.status}`)
process.exitCode = result.status ?? 1
} else {
console.log(`Rolled back to ${target} successfully.`)
}
}

View File

@@ -1,2 +1,95 @@
// Auto-generated stub
export async function up(): Promise<void> {}
import { readFileSync } from 'fs'
import { join } from 'path'
import { spawnSync } from 'child_process'
import { findGitRoot } from '../utils/git.js'
/**
* `claude up` — run the "# claude up" section from the nearest CLAUDE.md.
*
* Walks up from CWD looking for CLAUDE.md files, extracts the section
* under the `# claude up` heading, and executes it as a shell script.
*
* ANT-only command (USER_TYPE === "ant").
*/
export async function up(): Promise<void> {
const cwd = process.cwd()
const gitRoot = findGitRoot(cwd)
const searchDirs = gitRoot ? [gitRoot, cwd] : [cwd]
let upSection: string | null = null
for (const dir of searchDirs) {
const claudeMdPath = join(dir, 'CLAUDE.md')
try {
const content = readFileSync(claudeMdPath, 'utf-8')
upSection = extractUpSection(content)
if (upSection) {
console.log(`Found "# claude up" in ${claudeMdPath}`)
break
}
} catch {
// File not found — continue searching
}
}
if (!upSection) {
console.log(
'No "# claude up" section found in CLAUDE.md.\n' +
'Add a section like:\n\n' +
' # claude up\n' +
' ```bash\n' +
' npm install\n' +
' npm run build\n' +
' ```',
)
return
}
console.log('Running:\n')
console.log(upSection)
console.log()
const result = spawnSync('bash', ['-c', upSection], {
cwd,
stdio: 'inherit',
})
if (result.status !== 0) {
console.error(`\nclaude up failed with exit code ${result.status}`)
process.exitCode = result.status ?? 1
} else {
console.log('\nclaude up completed successfully.')
}
}
/**
* Extract the content under "# claude up" heading from markdown.
* Returns the text between `# claude up` and the next `#` heading (or EOF).
* Strips fenced code block markers if present.
*/
function extractUpSection(markdown: string): string | null {
const lines = markdown.split('\n')
let inSection = false
const sectionLines: string[] = []
for (const line of lines) {
if (/^#\s+claude\s+up\b/i.test(line)) {
inSection = true
continue
}
if (inSection && /^#\s/.test(line)) {
break
}
if (inSection) {
sectionLines.push(line)
}
}
if (sectionLines.length === 0) return null
// Strip fenced code block markers
let text = sectionLines.join('\n').trim()
text = text.replace(/^```\w*\n?/, '').replace(/\n?```\s*$/, '')
return text.trim() || null
}

View File

@@ -25,6 +25,7 @@ import ide from './commands/ide/index.js'
import init from './commands/init.js'
import initVerifiers from './commands/init-verifiers.js'
import keybindings from './commands/keybindings/index.js'
import lang from './commands/lang/index.js'
import login from './commands/login/index.js'
import logout from './commands/logout/index.js'
import installGitHubApp from './commands/install-github-app/index.js'
@@ -111,6 +112,13 @@ const ultraplan = feature('ULTRAPLAN')
? require('./commands/ultraplan.js').default
: null
const torch = feature('TORCH') ? require('./commands/torch.js').default : null
const daemonCmd =
feature('DAEMON') || feature('BG_SESSIONS')
? require('./commands/daemon/index.js').default
: null
const jobCmd = feature('TEMPLATES')
? require('./commands/job/index.js').default
: null
const peersCmd = feature('UDS_INBOX')
? (
require('./commands/peers/index.js') as typeof import('./commands/peers/index.js')
@@ -182,6 +190,7 @@ import sandboxToggle from './commands/sandbox-toggle/index.js'
import chrome from './commands/chrome/index.js'
import stickers from './commands/stickers/index.js'
import advisor from './commands/advisor.js'
import autonomy from './commands/autonomy.js'
import provider from './commands/provider.js'
import { logError } from './utils/log.js'
import { toError } from './utils/errors.js'
@@ -290,6 +299,7 @@ export const INTERNAL_ONLY_COMMANDS = [
const COMMANDS = memoize((): Command[] => [
addDir,
advisor,
autonomy,
provider,
agents,
branch,
@@ -315,6 +325,7 @@ const COMMANDS = memoize((): Command[] => [
ide,
init,
keybindings,
lang,
installGitHubApp,
installSlackApp,
mcp,
@@ -384,6 +395,8 @@ const COMMANDS = memoize((): Command[] => [
...(workflowsCmd ? [workflowsCmd] : []),
...(ultraplan ? [ultraplan] : []),
...(torch ? [torch] : []),
...(daemonCmd ? [daemonCmd] : []),
...(jobCmd ? [jobCmd] : []),
...(process.env.USER_TYPE === 'ant' && !process.env.IS_DEMO
? INTERNAL_ONLY_COMMANDS
: []),

View File

@@ -0,0 +1,246 @@
import { afterEach, beforeEach, describe, expect, test } from 'bun:test'
import autonomyCommand from '../autonomy'
import type { LocalCommandResult } from '../../types/command'
import {
resetStateForTests,
setOriginalCwd,
setProjectRoot,
} from '../../bootstrap/state'
function expectTextResult(
result: LocalCommandResult,
): asserts result is Extract<LocalCommandResult, { type: 'text' }> {
if (result.type !== 'text')
throw new Error(`Expected text result, got ${result.type}`)
}
import { listAutonomyFlows } from '../../utils/autonomyFlows'
import {
createAutonomyQueuedPrompt,
markAutonomyRunCompleted,
startManagedAutonomyFlowFromHeartbeatTask,
} from '../../utils/autonomyRuns'
import {
enqueuePendingNotification,
getCommandQueueSnapshot,
resetCommandQueue,
} from '../../utils/messageQueueManager'
import { cleanupTempDir, createTempDir } from '../../../tests/mocks/file-system'
let tempDir = ''
beforeEach(async () => {
tempDir = await createTempDir('autonomy-command-')
resetStateForTests()
resetCommandQueue()
setOriginalCwd(tempDir)
setProjectRoot(tempDir)
})
afterEach(async () => {
resetStateForTests()
resetCommandQueue()
if (tempDir) {
await cleanupTempDir(tempDir)
}
})
describe('/autonomy', () => {
test('status reports autonomy runs and managed flows separately', async () => {
const plainRun = await createAutonomyQueuedPrompt({
basePrompt: 'scheduled prompt',
trigger: 'scheduled-task',
rootDir: tempDir,
currentDir: tempDir,
sourceLabel: 'nightly',
})
expect(plainRun).not.toBeNull()
await markAutonomyRunCompleted(plainRun!.autonomy!.runId, tempDir)
await startManagedAutonomyFlowFromHeartbeatTask({
task: {
name: 'weekly-report',
interval: '7d',
prompt: 'Ship the weekly report',
steps: [
{
name: 'gather',
prompt: 'Gather weekly inputs',
},
{
name: 'draft',
prompt: 'Draft the weekly report',
},
],
},
rootDir: tempDir,
currentDir: tempDir,
})
const mod = await autonomyCommand.load()
const result = await mod.call('', {} as any)
expectTextResult(result)
expect(result.value).toContain('Autonomy runs: 2')
expect(result.value).toContain('Autonomy flows: 1')
expect(result.value).toContain('Completed: 1')
expect(result.value).toContain('Queued: 1')
})
test('runs subcommand lists recent autonomy runs', async () => {
const queued = await createAutonomyQueuedPrompt({
basePrompt: '<tick>12:00:00</tick>',
trigger: 'proactive-tick',
rootDir: tempDir,
currentDir: tempDir,
})
const mod = await autonomyCommand.load()
const result = await mod.call('runs 5', {} as any)
expectTextResult(result)
expect(result.value).toContain(queued!.autonomy!.runId)
expect(result.value).toContain('proactive-tick')
})
test('flows subcommand lists managed flows and flow subcommand shows detail', async () => {
await startManagedAutonomyFlowFromHeartbeatTask({
task: {
name: 'weekly-report',
interval: '7d',
prompt: 'Ship the weekly report',
steps: [
{
name: 'gather',
prompt: 'Gather weekly inputs',
},
{
name: 'draft',
prompt: 'Draft the weekly report',
},
],
},
rootDir: tempDir,
currentDir: tempDir,
})
const [flow] = await listAutonomyFlows(tempDir)
const mod = await autonomyCommand.load()
const flowsResult = await mod.call('flows 5', {} as any)
expectTextResult(flowsResult)
expect(flowsResult.value).toContain(flow!.flowId)
expect(flowsResult.value).toContain('managed')
const flowResult = await mod.call(`flow ${flow!.flowId}`, {} as any)
expectTextResult(flowResult)
expect(flowResult.value).toContain(`Flow: ${flow!.flowId}`)
expect(flowResult.value).toContain('Mode: managed')
expect(flowResult.value).toContain('Current step: gather')
})
test('flow resume queues the next waiting step', async () => {
const waitingStart = await startManagedAutonomyFlowFromHeartbeatTask({
task: {
name: 'weekly-report',
interval: '7d',
prompt: 'Ship the weekly report',
steps: [
{
name: 'gather',
prompt: 'Gather weekly inputs',
waitFor: 'manual',
},
{
name: 'draft',
prompt: 'Draft the weekly report',
},
],
},
rootDir: tempDir,
currentDir: tempDir,
})
expect(waitingStart).toBeNull()
const [flow] = await listAutonomyFlows(tempDir)
const mod = await autonomyCommand.load()
const result = await mod.call(`flow resume ${flow!.flowId}`, {} as any)
expectTextResult(result)
expect(result.value).toContain('Queued the next managed step')
expect(getCommandQueueSnapshot()).toHaveLength(1)
expect(getCommandQueueSnapshot()[0]!.autonomy?.flowId).toBe(flow!.flowId)
})
test('flow cancel removes queued managed steps and marks the flow cancelled', async () => {
const queued = await startManagedAutonomyFlowFromHeartbeatTask({
task: {
name: 'weekly-report',
interval: '7d',
prompt: 'Ship the weekly report',
steps: [
{
name: 'gather',
prompt: 'Gather weekly inputs',
},
{
name: 'draft',
prompt: 'Draft the weekly report',
},
],
},
rootDir: tempDir,
currentDir: tempDir,
})
expect(queued).not.toBeNull()
enqueuePendingNotification(queued!)
expect(getCommandQueueSnapshot()).toHaveLength(1)
const [flow] = await listAutonomyFlows(tempDir)
const mod = await autonomyCommand.load()
const result = await mod.call(`flow cancel ${flow!.flowId}`, {} as any)
const [cancelledFlow] = await listAutonomyFlows(tempDir)
expectTextResult(result)
expect(result.value).toContain('Cancelled flow')
expect(cancelledFlow!.status).toBe('cancelled')
expect(getCommandQueueSnapshot()).toHaveLength(0)
})
test('flow cancel refuses to rewrite a terminal managed flow', async () => {
const queued = await startManagedAutonomyFlowFromHeartbeatTask({
task: {
name: 'weekly-report',
interval: '7d',
prompt: 'Ship the weekly report',
steps: [
{
name: 'gather',
prompt: 'Gather weekly inputs',
},
],
},
rootDir: tempDir,
currentDir: tempDir,
})
await markAutonomyRunCompleted(queued!.autonomy!.runId, tempDir)
const [flow] = await listAutonomyFlows(tempDir)
const mod = await autonomyCommand.load()
const result = await mod.call(`flow cancel ${flow!.flowId}`, {} as any)
const [terminalFlow] = await listAutonomyFlows(tempDir)
expectTextResult(result)
expect(result.value).toContain('already terminal')
expect(terminalFlow!.status).toBe('succeeded')
})
test('invalid subcommands return usage text', async () => {
const mod = await autonomyCommand.load()
const result = await mod.call('unknown', {} as any)
expectTextResult(result)
expect(result.value).toContain('Usage: /autonomy')
})
})

View File

@@ -0,0 +1,48 @@
import { beforeEach, describe, expect, test } from 'bun:test'
import proactiveCommand from '../proactive'
import {
activateProactive,
deactivateProactive,
isProactiveActive,
} from '../../proactive/index'
beforeEach(() => {
deactivateProactive()
})
describe('/proactive baseline', () => {
test('invoking the command enables proactive mode and emits a system reminder', async () => {
const mod = await proactiveCommand.load()
let resultText: string | undefined
let options: Parameters<Parameters<typeof mod.call>[0]>[1] | undefined
await mod.call((result, opts) => {
resultText = result
options = opts
}, {} as any)
expect(isProactiveActive()).toBe(true)
expect(resultText).toContain('Proactive mode enabled')
expect(options?.display).toBe('system')
expect(options?.metaMessages?.[0]).toContain(
'Proactive mode is now enabled',
)
})
test('invoking the command again disables proactive mode', async () => {
const mod = await proactiveCommand.load()
activateProactive('test')
let resultText: string | undefined
let options: Parameters<Parameters<typeof mod.call>[0]>[1] | undefined
await mod.call((result, opts) => {
resultText = result
options = opts
}, {} as any)
expect(isProactiveActive()).toBe(false)
expect(resultText).toBe('Proactive mode disabled')
expect(options?.display).toBe('system')
})
})

View File

@@ -1,53 +0,0 @@
import * as React from 'react'
import type { LocalJSXCommandContext } from '../../commands.js'
import type { LocalJSXCommandOnDone } from '../../types/command.js'
import type { AppState } from '../../state/AppState.js'
/** Stub — install wizard is not yet restored. */
export async function computeDefaultInstallDir(): Promise<string> {
return ''
}
/** Stub — install wizard is not yet restored. */
export function NewInstallWizard(_props: {
defaultDir: string
onInstalled: (dir: string) => void
onCancel: () => void
onError: (message: string) => void
}): React.ReactNode {
return null
}
/**
* /assistant command implementation.
*
* Opens the Kairos assistant panel. In the current build the panel is
* rendered by the REPL layer when kairosActive is true; the slash command
* simply toggles visibility and prints a confirmation line.
*/
export async function call(
onDone: LocalJSXCommandOnDone,
context: LocalJSXCommandContext,
_args: string,
): Promise<React.ReactNode> {
const { setAppState, getAppState } = context
const current = getAppState()
const isVisible = (current as Record<string, unknown>).assistantPanelVisible
if (isVisible) {
setAppState((prev: AppState) => ({
...prev,
assistantPanelVisible: false,
} as AppState))
onDone('Assistant panel hidden.', { display: 'system' })
} else {
setAppState((prev: AppState) => ({
...prev,
assistantPanelVisible: true,
} as AppState))
onDone('Assistant panel opened.', { display: 'system' })
}
return null
}

View File

@@ -0,0 +1,175 @@
import { spawn } from 'child_process';
import * as React from 'react';
import { useEffect, useState } from 'react';
import { resolve } from 'path';
import { Box, Text } from '@anthropic/ink';
import { Dialog } from '../../components/design-system/Dialog.js';
import { ListItem } from '../../components/design-system/ListItem.js';
import { useRegisterOverlay } from '../../context/overlayContext.js';
import { useKeybindings } from '../../keybindings/useKeybinding.js';
import { findGitRoot } from '../../utils/git.js';
import { getKairosActive, setKairosActive } from '../../bootstrap/state.js';
import type { LocalJSXCommandContext } from '../../commands.js';
import type { LocalJSXCommandOnDone } from '../../types/command.js';
import type { AppState } from '../../state/AppState.js';
/**
* Compute the default directory for assistant daemon installation.
* Prefers git root of cwd; falls back to cwd itself.
*/
export async function computeDefaultInstallDir(): Promise<string> {
const cwd = process.cwd();
const gitRoot = findGitRoot(cwd);
return gitRoot || resolve(cwd);
}
interface WizardProps {
defaultDir: string;
onInstalled: (dir: string) => void;
onCancel: () => void;
onError: (message: string) => void;
}
/**
* Install wizard for assistant mode. Shown when `claude assistant` finds
* zero CCR sessions. Guides the user to start a daemon that registers
* a bridge → CCR cloud session.
*
* After installation, main.tsx tells the user to run `claude assistant`
* again in a few seconds (daemon needs time to register the bridge session).
*/
export function NewInstallWizard({ defaultDir, onInstalled, onCancel, onError }: WizardProps): React.ReactNode {
useRegisterOverlay('assistant-install-wizard');
const [focusIndex, setFocusIndex] = useState(0);
const [starting, setStarting] = useState(false);
useKeybindings(
{
'select:next': () => setFocusIndex(i => (i + 1) % 2),
'select:previous': () => setFocusIndex(i => (i - 1 + 2) % 2),
'select:accept': () => {
if (focusIndex === 0) {
startDaemon();
} else {
onCancel();
}
},
},
{ context: 'Select' },
);
function startDaemon(): void {
if (starting) return;
setStarting(true);
const dir = defaultDir || resolve('.');
try {
const execArgs = [...process.execArgv, process.argv[1]!, 'daemon', 'start', `--dir=${dir}`];
const child = spawn(process.execPath, execArgs, {
cwd: dir,
stdio: 'ignore',
detached: true,
});
child.unref();
child.on('error', err => {
onError(`Failed to start daemon: ${err.message}`);
});
// Give the daemon a moment to initialize, then report success.
// The daemon still needs several more seconds to register the bridge
// and create a CCR session — main.tsx will tell the user to reconnect.
setTimeout(() => {
onInstalled(dir);
}, 1500);
} catch (err) {
onError(`Failed to start daemon: ${err instanceof Error ? err.message : String(err)}`);
}
}
if (starting) {
return (
<Dialog title="Assistant Setup" onCancel={onCancel} hideInputGuide>
<Text>Starting daemon in {defaultDir}...</Text>
</Dialog>
);
}
return (
<Dialog title="Assistant Setup" onCancel={onCancel} hideInputGuide>
<Box flexDirection="column" gap={1}>
<Text>No active assistant sessions found.</Text>
<Text>
Start a daemon in <Text bold>{defaultDir || '.'}</Text> to create a cloud session?
</Text>
<Box flexDirection="column">
<ListItem isFocused={focusIndex === 0}>
<Text>Start assistant daemon</Text>
</ListItem>
<ListItem isFocused={focusIndex === 1}>
<Text>Cancel</Text>
</ListItem>
</Box>
<Text dimColor>Enter to select · Esc to cancel</Text>
</Box>
</Dialog>
);
}
/**
* /assistant command implementation.
*
* First invocation activates KAIROS (sets kairosActive, enables brief
* and proactive tools). Subsequent invocations toggle the assistant panel.
*/
export async function call(
onDone: LocalJSXCommandOnDone,
context: LocalJSXCommandContext,
_args: string,
): Promise<React.ReactNode> {
const { setAppState, getAppState } = context;
// First invocation: activate KAIROS
if (!getKairosActive()) {
setKairosActive(true);
setAppState(
(prev: AppState) =>
({
...prev,
kairosEnabled: true,
assistantPanelVisible: true,
}) as AppState,
);
onDone('KAIROS assistant mode activated.', { display: 'system' });
return null;
}
// Subsequent invocations: toggle panel visibility
const current = getAppState();
const isVisible = (current as Record<string, unknown>).assistantPanelVisible;
if (isVisible) {
setAppState(
(prev: AppState) =>
({
...prev,
assistantPanelVisible: false,
}) as AppState,
);
onDone('Assistant panel hidden.', { display: 'system' });
} else {
setAppState(
(prev: AppState) =>
({
...prev,
assistantPanelVisible: true,
}) as AppState,
);
onDone('Assistant panel opened.', { display: 'system' });
}
return null;
}

View File

@@ -1,25 +1,21 @@
import { feature } from 'bun:bundle'
import { getKairosActive } from '../../bootstrap/state.js'
import { getFeatureValue_CACHED_MAY_BE_STALE } from '../../services/analytics/growthbook.js'
/**
* Runtime gate for the /assistant command.
* Runtime gate for the /assistant command visibility.
*
* Build-time: feature('KAIROS') must be on (checked in commands.ts before
* the module is even required).
* Build-time: feature('KAIROS') must be on.
* Runtime: tengu_kairos_assistant GrowthBook flag (remote kill switch).
*
* Runtime: tengu_kairos_assistant GrowthBook flag acts as a remote kill
* switch, and kairosActive state must be true (set during bootstrap when
* the session qualifies for KAIROS features).
* Does NOT require kairosActive — the /assistant command is visible
* before activation so users can invoke it to activate KAIROS.
*/
export function isAssistantEnabled(): boolean {
if (!feature('KAIROS')) {
return false
}
if (
!getFeatureValue_CACHED_MAY_BE_STALE('tengu_kairos_assistant', false)
) {
if (!getFeatureValue_CACHED_MAY_BE_STALE('tengu_kairos_assistant', false)) {
return false
}
return getKairosActive()
return true
}

125
src/commands/autonomy.ts Normal file
View File

@@ -0,0 +1,125 @@
import type { Command, LocalCommandCall } from '../types/command.js'
import {
formatAutonomyFlowDetail,
formatAutonomyFlowsList,
formatAutonomyFlowsStatus,
getAutonomyFlowById,
listAutonomyFlows,
requestManagedAutonomyFlowCancel,
} from '../utils/autonomyFlows.js'
import {
formatAutonomyRunsList,
formatAutonomyRunsStatus,
listAutonomyRuns,
markAutonomyRunCancelled,
resumeManagedAutonomyFlowPrompt,
} from '../utils/autonomyRuns.js'
import {
enqueuePendingNotification,
removeByFilter,
} from '../utils/messageQueueManager.js'
function parseRunsLimit(raw?: string): number {
const parsed = Number.parseInt(raw ?? '', 10)
if (!Number.isFinite(parsed) || parsed <= 0) {
return 10
}
return Math.min(parsed, 50)
}
const call: LocalCommandCall = async (args: string) => {
const [subcommand = 'status', arg1, arg2] = args.trim().split(/\s+/, 3)
const runs = await listAutonomyRuns()
const flows = await listAutonomyFlows()
if (subcommand === 'runs') {
return {
type: 'text',
value: formatAutonomyRunsList(runs, parseRunsLimit(arg1)),
}
}
if (subcommand === 'flows') {
return {
type: 'text',
value: formatAutonomyFlowsList(flows, parseRunsLimit(arg1)),
}
}
if (subcommand === 'flow') {
if (arg1 === 'cancel') {
const flowId = arg2 ?? ''
const cancelled = await requestManagedAutonomyFlowCancel({ flowId })
if (!cancelled) {
return {
type: 'text',
value: 'Autonomy flow not found.',
}
}
if (!cancelled.accepted) {
return {
type: 'text',
value: `Autonomy flow ${flowId} is already terminal (${cancelled.flow.status}).`,
}
}
const removed = removeByFilter(cmd => cmd.autonomy?.flowId === flowId)
for (const command of removed) {
if (command.autonomy?.runId) {
await markAutonomyRunCancelled(command.autonomy.runId)
}
}
return {
type: 'text',
value:
cancelled.flow.status === 'running'
? `Cancellation requested for flow ${flowId}. The current step is still running, and no new steps will be started.`
: `Cancelled flow ${flowId}. Removed ${removed.length} queued step(s).`,
}
}
if (arg1 === 'resume') {
const flowId = arg2 ?? ''
const command = await resumeManagedAutonomyFlowPrompt({ flowId })
if (!command) {
return {
type: 'text',
value: 'Autonomy flow is not waiting or was not found.',
}
}
enqueuePendingNotification(command)
return {
type: 'text',
value: `Queued the next managed step for flow ${flowId}.`,
}
}
return {
type: 'text',
value: formatAutonomyFlowDetail(await getAutonomyFlowById(arg1 ?? '')),
}
}
if (subcommand !== 'status' && subcommand !== '') {
return {
type: 'text',
value:
'Usage: /autonomy [status|runs [limit]|flows [limit]|flow <id>|flow cancel <id>|flow resume <id>]',
}
}
return {
type: 'text',
value: [formatAutonomyRunsStatus(runs), formatAutonomyFlowsStatus(flows)].join('\n'),
}
}
const autonomy = {
type: 'local',
name: 'autonomy',
description:
'Inspect automatic autonomy runs recorded for proactive ticks and scheduled tasks',
supportsNonInteractive: true,
load: () => Promise.resolve({ call }),
} satisfies Command
export default autonomy

View File

@@ -0,0 +1,24 @@
import { describe, test, expect } from 'bun:test'
describe('/daemon command', () => {
test('index exports a valid Command', async () => {
const mod = await import('../index.js')
const cmd = mod.default
expect(cmd.name).toBe('daemon')
expect(cmd.type).toBe('local-jsx')
expect(typeof cmd.load).toBe('function')
expect(cmd.description).toContain('daemon')
})
test('daemon module exports call function', async () => {
const mod = await import('../daemon.js')
expect(typeof mod.call).toBe('function')
})
test('argumentHint lists subcommands', async () => {
const mod = await import('../index.js')
const cmd = mod.default
expect(cmd.argumentHint).toContain('status')
expect(cmd.argumentHint).toContain('bg')
})
})

View File

@@ -0,0 +1,57 @@
import type {
LocalJSXCommandOnDone,
LocalJSXCommandContext,
} from '../../types/command.js'
/**
* /daemon slash command — manages daemon and background sessions from the REPL.
*
* Subcommands: status | start | stop | bg | attach | logs | kill
* Default (no args): status
*/
export async function call(
onDone: LocalJSXCommandOnDone,
_context: LocalJSXCommandContext,
args: string,
): Promise<React.ReactNode> {
const parts = args ? args.trim().split(/\s+/) : []
const sub = parts[0] || 'status'
// attach is interactive/blocking — not available inside the REPL
if (sub === 'attach') {
onDone(
'Use `claude daemon attach` from the CLI. Attach is not available inside the REPL.',
{ display: 'system' },
)
return null
}
// For all other subcommands, capture console output and return via onDone
const lines = await captureConsole(async () => {
if (sub === 'bg') {
const bg = await import('../../cli/bg.js')
await bg.handleBgStart(parts.slice(1))
} else {
const { daemonMain } = await import('../../daemon/main.js')
await daemonMain([sub, ...parts.slice(1)])
}
})
onDone(lines.join('\n') || 'Done.', { display: 'system' })
return null
}
async function captureConsole(fn: () => Promise<void>): Promise<string[]> {
const lines: string[] = []
const origLog = console.log
const origError = console.error
console.log = (...a: unknown[]) => lines.push(a.map(String).join(' '))
console.error = (...a: unknown[]) => lines.push(a.map(String).join(' '))
try {
await fn()
} finally {
console.log = origLog
console.error = origError
}
return lines
}

View File

@@ -0,0 +1,17 @@
import type { Command } from '../../commands.js'
import { feature } from 'bun:bundle'
const daemon = {
type: 'local-jsx',
name: 'daemon',
description: 'Manage background sessions and daemon',
argumentHint: '[status|start|stop|bg|attach|logs|kill]',
isEnabled: () => {
if (feature('DAEMON')) return true
if (feature('BG_SESSIONS')) return true
return false
},
load: () => import('./daemon.js'),
} satisfies Command
export default daemon

View File

@@ -1,6 +1,7 @@
import { feature } from 'bun:bundle'
import type { Command } from '../commands.js'
import { maybeMarkProjectOnboardingComplete } from '../projectOnboardingState.js'
import { AUTONOMY_AGENTS_PATH_POSIX } from '../utils/autonomyAuthority.js'
import { isEnvTruthy } from '../utils/envUtils.js'
const OLD_INIT_PROMPT = `Please analyze this codebase and create a CLAUDE.md file, which will be given to future instances of Claude Code to operate in this repository.
@@ -43,7 +44,7 @@ Use AskUserQuestion to find out what the user wants:
## Phase 2: Explore the codebase
Launch a subagent to survey the codebase, and ask it to read key files to understand the project: manifest files (package.json, Cargo.toml, pyproject.toml, go.mod, pom.xml, etc.), README, Makefile/build configs, CI config, existing CLAUDE.md, .claude/rules/, AGENTS.md, .cursor/rules or .cursorrules, .github/copilot-instructions.md, .windsurfrules, .clinerules, .mcp.json.
Launch a subagent to survey the codebase, and ask it to read key files to understand the project: manifest files (package.json, Cargo.toml, pyproject.toml, go.mod, pom.xml, etc.), README, Makefile/build configs, CI config, existing CLAUDE.md, .claude/rules/, ${AUTONOMY_AGENTS_PATH_POSIX}, .cursor/rules or .cursorrules, .github/copilot-instructions.md, .windsurfrules, .clinerules, .mcp.json.
Detect:
- Build, test, and lint commands (especially non-standard ones)
@@ -105,7 +106,7 @@ Include:
- Repo etiquette (branch naming, PR conventions, commit style)
- Required env vars or setup steps
- Non-obvious gotchas or architectural decisions
- Important parts from existing AI coding tool configs if they exist (AGENTS.md, .cursor/rules, .cursorrules, .github/copilot-instructions.md, .windsurfrules, .clinerules)
- Important parts from existing AI coding tool configs if they exist (${AUTONOMY_AGENTS_PATH_POSIX}, .cursor/rules, .cursorrules, .github/copilot-instructions.md, .windsurfrules, .clinerules)
Exclude:
- File-by-file structure or component lists (Claude can discover these by reading the codebase)

View File

@@ -0,0 +1,25 @@
import { describe, test, expect } from 'bun:test'
describe('/job command', () => {
test('index exports a valid Command', async () => {
const mod = await import('../index.js')
const cmd = mod.default
expect(cmd.name).toBe('job')
expect(cmd.type).toBe('local-jsx')
expect(typeof cmd.load).toBe('function')
expect(cmd.description).toContain('job')
})
test('job module exports call function', async () => {
const mod = await import('../job.js')
expect(typeof mod.call).toBe('function')
})
test('argumentHint lists subcommands', async () => {
const mod = await import('../index.js')
const cmd = mod.default
expect(cmd.argumentHint).toContain('list')
expect(cmd.argumentHint).toContain('new')
expect(cmd.argumentHint).toContain('status')
})
})

16
src/commands/job/index.ts Normal file
View File

@@ -0,0 +1,16 @@
import type { Command } from '../../commands.js'
import { feature } from 'bun:bundle'
const job = {
type: 'local-jsx',
name: 'job',
description: 'Manage template jobs',
argumentHint: '[list|new|reply|status]',
isEnabled: () => {
if (feature('TEMPLATES')) return true
return false
},
load: () => import('./job.js'),
} satisfies Command
export default job

34
src/commands/job/job.tsx Normal file
View File

@@ -0,0 +1,34 @@
import type { LocalJSXCommandOnDone, LocalJSXCommandContext } from '../../types/command.js'
/**
* /job slash command — manages template jobs from inside the REPL.
*
* Subcommands: list | new <template> [args] | reply <id> <text> | status <id>
* Default (no args): list
*/
export async function call(
onDone: LocalJSXCommandOnDone,
_context: LocalJSXCommandContext,
args: string,
): Promise<React.ReactNode> {
const parts = args ? args.trim().split(/\s+/) : []
const sub = parts[0] || 'list'
// Capture console output so we can return it as onDone text
const lines: string[] = []
const origLog = console.log
const origError = console.error
console.log = (...a: unknown[]) => lines.push(a.map(String).join(' '))
console.error = (...a: unknown[]) => lines.push(a.map(String).join(' '))
try {
const { templatesMain } = await import('../../cli/handlers/templateJobs.js')
await templatesMain([sub, ...parts.slice(1)])
} finally {
console.log = origLog
console.error = origError
}
onDone(lines.join('\n') || 'Done.', { display: 'system' })
return null
}

View File

@@ -0,0 +1,12 @@
import type { Command } from '../../commands.js'
const lang = {
type: 'local-jsx',
name: 'lang',
description: 'Set display language (en/zh/auto)',
immediate: true,
argumentHint: '<en|zh|auto>',
load: () => import('./lang.js'),
} satisfies Command
export default lang

49
src/commands/lang/lang.ts Normal file
View File

@@ -0,0 +1,49 @@
import type { ToolUseContext } from '../../Tool.js'
import type {
LocalJSXCommandContext,
LocalJSXCommandOnDone,
} from '../../types/command.js'
import { getGlobalConfig, saveGlobalConfig } from '../../utils/config.js'
import {
type PreferredLanguage,
getLanguageDisplayName,
getResolvedLanguage,
} from '../../utils/language.js'
const VALID_LANGS: readonly PreferredLanguage[] = ['en', 'zh', 'auto']
export async function call(
onDone: LocalJSXCommandOnDone,
_context: ToolUseContext & LocalJSXCommandContext,
args: string,
): Promise<null> {
const arg = args.trim().toLowerCase()
if (!arg) {
const pref = getGlobalConfig().preferredLanguage ?? 'auto'
const resolved = getResolvedLanguage()
const suffix =
pref === 'auto' ? `${getLanguageDisplayName(resolved)}` : ''
onDone(`Language: ${getLanguageDisplayName(pref)}${suffix}`, {
display: 'system',
})
return null
}
if (!VALID_LANGS.includes(arg as PreferredLanguage)) {
onDone(`Invalid language "${arg}". Use: en, zh, or auto`, {
display: 'system',
})
return null
}
const lang = arg as PreferredLanguage
saveGlobalConfig(current => ({ ...current, preferredLanguage: lang }))
const resolved = getResolvedLanguage()
const suffix = lang === 'auto' ? `${getLanguageDisplayName(resolved)}` : ''
onDone(`Language set to ${getLanguageDisplayName(lang)}${suffix}`, {
display: 'system',
})
return null
}

View File

@@ -1,6 +1,11 @@
import type { LocalCommandCall } from '../../types/command.js'
import { getSlaveClient } from '../../hooks/useMasterMonitor.js'
import { getPipeIpc } from '../../utils/pipeTransport.js'
import {
addSendOverride,
removeSendOverride,
removeMasterPipeMute,
} from '../../utils/pipeMuteState.js'
export const call: LocalCommandCall = async (args, context) => {
const currentState = context.getAppState()
@@ -48,6 +53,12 @@ export const call: LocalCommandCall = async (args, context) => {
}
try {
// Temporarily override mute for this slave so its response is visible.
// Override lasts until the slave emits 'done' or 'error' (cleared by
// useMasterMonitor's attachPipeEntryEmitter handler).
addSendOverride(targetName)
removeMasterPipeMute(targetName)
client.send({ type: 'relay_unmute' })
client.send({
type: 'prompt',
data: message,
@@ -89,6 +100,8 @@ export const call: LocalCommandCall = async (args, context) => {
value: `Sent to "${targetName}": ${message.slice(0, 100)}${message.length > 100 ? '...' : ''}`,
}
} catch (err) {
// Roll back override on send failure to prevent permanent unmute
removeSendOverride(targetName)
return {
type: 'text',
value: `Failed to send to "${targetName}": ${err instanceof Error ? err.message : String(err)}`,

View File

@@ -1 +1,19 @@
export default null
import type { Command, LocalJSXCommandOnDone } from '../types/command.js'
import type { ReactNode } from 'react'
const call = async (onDone: LocalJSXCommandOnDone): Promise<ReactNode> => {
onDone(
'torch: Reserved internal debug command. No implementation is available in this build.',
{ display: 'system' },
)
return null
}
export default {
type: 'local-jsx',
name: 'torch',
description: '[INTERNAL] Development debug command (reserved)',
isEnabled: () => true,
isHidden: true,
load: () => Promise.resolve({ call }),
} satisfies Command

View File

@@ -0,0 +1,61 @@
/**
* Tests for daemon/main.ts subcommand routing.
*
* The `status` and `bg` subcommands trigger dynamic imports of `cli/bg.ts`
* which depends on `envUtils.ts` → `lodash-es/memoize.js` (unavailable in
* raw test context without `bun run dev`'s define flags). We test only the
* self-contained subcommands: help and unknown.
*/
import { describe, test, expect, beforeEach, afterEach } from 'bun:test'
describe('daemonMain subcommand routing', () => {
const origLog = console.log
const origError = console.error
let logLines: string[]
beforeEach(() => {
logLines = []
console.log = (...a: unknown[]) => logLines.push(a.map(String).join(' '))
console.error = (...a: unknown[]) => logLines.push(a.map(String).join(' '))
})
afterEach(() => {
console.log = origLog
console.error = origError
process.exitCode = 0
})
test('unknown subcommand sets exitCode to 1', async () => {
const { daemonMain } = await import('../main.js')
await daemonMain(['unknown-command-xyz'])
expect(process.exitCode).toBe(1)
})
test('help subcommand prints usage', async () => {
const { daemonMain } = await import('../main.js')
await daemonMain(['help'])
const output = logLines.join('\n')
expect(output).toContain('SUBCOMMANDS')
expect(output).toContain('status')
expect(output).toContain('start')
expect(output).toContain('stop')
expect(output).toContain('bg')
expect(output).toContain('attach')
expect(output).toContain('logs')
expect(output).toContain('kill')
})
test('--help is alias for help', async () => {
const { daemonMain } = await import('../main.js')
await daemonMain(['--help'])
const output = logLines.join('\n')
expect(output).toContain('SUBCOMMANDS')
})
test('-h is alias for help', async () => {
const { daemonMain } = await import('../main.js')
await daemonMain(['-h'])
const output = logLines.join('\n')
expect(output).toContain('SUBCOMMANDS')
})
})

View File

@@ -0,0 +1,185 @@
/**
* Tests for src/daemon/state.ts
*
* Uses real temp directories and CLAUDE_CONFIG_DIR env var
* instead of mocking fs/envUtils, to avoid cross-test mock pollution.
*/
import { describe, expect, test, beforeEach, afterAll } from 'bun:test'
import { mkdtempSync, rmSync, existsSync, readFileSync } from 'fs'
import { join } from 'path'
import { tmpdir } from 'os'
import { getClaudeConfigHomeDir } from '../../utils/envUtils.js'
// ─── setup: real temp dir via env var ──────────────────────────────────────
const tempBase = mkdtempSync(join(tmpdir(), 'daemon-state-test-'))
beforeEach(() => {
// Clear lodash memoize cache so CLAUDE_CONFIG_DIR env var takes effect
if (
typeof getClaudeConfigHomeDir === 'function' &&
'cache' in getClaudeConfigHomeDir
) {
;(getClaudeConfigHomeDir as any).cache.clear?.()
}
const tempHome = mkdtempSync(join(tempBase, 'home-'))
process.env.CLAUDE_CONFIG_DIR = tempHome
})
afterAll(() => {
delete process.env.CLAUDE_CONFIG_DIR
// Clear memoize cache after all tests so other files see fresh state
if (
typeof getClaudeConfigHomeDir === 'function' &&
'cache' in getClaudeConfigHomeDir
) {
;(getClaudeConfigHomeDir as any).cache.clear?.()
}
try {
rmSync(tempBase, { recursive: true, force: true })
} catch {
// best-effort cleanup
}
})
// ─── import ─────────────────────────────────────────────────────────────────
const {
getDaemonStateFilePath,
writeDaemonState,
readDaemonState,
removeDaemonState,
queryDaemonStatus,
} = await import('../state.js')
// ─── tests ─────────────────────────────────────────────────────────────────
describe('getDaemonStateFilePath', () => {
test('returns default path with remote-control name', () => {
const p = getDaemonStateFilePath()
expect(p).toContain('daemon')
expect(p).toContain('remote-control.json')
})
test('returns path with custom name', () => {
const p = getDaemonStateFilePath('my-daemon')
expect(p).toContain('my-daemon.json')
})
})
describe('writeDaemonState', () => {
test('writes state JSON to disk', () => {
const state = {
pid: 1234,
cwd: '/test',
startedAt: '2026-01-01T00:00:00Z',
workerKinds: ['rcs'],
lastStatus: 'running' as const,
}
writeDaemonState(state, 'test')
const filePath = getDaemonStateFilePath('test')
expect(existsSync(filePath)).toBe(true)
const parsed = JSON.parse(readFileSync(filePath, 'utf-8'))
expect(parsed.pid).toBe(1234)
expect(parsed.cwd).toBe('/test')
})
test('creates directory recursively', () => {
writeDaemonState(
{
pid: 1,
cwd: '/',
startedAt: '',
workerKinds: [],
lastStatus: 'running',
},
'dir-test',
)
const filePath = getDaemonStateFilePath('dir-test')
expect(existsSync(filePath)).toBe(true)
})
})
describe('readDaemonState', () => {
test('returns null when no state file', () => {
expect(readDaemonState('nonexistent')).toBeNull()
})
test('returns parsed state when file exists', () => {
const state = {
pid: 42,
cwd: '/x',
startedAt: '',
workerKinds: [],
lastStatus: 'running' as const,
}
writeDaemonState(state, 'read-test')
const result = readDaemonState('read-test')
expect(result).not.toBeNull()
expect(result!.pid).toBe(42)
})
})
describe('removeDaemonState', () => {
test('removes existing state file', () => {
writeDaemonState(
{
pid: 1,
cwd: '/',
startedAt: '',
workerKinds: [],
lastStatus: 'running',
},
'rm-test',
)
const filePath = getDaemonStateFilePath('rm-test')
expect(existsSync(filePath)).toBe(true)
removeDaemonState('rm-test')
expect(existsSync(filePath)).toBe(false)
})
test('does not throw when file does not exist', () => {
expect(() => removeDaemonState('no-file')).not.toThrow()
})
})
describe('queryDaemonStatus', () => {
test('returns stopped when no state file', () => {
const result = queryDaemonStatus('empty')
expect(result.status).toBe('stopped')
expect(result.state).toBeUndefined()
})
test('returns running when PID is alive (current process)', () => {
writeDaemonState(
{
pid: process.pid,
cwd: process.cwd(),
startedAt: new Date().toISOString(),
workerKinds: ['test'],
lastStatus: 'running',
},
'alive-test',
)
const result = queryDaemonStatus('alive-test')
expect(result.status).toBe('running')
expect(result.state).toBeDefined()
expect(result.state!.pid).toBe(process.pid)
})
test('returns stale when PID is dead and cleans up', () => {
writeDaemonState(
{
pid: 999999,
cwd: '/',
startedAt: '',
workerKinds: [],
lastStatus: 'running',
},
'stale-test',
)
const result = queryDaemonStatus('stale-test')
expect(result.status).toBe('stale')
expect(existsSync(getDaemonStateFilePath('stale-test'))).toBe(false)
})
})

View File

@@ -1,6 +1,12 @@
import { spawn, type ChildProcess } from 'child_process'
import { resolve } from 'path'
import { errorMessage } from '../utils/errors.js'
import {
writeDaemonState,
removeDaemonState,
queryDaemonStatus,
stopDaemonByPid,
} from './state.js'
/**
* Exit code used by workers for permanent (non-retryable) failures.
@@ -29,30 +35,62 @@ interface WorkerState {
* Daemon supervisor entry point. Called from `cli.tsx` via:
* `claude daemon [subcommand]`
*
* Starts and supervises long-running workers. Currently spawns one
* `remoteControl` worker that runs the headless bridge server.
* Manages the daemon supervisor AND background sessions under one namespace.
*
* Subcommands:
* (none) — start the supervisor with default workers
* start — same as no subcommand
* status — print worker status (TODO: IPC)
* stop send SIGTERM to supervisor (TODO: PID file)
* (none) — unified status (supervisor + sessions)
* start — start the supervisor with default workers
* stop — send SIGTERM to supervisor
* statusunified status (supervisor + sessions)
* ps — alias for status
* bg — start a background session
* attach — attach to a background session
* logs — show session logs
* kill — kill a session
*/
export async function daemonMain(args: string[]): Promise<void> {
const subcommand = args[0] || 'start'
const subcommand = args[0] || 'status'
switch (subcommand) {
// --- Supervisor management ---
case 'start':
await runSupervisor(args.slice(1))
break
case 'status':
console.log('daemon status: not yet implemented (requires IPC)')
break
case 'stop':
console.log('daemon stop: not yet implemented (requires PID file)')
await handleDaemonStop()
break
// --- Unified status ---
case 'status':
case 'ps':
await showUnifiedStatus()
break
// --- Session management (delegates to bg.ts) ---
case 'bg': {
const bg = await import('../cli/bg.js')
await bg.handleBgStart(args.slice(1))
break
}
case 'attach': {
const bg = await import('../cli/bg.js')
await bg.attachHandler(args[1])
break
}
case 'logs': {
const bg = await import('../cli/bg.js')
await bg.logsHandler(args[1])
break
}
case 'kill': {
const bg = await import('../cli/bg.js')
await bg.killHandler(args[1])
break
}
case '--help':
case '-h':
case 'help':
printHelp()
break
default:
@@ -64,17 +102,25 @@ export async function daemonMain(args: string[]): Promise<void> {
function printHelp(): void {
console.log(`
Claude Code Daemon — persistent background supervisor
Claude Code Daemon — background process management
USAGE
claude daemon [subcommand] [options]
claude daemon [subcommand]
SUBCOMMANDS
start Start the daemon supervisor (default)
status Show worker status
status Show daemon and session status (default)
start Start the daemon supervisor
stop Stop the daemon
bg Start a background session
attach Attach to a background session
logs Show session logs
kill Kill a session
help Show this help
OPTIONS
REPL
/daemon [subcommand] Same commands available in interactive mode
OPTIONS (for start)
--dir <path> Working directory (default: current)
--spawn-mode <mode> Worker spawn mode: same-dir | worktree (default: same-dir)
--capacity <N> Max concurrent sessions per worker (default: 4)
@@ -85,6 +131,63 @@ OPTIONS
`)
}
/**
* Show unified status: daemon supervisor + background sessions.
*/
async function showUnifiedStatus(): Promise<void> {
// 1. Daemon supervisor status
const result = queryDaemonStatus()
console.log('=== Daemon Supervisor ===')
switch (result.status) {
case 'running': {
const s = result.state!
console.log(` Status: running`)
console.log(` PID: ${s.pid}`)
console.log(` CWD: ${s.cwd}`)
console.log(` Started: ${s.startedAt}`)
console.log(` Workers: ${s.workerKinds.join(', ')}`)
break
}
case 'stopped':
console.log(' Status: stopped')
break
case 'stale':
console.log(' Status: stale (cleaned up)')
break
}
// 2. Background sessions
console.log('\n=== Background Sessions ===')
const bg = await import('../cli/bg.js')
await bg.psHandler([])
}
/**
* Stop a running daemon from another CLI process.
*/
async function handleDaemonStop(): Promise<void> {
const result = queryDaemonStatus()
if (result.status === 'stopped') {
console.log('daemon is not running')
return
}
if (result.status === 'stale') {
console.log('daemon was stale (cleaned up)')
return
}
console.log(`stopping daemon (PID: ${result.state!.pid})...`)
const stopped = await stopDaemonByPid()
if (stopped) {
console.log('daemon stopped')
} else {
console.log('daemon could not be stopped (may have already exited)')
}
}
/**
* Parse supervisor arguments from CLI.
*/
@@ -140,12 +243,22 @@ async function runSupervisor(args: string[]): Promise<void> {
},
]
// Write daemon state file so other CLI processes can query/stop us
writeDaemonState({
pid: process.pid,
cwd: dir,
startedAt: new Date().toISOString(),
workerKinds: workers.map(w => w.kind),
lastStatus: 'running',
})
const controller = new AbortController()
// Graceful shutdown
const shutdown = () => {
console.log('[daemon] supervisor shutting down...')
controller.abort()
removeDaemonState()
for (const w of workers) {
if (w.process && !w.process.killed) {
w.process.kill('SIGTERM')

157
src/daemon/state.ts Normal file
View File

@@ -0,0 +1,157 @@
import { readFileSync, writeFileSync, mkdirSync, unlinkSync } from 'fs'
import { join, dirname } from 'path'
import { getClaudeConfigHomeDir } from '../utils/envUtils.js'
/**
* Daemon state persisted to disk so that `status` / `stop` can work
* from a different CLI process than the one that started the daemon.
*/
export interface DaemonStateData {
pid: number
cwd: string
startedAt: string
workerKinds: string[]
lastStatus: 'running' | 'stopped' | 'error'
}
export type DaemonStatus = 'running' | 'stopped' | 'stale'
/**
* Returns the path to the daemon state file for a given daemon name.
*/
export function getDaemonStateFilePath(name = 'remote-control'): string {
return join(getClaudeConfigHomeDir(), 'daemon', `${name}.json`)
}
/**
* Write daemon state to disk. Called by the supervisor on startup.
*/
export function writeDaemonState(
state: DaemonStateData,
name = 'remote-control',
): void {
const filePath = getDaemonStateFilePath(name)
mkdirSync(dirname(filePath), { recursive: true })
writeFileSync(filePath, JSON.stringify(state, null, 2), 'utf-8')
}
/**
* Read daemon state from disk. Returns null if no state file exists.
*/
export function readDaemonState(
name = 'remote-control',
): DaemonStateData | null {
const filePath = getDaemonStateFilePath(name)
try {
const raw = readFileSync(filePath, 'utf-8')
return JSON.parse(raw) as DaemonStateData
} catch {
return null
}
}
/**
* Remove the daemon state file.
*/
export function removeDaemonState(name = 'remote-control'): void {
const filePath = getDaemonStateFilePath(name)
try {
unlinkSync(filePath)
} catch {
// File may not exist — that's fine
}
}
/**
* Check if a process with the given PID is alive.
*/
function isProcessAlive(pid: number): boolean {
try {
process.kill(pid, 0)
return true
} catch {
return false
}
}
/**
* Query the daemon status by reading the state file and probing the PID.
*
* Returns:
* - { status: 'running', state } — PID is alive
* - { status: 'stopped' } — no state file
* - { status: 'stale' } — state file exists but PID is dead (auto-cleaned)
*/
export function queryDaemonStatus(name = 'remote-control'): {
status: DaemonStatus
state?: DaemonStateData
} {
const state = readDaemonState(name)
if (!state) {
return { status: 'stopped' }
}
if (isProcessAlive(state.pid)) {
return { status: 'running', state }
}
// Stale — process is dead but state file remains
removeDaemonState(name)
return { status: 'stale' }
}
/**
* Stop a running daemon by sending SIGTERM, waiting, then SIGKILL if needed.
* Cleans up the state file afterward.
*
* @returns true if the daemon was stopped, false if it wasn't running
*/
export async function stopDaemonByPid(
name = 'remote-control',
timeoutMs = 10_000,
): Promise<boolean> {
const state = readDaemonState(name)
if (!state) {
return false
}
const { pid } = state
if (!isProcessAlive(pid)) {
removeDaemonState(name)
return false
}
// Send SIGTERM
try {
process.kill(pid, 'SIGTERM')
} catch {
removeDaemonState(name)
return false
}
// Wait for exit with timeout
const deadline = Date.now() + timeoutMs
const pollInterval = 200
while (Date.now() < deadline) {
if (!isProcessAlive(pid)) {
removeDaemonState(name)
return true
}
await new Promise(resolve => setTimeout(resolve, pollInterval))
}
// Force kill
try {
process.kill(pid, 'SIGKILL')
} catch {
// Already dead
}
// Brief wait for SIGKILL to take effect
await new Promise(resolve => setTimeout(resolve, 500))
removeDaemonState(name)
return true
}

View File

@@ -199,8 +199,13 @@ async function main(): Promise<void> {
return
}
// Fast-path for `claude daemon [subcommand]`: long-running supervisor.
if (feature('DAEMON') && args[0] === 'daemon') {
// Fast-path for `claude daemon [subcommand]`: unified daemon + session management.
// Handles both supervisor (start/stop) and background session (bg/attach/logs/kill)
// subcommands under one namespace.
if (
(feature('DAEMON') || feature('BG_SESSIONS')) &&
args[0] === 'daemon'
) {
profileCheckpoint('cli_daemon_path')
const { enableConfigs } = await import('../utils/config.js')
enableConfigs()
@@ -211,51 +216,65 @@ async function main(): Promise<void> {
return
}
// Fast-path for `claude ps|logs|attach|kill` and `--bg`/`--background`.
// Session management against the ~/.claude/sessions/ registry. Flag
// literals are inlined so bg.js only loads when actually dispatching.
// Fast-path for `--bg`/`--background` shortcut → daemon bg.
if (
feature('BG_SESSIONS') &&
(args.includes('--bg') || args.includes('--background'))
) {
profileCheckpoint('cli_daemon_path')
const { enableConfigs } = await import('../utils/config.js')
enableConfigs()
const bg = await import('../cli/bg.js')
await bg.handleBgStart(
args.filter(a => a !== '--bg' && a !== '--background'),
)
return
}
// Backward-compat: ps/logs/attach/kill → daemon <sub> (deprecated)
if (
feature('BG_SESSIONS') &&
(args[0] === 'ps' ||
args[0] === 'logs' ||
args[0] === 'attach' ||
args[0] === 'kill' ||
args.includes('--bg') ||
args.includes('--background'))
args[0] === 'kill')
) {
profileCheckpoint('cli_bg_path')
const mapped = args[0] === 'ps' ? 'status' : args[0]
console.error(
`[deprecated] Use: claude daemon ${mapped}${args[1] ? ' ' + args[1] : ''}`,
)
profileCheckpoint('cli_daemon_path')
const { enableConfigs } = await import('../utils/config.js')
enableConfigs()
const bg = await import('../cli/bg.js')
switch (args[0]) {
case 'ps':
await bg.psHandler(args.slice(1))
break
case 'logs':
await bg.logsHandler(args[1])
break
case 'attach':
await bg.attachHandler(args[1])
break
case 'kill':
await bg.killHandler(args[1])
break
default:
await bg.handleBgFlag(args)
}
const { initSinks } = await import('../utils/sinks.js')
initSinks()
const { daemonMain } = await import('../daemon/main.js')
await daemonMain([args[0] === 'ps' ? 'status' : args[0]!, ...args.slice(1)])
return
}
// Fast-path for template job commands.
// Fast-path for `claude job <subcommand>`: template jobs.
if (feature('TEMPLATES') && args[0] === 'job') {
profileCheckpoint('cli_templates_path')
const { templatesMain } = await import('../cli/handlers/templateJobs.js')
await templatesMain(args.slice(1))
// process.exit (not return) — mountFleetView's Ink TUI can leave event
// loop handles that prevent natural exit.
// eslint-disable-next-line custom-rules/no-process-exit
process.exit(0)
}
// Backward-compat: new/list/reply → job <sub> (deprecated)
if (
feature('TEMPLATES') &&
(args[0] === 'new' || args[0] === 'list' || args[0] === 'reply')
) {
console.error(
`[deprecated] Use: claude job ${args[0]} ${args.slice(1).join(' ')}`.trim(),
)
profileCheckpoint('cli_templates_path')
const { templatesMain } = await import('../cli/handlers/templateJobs.js')
await templatesMain(args)
// process.exit (not return) — mountFleetView's Ink TUI can leave event
// loop handles that prevent natural exit.
// eslint-disable-next-line custom-rules/no-process-exit
process.exit(0)
}

View File

@@ -48,7 +48,6 @@ export function useAwaySummary(
'tengu_sedge_lantern',
false,
)
useEffect(() => {
if (!feature('AWAY_SUMMARY')) return
if (!gbEnabled) return

View File

@@ -18,6 +18,11 @@ import {
type PipeIpcSlaveState,
} from '../utils/pipeTransport.js'
import { logForDebugging } from '../utils/debug.js'
import {
isMasterPipeMuted,
hasSendOverride,
removeSendOverride,
} from '../utils/pipeMuteState.js'
/** Session history entry for pipe IPC monitoring. */
export type SessionEntry = {
@@ -113,6 +118,28 @@ function isMonitoredPipeEntryType(type: string): boolean {
return MONITORED_PIPE_ENTRY_TYPES.includes(type)
}
/** Business message types that should be dropped when a slave is muted. */
const MUTED_DROPPABLE_TYPES = new Set([
'prompt_ack',
'stream',
'tool_start',
'tool_result',
'done',
'error',
'permission_request',
'permission_cancel',
])
/**
* Centralized mute check used by both attachPipeEntryEmitter and
* useMasterMonitor's inline handler — keeps the two gates in sync.
*/
export function shouldDropMutedMessage(slaveName: string, msgType: string): boolean {
if (hasSendOverride(slaveName)) return false
if (!isMasterPipeMuted(slaveName)) return false
return MUTED_DROPPABLE_TYPES.has(msgType)
}
function pipeMessageToSessionEntry(
slaveName: string,
msg: PipeMessage,
@@ -153,6 +180,35 @@ function attachPipeEntryEmitter(name: string, client: PipeClient): void {
if (typeof client.on !== 'function') return
const handler = (msg: PipeMessage) => {
if (!isMonitoredPipeEntryType(msg.type)) return
// Mute gate: drop business messages from muted slaves
if (shouldDropMutedMessage(name, msg.type)) {
// Auto-deny permission_request to prevent slave deadlock
if (msg.type === 'permission_request') {
try {
const payload = JSON.parse(msg.data ?? '{}')
if (payload.requestId) {
client.send({
type: 'permission_response',
data: JSON.stringify({
requestId: payload.requestId,
behavior: 'deny',
feedback: 'Permission auto-denied: pipe is logically disconnected.',
}),
})
}
} catch {
// Malformed payload — safe to ignore
}
}
return
}
// Clear /send override when slave turn completes
if ((msg.type === 'done' || msg.type === 'error') && hasSendOverride(name)) {
removeSendOverride(name)
}
emitPipeEntry(name, pipeMessageToSessionEntry(name, msg))
}
_pipeEntryHandlers.set(name, handler)
@@ -166,14 +222,14 @@ function emitSlaveClientRegistryChanged(): void {
}
}
function subscribeToSlaveClientRegistry(listener: () => void): () => void {
export function subscribeToSlaveClientRegistry(listener: () => void): () => void {
_slaveClientRegistryListeners.add(listener)
return () => {
_slaveClientRegistryListeners.delete(listener)
}
}
function getSlaveClientRegistryVersion(): number {
export function getSlaveClientRegistryVersion(): number {
return _slaveClientRegistryVersion
}
@@ -248,13 +304,23 @@ export function useMasterMonitor(): void {
for (const [slaveName, client] of _slaveClients.entries()) {
const handler = (msg: PipeMessage) => {
const entry = pipeMessageToSessionEntry(slaveName, msg)
// Only record relevant message types
if (!isMonitoredPipeEntryType(msg.type)) {
return
}
// Mute gate (second gate, same helper as attachPipeEntryEmitter)
if (shouldDropMutedMessage(slaveName, msg.type)) {
return
}
// Clear /send override when slave turn completes
if ((msg.type === 'done' || msg.type === 'error') && hasSendOverride(slaveName)) {
removeSendOverride(slaveName)
}
const entry = pipeMessageToSessionEntry(slaveName, msg)
setAppState(prev => {
const slave = getPipeIpc(prev).slaves[slaveName]
if (!slave) return prev
@@ -294,6 +360,8 @@ export function useMasterMonitor(): void {
// Handle slave disconnect
const onDisconnect = () => {
logForDebugging(`[MasterMonitor] Slave "${slaveName}" disconnected`)
// Clear any lingering /send override before removing client
removeSendOverride(slaveName)
removeSlaveClient(slaveName)
setAppState(prev => {
const { [slaveName]: _removed, ...remainingSlaves } =

View File

@@ -246,6 +246,15 @@ function registerMessageHandlers(
}
})
// Handle relay mute/unmute from master
server.onMessage((msg: PipeMessage, _reply) => {
if (msg.type === 'relay_mute') {
pp().setRelayMuted(true)
} else if (msg.type === 'relay_unmute') {
pp().setRelayMuted(false)
}
})
// Handle detach
server.onMessage((msg: PipeMessage, _reply) => {
if (msg.type !== 'detach') return

View File

@@ -0,0 +1,141 @@
/**
* usePipeMuteSync — Sync master's UI selection state to slave relay mute flags.
*
* Watches routeMode, selectedPipes, slave client registry, and send-override
* changes. When a slave is deselected or routeMode switches to 'local', sends
* relay_mute. When re-selected, sends relay_unmute. Also maintains the
* master-side muted set for in-flight message filtering.
*
* Feature-gated by UDS_INBOX (conditional import in REPL.tsx).
*/
import { useEffect, useRef, useSyncExternalStore } from 'react'
import { useAppState } from '../state/AppState.js'
import { getPipeIpc } from '../utils/pipeTransport.js'
import {
setMasterMutedPipes,
clearMasterMutedPipes,
hasSendOverride,
clearSendOverrides,
subscribeSendOverride,
getSendOverrideVersion,
} from '../utils/pipeMuteState.js'
import {
getAllSlaveClients,
subscribeToSlaveClientRegistry,
getSlaveClientRegistryVersion,
} from './useMasterMonitor.js'
type UsePipeMuteSyncDeps = {
setToolUseConfirmQueue: (action: React.SetStateAction<Record<string, unknown>[]>) => void
}
export function usePipeMuteSync({
setToolUseConfirmQueue,
}: UsePipeMuteSyncDeps): void {
// Subscribe to individual scalars to avoid object-selector re-render churn
// (AppState.tsx warns against object-returning selectors)
const routeMode = useAppState(
s => (getPipeIpc(s).routeMode as 'selected' | 'local') ?? 'selected',
)
const selectedPipes: string[] = useAppState(
s => (getPipeIpc(s).selectedPipes as string[]) ?? [],
)
// Subscribe to slave client registry changes
const registryVersion = useSyncExternalStore(
subscribeToSlaveClientRegistry,
getSlaveClientRegistryVersion,
getSlaveClientRegistryVersion,
)
// Subscribe to send-override changes so mute recalculates after /send completes
const sendOverrideVersion = useSyncExternalStore(
subscribeSendOverride,
getSendOverrideVersion,
getSendOverrideVersion,
)
const prevMutedRef = useRef<Set<string>>(new Set())
useEffect(() => {
const slaves = getAllSlaveClients()
// Compute which slaves should be muted now
const nextMuted = new Set<string>()
if (routeMode === 'local') {
// All connected slaves muted
for (const name of slaves.keys()) {
if (!hasSendOverride(name)) {
nextMuted.add(name)
}
}
} else {
// routeMode === 'selected': mute slaves NOT in selectedPipes
const selectedSet = new Set(selectedPipes)
for (const name of slaves.keys()) {
if (!selectedSet.has(name) && !hasSendOverride(name)) {
nextMuted.add(name)
}
}
}
// Step 1: Update master-side muted set FIRST (before sending control packets)
setMasterMutedPipes(nextMuted)
const prevMuted = prevMutedRef.current
// Step 2: For newly muted slaves — abort pending permissions, then send relay_mute
for (const name of nextMuted) {
if (!prevMuted.has(name)) {
// Abort pending permission prompts for this slave
setToolUseConfirmQueue((queue: Record<string, unknown>[]) => {
const toAbort = queue.filter(
(item: Record<string, unknown>) => item.pipeName === name,
)
for (const item of toAbort) {
try {
;(item.onAbort as (() => void) | undefined)?.()
} catch {
// onAbort may throw if client disconnected — safe to ignore
}
}
return queue.filter((item: Record<string, unknown>) => item.pipeName !== name)
})
// Send relay_mute to slave
const client = slaves.get(name)
if (client?.connected) {
try {
client.send({ type: 'relay_mute' })
} catch {
// send may fail if socket is closing — non-fatal
}
}
}
}
// Step 3: For newly unmuted slaves — send relay_unmute
for (const name of prevMuted) {
if (!nextMuted.has(name)) {
const client = slaves.get(name)
if (client?.connected) {
try {
client.send({ type: 'relay_unmute' })
} catch {
// non-fatal
}
}
}
}
prevMutedRef.current = nextMuted
}, [routeMode, selectedPipes, registryVersion, sendOverrideVersion, setToolUseConfirmQueue])
// Cleanup on unmount: clear all master-side mute state
useEffect(() => {
return () => {
clearMasterMutedPipes()
clearSendOverrides()
}
}, [])
}

View File

@@ -90,6 +90,7 @@ export function usePipePermissionForward({
input: payload.input,
toolUseContext,
toolUseID: `pipe:${payload.requestId}`,
pipeName,
permissionResult: payload.permissionResult,
permissionPromptStartTimeMs:
payload.permissionPromptStartTimeMs,

View File

@@ -6,7 +6,7 @@
* `getPipeRelay()` singleton set by usePipeIpc's attach handler.
*/
import { useRef, useCallback } from 'react'
import { getPipeRelay } from '../utils/pipePermissionRelay.js'
import { getPipeRelay, isRelayMuted } from '../utils/pipePermissionRelay.js'
import type { PipeMessage } from '../utils/pipeTransport.js'
export type PipeRelayHandle = {
@@ -29,6 +29,9 @@ export function usePipeRelay(): PipeRelayHandle {
if (typeof relay !== 'function') {
return false
}
if (isRelayMuted()) {
return false
}
relay(message)
return true
},

View File

@@ -7,9 +7,12 @@ import {
} from '../tasks/InProcessTeammateTask/InProcessTeammateTask.js'
import { isKairosCronEnabled } from '@claude-code-best/builtin-tools/tools/ScheduleCronTool/prompt.js'
import type { Message } from '../types/message.js'
import { getCwd } from '../utils/cwd.js'
import { getCronJitterConfig } from '../utils/cronJitterConfig.js'
import { createCronScheduler } from '../utils/cronScheduler.js'
import { removeCronTasks } from '../utils/cronTasks.js'
import { createAutonomyQueuedPrompt } from '../utils/autonomyRuns.js'
import { markAutonomyRunFailed } from '../utils/autonomyRuns.js'
import { logForDebugging } from '../utils/debug.js'
import { enqueuePendingNotification } from '../utils/messageQueueManager.js'
import { createScheduledTaskFireMessage } from '../utils/messages.js'
@@ -68,50 +71,92 @@ export function useScheduledTasks({
// forward isMeta, so their messages remain visible in the
// transcript. This is acceptable since normal mode is not the
// primary use case for scheduled tasks.
const enqueueForLead = (prompt: string) =>
enqueuePendingNotification({
value: prompt,
mode: 'prompt',
priority: 'later',
isMeta: true,
// Threaded through to cc_workload= in the billing-header
// attribution block so the API can serve cron-initiated requests
// at lower QoS when capacity is tight. No human is actively
// waiting on this response.
const enqueueForLead = async (prompt: string) => {
const command = await createAutonomyQueuedPrompt({
basePrompt: prompt,
trigger: 'scheduled-task',
currentDir: getCwd(),
workload: WORKLOAD_CRON,
})
if (!command) {
return
}
enqueuePendingNotification(command)
}
const scheduler = createCronScheduler({
// Missed-task surfacing (onFire fallback). Teammate crons are always
// session-only (durable:false) so they never appear in the missed list,
// which is populated from disk at scheduler startup — this path only
// handles team-lead durable crons.
onFire: enqueueForLead,
onFire: prompt => {
void enqueueForLead(prompt)
},
// Normal fires receive the full CronTask so we can route by agentId.
onFireTask: task => {
if (task.agentId) {
const teammate = findTeammateTaskByAgentId(
task.agentId,
store.getState().tasks,
)
if (teammate && !isTerminalTaskStatus(teammate.status)) {
injectUserMessageToTeammate(teammate.id, task.prompt, setAppState)
void (async () => {
if (task.agentId) {
const teammate = findTeammateTaskByAgentId(
task.agentId,
store.getState().tasks,
)
if (teammate && !isTerminalTaskStatus(teammate.status)) {
const command = await createAutonomyQueuedPrompt({
basePrompt: task.prompt,
trigger: 'scheduled-task',
currentDir: getCwd(),
sourceId: task.id,
sourceLabel: task.prompt,
workload: WORKLOAD_CRON,
})
if (!command) {
return
}
const injected = injectUserMessageToTeammate(
teammate.id,
command.value as string,
{
autonomyRunId: command.autonomy?.runId,
origin: command.origin,
},
setAppState,
)
if (!injected && command.autonomy?.runId) {
await markAutonomyRunFailed(
command.autonomy.runId,
`Teammate ${task.agentId} exited before the scheduled message could be delivered.`,
)
}
return
}
// Teammate is gone — clean up the orphaned cron so it doesn't keep
// firing into nowhere every tick. One-shots would auto-delete on
// fire anyway, but recurring crons would loop until auto-expiry.
logForDebugging(
`[ScheduledTasks] teammate ${task.agentId} gone, removing orphaned cron ${task.id}`,
)
void removeCronTasks([task.id])
return
}
// Teammate is gone — clean up the orphaned cron so it doesn't keep
// firing into nowhere every tick. One-shots would auto-delete on
// fire anyway, but recurring crons would loop until auto-expiry.
logForDebugging(
`[ScheduledTasks] teammate ${task.agentId} gone, removing orphaned cron ${task.id}`,
const command = await createAutonomyQueuedPrompt({
basePrompt: task.prompt,
trigger: 'scheduled-task',
currentDir: getCwd(),
sourceId: task.id,
sourceLabel: task.prompt,
workload: WORKLOAD_CRON,
})
if (!command) {
return
}
const msg = createScheduledTaskFireMessage(
`Running scheduled task (${formatCronFireTime(new Date())})`,
)
void removeCronTasks([task.id])
return
}
const msg = createScheduledTaskFireMessage(
`Running scheduled task (${formatCronFireTime(new Date())})`,
)
setMessages(prev => [...prev, msg])
enqueueForLead(task.prompt)
setMessages(prev => [...prev, msg])
enqueuePendingNotification(command)
})()
},
isLoading: () => isLoadingRef.current,
assistantMode,

View File

@@ -0,0 +1,140 @@
/**
* Tests for src/jobs/classifier.ts
*
* Uses real temp directories instead of mocking fs to avoid
* cross-test mock pollution in bun test.
*
* classifier.ts takes jobDir as a parameter, so no envUtils mock needed.
*/
import { describe, expect, test, beforeEach, afterAll } from 'bun:test'
import { mkdtempSync, mkdirSync, writeFileSync, readFileSync, rmSync } from 'fs'
import { join } from 'path'
import { tmpdir } from 'os'
import type { AssistantMessage } from '../../types/message.js'
import { classifyAndWriteState } from '../classifier.js'
// ─── setup: real temp dir ──────────────────────────────────────────────────
let tempBase: string
let jobDir: string
let stateFile: string
tempBase = mkdtempSync(join(tmpdir(), 'classifier-test-'))
function freshJobDir(): void {
jobDir = mkdtempSync(join(tempBase, 'job-'))
stateFile = join(jobDir, 'state.json')
}
// ─── helpers ────────────────────────────────────────────────────────────────
function makeAssistantMessage(
content: any[],
extra: Record<string, any> = {},
): AssistantMessage {
return {
type: 'assistant',
uuid: '00000000-0000-0000-0000-000000000000' as any,
message: {
role: 'assistant',
content,
...extra,
},
} as any
}
// ─── lifecycle ─────────────────────────────────────────────────────────────
beforeEach(() => {
freshJobDir()
})
afterAll(() => {
try {
rmSync(tempBase, { recursive: true, force: true })
} catch {
// best-effort cleanup
}
})
// ─── tests ──────────────────────────────────────────────────────────────────
describe('classifyAndWriteState', () => {
test('does nothing when state.json is missing', async () => {
await classifyAndWriteState(jobDir, [])
// stateFile should still not exist
let exists = false
try {
readFileSync(stateFile, 'utf-8')
exists = true
} catch {
// expected
}
expect(exists).toBe(false)
})
test('sets status to running when last message has tool_use block', async () => {
writeFileSync(
stateFile,
JSON.stringify({ status: 'created', updatedAt: '2026-01-01' }),
'utf-8',
)
const msg = makeAssistantMessage([
{ type: 'text', text: 'Let me check...' },
{ type: 'tool_use', id: 'toolu_1', name: 'bash', input: {} },
])
await classifyAndWriteState(jobDir, [msg])
const state = JSON.parse(readFileSync(stateFile, 'utf-8'))
expect(state.status).toBe('running')
})
test('sets status to completed when stop_reason is end_turn', async () => {
writeFileSync(
stateFile,
JSON.stringify({ status: 'running', updatedAt: '2026-01-01' }),
'utf-8',
)
const msg = makeAssistantMessage([{ type: 'text', text: 'All done.' }], {
stop_reason: 'end_turn',
})
await classifyAndWriteState(jobDir, [msg])
const state = JSON.parse(readFileSync(stateFile, 'utf-8'))
expect(state.status).toBe('completed')
})
test('sets status to running for empty messages (state exists)', async () => {
writeFileSync(
stateFile,
JSON.stringify({ status: 'created', updatedAt: '2026-01-01' }),
'utf-8',
)
await classifyAndWriteState(jobDir, [])
const state = JSON.parse(readFileSync(stateFile, 'utf-8'))
expect(state.status).toBe('running')
})
test('sets status to running when stop_reason is max_tokens', async () => {
writeFileSync(
stateFile,
JSON.stringify({ status: 'running', updatedAt: '2026-01-01' }),
'utf-8',
)
const msg = makeAssistantMessage([{ type: 'text', text: 'I need more' }], {
stop_reason: 'max_tokens',
})
await classifyAndWriteState(jobDir, [msg])
const state = JSON.parse(readFileSync(stateFile, 'utf-8'))
expect(state.status).toBe('running')
})
})

View File

@@ -0,0 +1,91 @@
/**
* Tests for src/jobs/state.ts
*
* Uses real temp directories and CLAUDE_CONFIG_DIR env var
* instead of mocking fs, to avoid cross-test mock pollution.
*/
import { describe, expect, test, beforeEach, afterAll } from 'bun:test'
import { mkdtempSync, rmSync, readFileSync, existsSync } from 'fs'
import { join } from 'path'
import { tmpdir } from 'os'
// ─── setup: real temp dir via env var ──────────────────────────────────────
const tempBase = mkdtempSync(join(tmpdir(), 'jobs-state-test-'))
beforeEach(() => {
// Each test gets a fresh config dir
const tempHome = mkdtempSync(join(tempBase, 'home-'))
process.env.CLAUDE_CONFIG_DIR = tempHome
})
afterAll(() => {
delete process.env.CLAUDE_CONFIG_DIR
try {
rmSync(tempBase, { recursive: true, force: true })
} catch {
// best-effort cleanup
}
})
// ─── import ─────────────────────────────────────────────────────────────────
const { createJob, readJobState, appendJobReply, getJobDir } = await import(
'../state.js'
)
// ─── tests ──────────────────────────────────────────────────────────────────
describe('createJob', () => {
test('creates job directory and writes state, template, and input files', () => {
const dir = createJob('job-1', 'my-template', '# Template', 'hello', [
'--flag',
])
expect(dir).toContain('job-1')
expect(existsSync(dir)).toBe(true)
const stateFile = join(dir, 'state.json')
expect(existsSync(stateFile)).toBe(true)
const state = JSON.parse(readFileSync(stateFile, 'utf-8'))
expect(state.jobId).toBe('job-1')
expect(state.templateName).toBe('my-template')
expect(state.status).toBe('created')
expect(state.args).toEqual(['--flag'])
expect(readFileSync(join(dir, 'template.md'), 'utf-8')).toBe('# Template')
expect(readFileSync(join(dir, 'input.txt'), 'utf-8')).toBe('hello')
})
})
describe('readJobState', () => {
test('returns null when job does not exist', () => {
expect(readJobState('nonexistent')).toBeNull()
})
test('returns parsed state when job exists', () => {
createJob('job-2', 'tpl', 'content', 'input', [])
const result = readJobState('job-2')
expect(result).not.toBeNull()
expect(result!.jobId).toBe('job-2')
expect(result!.status).toBe('created')
})
})
describe('appendJobReply', () => {
test('returns false when job does not exist', () => {
expect(appendJobReply('no-job', 'hello')).toBe(false)
})
test('appends reply and updates state', () => {
createJob('job-3', 'tpl', 'content', 'input', [])
const result = appendJobReply('job-3', 'my reply')
expect(result).toBe(true)
const dir = getJobDir('job-3')
const repliesPath = join(dir, 'replies.jsonl')
expect(existsSync(repliesPath)).toBe(true)
const replyLine = JSON.parse(readFileSync(repliesPath, 'utf-8').trim())
expect(replyLine.text).toBe('my reply')
})
})

View File

@@ -0,0 +1,87 @@
/**
* Tests for src/jobs/templates.ts
*
* Uses real temp directories and CLAUDE_CONFIG_DIR env var
* instead of mocking fs, to avoid cross-test mock pollution.
*/
import { describe, expect, test, beforeEach, afterAll } from 'bun:test'
import { mkdtempSync, mkdirSync, writeFileSync, rmSync } from 'fs'
import { join } from 'path'
import { tmpdir } from 'os'
// ─── setup: real temp dir via env var ──────────────────────────────────────
const tempBase = mkdtempSync(join(tmpdir(), 'jobs-templates-test-'))
beforeEach(() => {
const tempHome = mkdtempSync(join(tempBase, 'home-'))
process.env.CLAUDE_CONFIG_DIR = tempHome
})
afterAll(() => {
delete process.env.CLAUDE_CONFIG_DIR
try {
rmSync(tempBase, { recursive: true, force: true })
} catch {
// best-effort cleanup
}
})
// ─── import ─────────────────────────────────────────────────────────────────
const { listTemplates, loadTemplate } = await import('../templates.js')
// ─── tests ──────────────────────────────────────────────────────────────────
describe('listTemplates', () => {
test('returns empty array when no template dirs exist', () => {
const result = listTemplates()
expect(result).toEqual([])
})
test('discovers templates from user-level dir', () => {
const userDir = join(process.env.CLAUDE_CONFIG_DIR!, 'templates')
mkdirSync(userDir, { recursive: true })
writeFileSync(
join(userDir, 'greeting.md'),
'---\ndescription: A greeting template\n---\nHello {{name}}',
'utf-8',
)
const result = listTemplates()
expect(result.length).toBe(1)
expect(result[0]!.name).toBe('greeting')
expect(result[0]!.description).toBe('A greeting template')
expect(result[0]!.content).toBe('Hello {{name}}')
})
test('skips non-md files', () => {
const userDir = join(process.env.CLAUDE_CONFIG_DIR!, 'templates')
mkdirSync(userDir, { recursive: true })
writeFileSync(join(userDir, 'notes.txt'), 'not a template', 'utf-8')
writeFileSync(join(userDir, 'data.json'), '{}', 'utf-8')
const result = listTemplates()
expect(result).toEqual([])
})
})
describe('loadTemplate', () => {
test('returns null when template not found', () => {
expect(loadTemplate('nonexistent')).toBeNull()
})
test('returns template by name', () => {
const userDir = join(process.env.CLAUDE_CONFIG_DIR!, 'templates')
mkdirSync(userDir, { recursive: true })
writeFileSync(
join(userDir, 'deploy.md'),
'---\ndescription: Deploy script\n---\nrun deploy',
'utf-8',
)
const result = loadTemplate('deploy')
expect(result).not.toBeNull()
expect(result!.name).toBe('deploy')
})
})

View File

@@ -1,3 +1,67 @@
// Auto-generated stub — replace with real implementation
export {};
export const classifyAndWriteState: (...args: unknown[]) => Promise<void> = () => Promise.resolve();
import { readFileSync, writeFileSync } from 'fs'
import { join } from 'path'
import type { AssistantMessage } from '../types/message.js'
/**
* Classify the job status from the turn's assistant messages and update state.json.
*
* Called by stopHooks.ts after each repl_main_thread turn when CLAUDE_JOB_DIR is set.
* Only the main thread calls this (not subagents).
*
* @param jobDir - Path to the job directory (from CLAUDE_JOB_DIR env)
* @param assistantMessages - Assistant messages from this turn
*/
export async function classifyAndWriteState(
jobDir: string,
assistantMessages: AssistantMessage[],
): Promise<void> {
const stateFile = join(jobDir, 'state.json')
let state: Record<string, unknown>
try {
state = JSON.parse(readFileSync(stateFile, 'utf-8'))
} catch {
// No state file or corrupt — not a valid job directory
return
}
const newStatus = classifyStatus(assistantMessages)
state.status = newStatus
state.updatedAt = new Date().toISOString()
writeFileSync(stateFile, JSON.stringify(state, null, 2), 'utf-8')
}
/**
* Determine job status from assistant messages.
*
* - Has tool_use blocks → still running (tools executing)
* - stop_reason === 'end_turn' → completed (model finished)
* - Otherwise → running
*/
function classifyStatus(messages: AssistantMessage[]): string {
if (messages.length === 0) return 'running'
const lastMessage = messages[messages.length - 1]!
const content = lastMessage.message?.content
// Check if the last message has tool_use blocks (still executing)
if (Array.isArray(content)) {
const hasToolUse = content.some(
block =>
typeof block === 'object' &&
block !== null &&
'type' in block &&
block.type === 'tool_use',
)
if (hasToolUse) return 'running'
}
// Check stop_reason via index signature
const stopReason = (lastMessage.message as Record<string, unknown>)
?.stop_reason
if (stopReason === 'end_turn') return 'completed'
if (stopReason === 'max_tokens') return 'running'
return 'running'
}

102
src/jobs/state.ts Normal file
View File

@@ -0,0 +1,102 @@
import { appendFileSync, mkdirSync, readFileSync, writeFileSync } from 'fs'
import { join } from 'path'
import { getClaudeConfigHomeDir } from '../utils/envUtils.js'
export interface JobState {
jobId: string
templateName: string
createdAt: string
updatedAt: string
status: 'created' | 'running' | 'completed' | 'failed'
args: string[]
}
function getJobsDir(): string {
return join(getClaudeConfigHomeDir(), 'jobs')
}
export function getJobDir(jobId: string): string {
return join(getJobsDir(), jobId)
}
/**
* Create a new job directory with initial state.
*/
export function createJob(
jobId: string,
templateName: string,
templateContent: string,
inputText: string,
args: string[],
): string {
const dir = getJobDir(jobId)
mkdirSync(dir, { recursive: true })
const now = new Date().toISOString()
const state: JobState = {
jobId,
templateName,
createdAt: now,
updatedAt: now,
status: 'created',
args,
}
writeFileSync(
join(dir, 'state.json'),
JSON.stringify(state, null, 2),
'utf-8',
)
writeFileSync(join(dir, 'template.md'), templateContent, 'utf-8')
writeFileSync(join(dir, 'input.txt'), inputText, 'utf-8')
return dir
}
/**
* Read job state from disk.
*/
export function readJobState(jobId: string): JobState | null {
try {
const raw = readFileSync(join(getJobDir(jobId), 'state.json'), 'utf-8')
const parsed: unknown = JSON.parse(raw)
if (typeof parsed !== 'object' || parsed === null) return null
const obj = parsed as Record<string, unknown>
if (typeof obj.jobId !== 'string' || typeof obj.status !== 'string') {
return null
}
return obj as unknown as JobState
} catch {
return null
}
}
/**
* Append a reply to a job.
*/
export function appendJobReply(jobId: string, text: string): boolean {
const dir = getJobDir(jobId)
const state = readJobState(jobId)
if (!state) return false
const repliesPath = join(dir, 'replies.jsonl')
const entry = JSON.stringify({
text,
timestamp: new Date().toISOString(),
})
try {
appendFileSync(repliesPath, entry + '\n', 'utf-8')
} catch {
writeFileSync(repliesPath, entry + '\n', 'utf-8')
}
const updated = { ...state, updatedAt: new Date().toISOString() }
writeFileSync(
join(dir, 'state.json'),
JSON.stringify(updated, null, 2),
'utf-8',
)
return true
}

86
src/jobs/templates.ts Normal file
View File

@@ -0,0 +1,86 @@
import { readdirSync, readFileSync } from 'fs'
import { join, basename } from 'path'
import { parseFrontmatter } from '../utils/frontmatterParser.js'
import type { FrontmatterData } from '../utils/frontmatterParser.js'
import { getClaudeConfigHomeDir } from '../utils/envUtils.js'
import {
getProjectDirsUpToHome,
extractDescriptionFromMarkdown,
type ClaudeConfigDirectory,
} from '../utils/markdownConfigLoader.js'
export interface TemplateInfo {
name: string
description: string
filePath: string
frontmatter: FrontmatterData
content: string
}
/**
* Discover .claude/templates directories from CWD up to git root,
* plus the user-level ~/.claude/templates.
*/
function getTemplatesDirs(): string[] {
const projectDirs = getProjectDirsUpToHome(
'templates' as ClaudeConfigDirectory,
process.cwd(),
)
// User-level dir (getProjectDirsUpToHome stops before home)
const userDir = join(getClaudeConfigHomeDir(), 'templates')
try {
readdirSync(userDir)
return [...projectDirs, userDir]
} catch {
return projectDirs
}
}
/**
* List all available templates.
*/
export function listTemplates(): TemplateInfo[] {
const templates: TemplateInfo[] = []
const seenNames = new Set<string>()
for (const dir of getTemplatesDirs()) {
let files: string[]
try {
files = readdirSync(dir)
} catch {
continue
}
for (const file of files) {
if (!file.endsWith('.md')) continue
const name = basename(file, '.md')
if (seenNames.has(name)) continue
seenNames.add(name)
const filePath = join(dir, file)
try {
const raw = readFileSync(filePath, 'utf-8')
const { frontmatter, content } = parseFrontmatter(raw, filePath)
const description =
(typeof frontmatter.description === 'string'
? frontmatter.description
: '') || extractDescriptionFromMarkdown(content, 'No description')
templates.push({ name, description, filePath, frontmatter, content })
} catch {
// Skip unreadable files
}
}
}
return templates
}
/**
* Load a specific template by name.
*/
export function loadTemplate(name: string): TemplateInfo | null {
const all = listTemplates()
return all.find(t => t.name === name) ?? null
}

View File

@@ -1802,9 +1802,11 @@ async function run(): Promise<CommanderCommand> {
}
if (
feature("KAIROS") &&
assistantModule?.isAssistantMode() &&
assistantModule &&
(assistantModule.isAssistantForced() ||
(options as Record<string, unknown>).assistant === true) &&
// Spawned teammates share the leader's cwd + settings.json, so
// isAssistantMode() is true for them too. --agent-id being set
// the flag is true for them too. --agent-id being set
// means we ARE a spawned teammate (extractTeammateOptions runs
// ~170 lines later so check the raw commander option) — don't
// re-init the team or override teammateMode/proactive/brief.

View File

@@ -0,0 +1,80 @@
import { beforeEach, describe, expect, test } from 'bun:test'
import {
activateProactive,
deactivateProactive,
getActivationSource,
getNextTickAt,
isContextBlocked,
isProactiveActive,
isProactivePaused,
pauseProactive,
resumeProactive,
setContextBlocked,
setNextTickAt,
shouldTick,
subscribeToProactiveChanges,
} from '../index'
function resetProactiveState() {
activateProactive('reset')
setContextBlocked(false)
setNextTickAt(null)
deactivateProactive()
}
beforeEach(() => {
resetProactiveState()
})
describe('proactive state baseline', () => {
test('activateProactive enables proactive mode and records the source', () => {
activateProactive('baseline_test')
expect(isProactiveActive()).toBe(true)
expect(isProactivePaused()).toBe(false)
expect(isContextBlocked()).toBe(false)
expect(getActivationSource()).toBe('baseline_test')
expect(shouldTick()).toBe(true)
})
test('pauseProactive suppresses ticking and clears nextTickAt', () => {
activateProactive('pause_case')
setNextTickAt(Date.now() + 30_000)
pauseProactive()
expect(isProactivePaused()).toBe(true)
expect(getNextTickAt()).toBeNull()
expect(shouldTick()).toBe(false)
resumeProactive()
expect(isProactivePaused()).toBe(false)
expect(shouldTick()).toBe(true)
})
test('setContextBlocked clears nextTickAt and blocks ticking', () => {
activateProactive('blocked_case')
setNextTickAt(Date.now() + 5_000)
setContextBlocked(true)
expect(isContextBlocked()).toBe(true)
expect(getNextTickAt()).toBeNull()
expect(shouldTick()).toBe(false)
})
test('subscribers are notified on state changes', () => {
let notifications = 0
const unsubscribe = subscribeToProactiveChanges(() => {
notifications += 1
})
activateProactive('subscriber_case')
setNextTickAt(Date.now() + 1_000)
setContextBlocked(true)
deactivateProactive()
unsubscribe()
expect(notifications).toBeGreaterThanOrEqual(3)
})
})

View File

@@ -6,7 +6,10 @@
* proactive mode is active and not blocked.
*/
import { useEffect, useRef } from 'react'
import type { QueuedCommand } from '../types/textInputTypes.js'
import { TICK_TAG } from '../constants/xml.js'
import { getCwd } from '../utils/cwd.js'
import { createProactiveAutonomyCommands } from '../utils/autonomyRuns.js'
import {
isProactiveActive,
isProactivePaused,
@@ -24,8 +27,7 @@ type UseProactiveOpts = {
queuedCommandsLength: number
hasActiveLocalJsxUI: boolean
isInPlanMode: boolean
onSubmitTick: (prompt: string) => void
onQueueTick: (prompt: string) => void
onQueueTick: (command: QueuedCommand) => void
}
export function useProactive(opts: UseProactiveOpts): void {
@@ -70,14 +72,19 @@ export function useProactive(opts: UseProactiveOpts): void {
return
}
const tickContent = `<${TICK_TAG}>${new Date().toLocaleTimeString()}</${TICK_TAG}>`
// If nothing is in the queue, submit directly; otherwise queue
if (queuedCommandsLength === 0) {
optsRef.current.onSubmitTick(tickContent)
} else {
optsRef.current.onQueueTick(tickContent)
}
void (async () => {
const commands = await createProactiveAutonomyCommands({
basePrompt: `<${TICK_TAG}>${new Date().toLocaleTimeString()}</${TICK_TAG}>`,
currentDir: getCwd(),
})
for (const command of commands) {
// Always queue proactive turns. This avoids races where the prompt
// is built asynchronously, a user turn starts meanwhile, and a
// direct-submit path would silently drop the autonomy turn after
// consuming its heartbeat due-state.
optsRef.current.onQueueTick(command)
}
})()
// Schedule next tick
scheduleTick()

View File

@@ -14,18 +14,27 @@ import { dirname, join } from 'path';
import { tmpdir } from 'os';
import figures from 'figures';
// eslint-disable-next-line custom-rules/prefer-use-keybindings -- / n N Esc [ v are bare letters in transcript modal context, same class as g/G/j/k in ScrollKeybindingHandler
import { useInput } from '@anthropic/ink'
import { useSearchInput } from '../hooks/useSearchInput.js'
import { useTerminalSize } from '../hooks/useTerminalSize.js'
import { useSearchHighlight } from '@anthropic/ink'
import type { JumpHandle } from '../components/VirtualMessageList.js'
import { renderMessagesToPlainText } from '../utils/exportRenderer.js'
import { openFileInExternalEditor } from '../utils/editor.js'
import { writeFile } from 'fs/promises'
import { type TabStatusKind, Box, Text, useStdin, useTheme, useTerminalFocus, useTerminalTitle, useTabStatus } from '@anthropic/ink'
import { CostThresholdDialog } from '../components/CostThresholdDialog.js'
import { IdleReturnDialog } from '../components/IdleReturnDialog.js'
import * as React from 'react'
import { useInput } from '@anthropic/ink';
import { useSearchInput } from '../hooks/useSearchInput.js';
import { useTerminalSize } from '../hooks/useTerminalSize.js';
import { useSearchHighlight } from '@anthropic/ink';
import type { JumpHandle } from '../components/VirtualMessageList.js';
import { renderMessagesToPlainText } from '../utils/exportRenderer.js';
import { openFileInExternalEditor } from '../utils/editor.js';
import { writeFile } from 'fs/promises';
import {
type TabStatusKind,
Box,
Text,
useStdin,
useTheme,
useTerminalFocus,
useTerminalTitle,
useTabStatus,
} from '@anthropic/ink';
import { CostThresholdDialog } from '../components/CostThresholdDialog.js';
import { IdleReturnDialog } from '../components/IdleReturnDialog.js';
import * as React from 'react';
import {
useEffect,
useMemo,
@@ -35,14 +44,11 @@ import {
useDeferredValue,
useLayoutEffect,
type RefObject,
} from 'react'
import { useNotifications } from '../context/notifications.js'
import { sendNotification } from '../services/notifier.js'
import {
startPreventSleep,
stopPreventSleep,
} from '../services/preventSleep.js'
import { useTerminalNotification, hasCursorUpViewportYankBug } from '@anthropic/ink'
} from 'react';
import { useNotifications } from '../context/notifications.js';
import { sendNotification } from '../services/notifier.js';
import { startPreventSleep, stopPreventSleep } from '../services/preventSleep.js';
import { useTerminalNotification, hasCursorUpViewportYankBug } from '@anthropic/ink';
import {
createFileStateCacheWithSizeLimit,
mergeFileStateCaches,
@@ -72,6 +78,11 @@ import { QueryGuard } from '../utils/QueryGuard.js';
import { isEnvTruthy } from '../utils/envUtils.js';
import { formatTokens, truncateToWidth } from '../utils/format.js';
import { consumeEarlyInput } from '../utils/earlyInput.js';
import {
finalizeAutonomyRunCompleted,
finalizeAutonomyRunFailed,
markAutonomyRunRunning,
} from '../utils/autonomyRuns.js';
import { setMemberActive } from '../utils/swarm/teamHelpers.js';
import {
@@ -155,9 +166,10 @@ import { CancelRequestHandler } from '../hooks/useCancelRequest.js';
import { useBackgroundTaskNavigation } from '../hooks/useBackgroundTaskNavigation.js';
import { useSwarmInitialization } from '../hooks/useSwarmInitialization.js';
import { useTeammateViewAutoExit } from '../hooks/useTeammateViewAutoExit.js';
import { errorMessage } from '../utils/errors.js';
import { errorMessage, toError } from '../utils/errors.js';
import { isHumanTurn } from '../utils/messagePredicates.js';
import { logError } from '../utils/log.js';
import { getCwd } from '../utils/cwd.js';
// Dead code elimination: conditional imports
/* eslint-disable custom-rules/no-process-env-top-level, @typescript-eslint/no-require-imports */
const useVoiceIntegration: typeof import('../hooks/useVoiceIntegration.js').useVoiceIntegration = feature('VOICE_MODE')
@@ -346,6 +358,7 @@ const usePipeRelay = feature('UDS_INBOX')
const usePipePermissionForward = feature('UDS_INBOX')
? require('../hooks/usePipePermissionForward.js').usePipePermissionForward
: () => undefined;
const usePipeMuteSync = feature('UDS_INBOX') ? require('../hooks/usePipeMuteSync.js').usePipeMuteSync : () => undefined;
const usePipeRouter = feature('UDS_INBOX')
? require('../hooks/usePipeRouter.js').usePipeRouter
: () => ({ routeToSelectedPipes: () => false });
@@ -465,21 +478,13 @@ import { UltraplanChoiceDialog } from '../components/ultraplan/UltraplanChoiceDi
import { UltraplanLaunchDialog } from '../components/ultraplan/UltraplanLaunchDialog.js';
import { launchUltraplan } from '../commands/ultraplan.js';
// Session manager removed - using AppState now
import type { RemoteSessionConfig } from '../remote/RemoteSessionManager.js'
import { REMOTE_SAFE_COMMANDS } from '../commands.js'
import type { RemoteMessageContent } from '../utils/teleport/api.js'
import {
FullscreenLayout,
useUnseenDivider,
computeUnseenDivider,
} from '../components/FullscreenLayout.js'
import {
isFullscreenEnvEnabled,
maybeGetTmuxMouseHint,
isMouseTrackingEnabled,
} from '../utils/fullscreen.js'
import { AlternateScreen } from '@anthropic/ink'
import { ScrollKeybindingHandler } from '../components/ScrollKeybindingHandler.js'
import type { RemoteSessionConfig } from '../remote/RemoteSessionManager.js';
import { REMOTE_SAFE_COMMANDS } from '../commands.js';
import type { RemoteMessageContent } from '../utils/teleport/api.js';
import { FullscreenLayout, useUnseenDivider, computeUnseenDivider } from '../components/FullscreenLayout.js';
import { isFullscreenEnvEnabled, maybeGetTmuxMouseHint, isMouseTrackingEnabled } from '../utils/fullscreen.js';
import { AlternateScreen } from '@anthropic/ink';
import { ScrollKeybindingHandler } from '../components/ScrollKeybindingHandler.js';
import {
useMessageActions,
MessageActionsKeybindings,
@@ -487,13 +492,10 @@ import {
type MessageActionsState,
type MessageActionsNav,
type MessageActionCaps,
} from '../components/messageActions.js'
import { setClipboard } from '@anthropic/ink'
import type { ScrollBoxHandle } from '@anthropic/ink'
import {
createAttachmentMessage,
getQueuedCommandAttachments,
} from '../utils/attachments.js'
} from '../components/messageActions.js';
import { setClipboard } from '@anthropic/ink';
import type { ScrollBoxHandle } from '@anthropic/ink';
import { createAttachmentMessage, getQueuedCommandAttachments } from '../utils/attachments.js';
// Stable empty array for hooks that accept MCPServerConnection[] — avoids
// creating a new [] literal on every render in remote mode, which would
@@ -1952,7 +1954,8 @@ export function REPL({
const content = lastAssistant.message?.content;
const contentArray = Array.isArray(content) ? content : [];
const inProgressToolUses = contentArray.filter(
(b): b is ContentBlock & { type: 'tool_use'; id: string } => b.type === 'tool_use' && inProgressToolUseIDs.has((b as { id: string }).id),
(b): b is ContentBlock & { type: 'tool_use'; id: string } =>
b.type === 'tool_use' && inProgressToolUseIDs.has((b as { id: string }).id),
);
return (
inProgressToolUses.length > 0 &&
@@ -3066,7 +3069,10 @@ export function REPL({
if (feature('PROACTIVE') || feature('KAIROS')) {
proactiveModule?.setContextBlocked(false);
}
} else if (newMessage.type === 'progress' && isEphemeralToolProgress(((newMessage as unknown as { data?: { type?: string } }).data?.type))) {
} else if (
newMessage.type === 'progress' &&
isEphemeralToolProgress((newMessage as unknown as { data?: { type?: string } }).data?.type)
) {
// Replace the previous ephemeral progress tick for the same tool
// call instead of appending. Sleep/Bash emit a tick per second and
// only the last one is rendered; appending blows up the messages
@@ -3198,7 +3204,10 @@ export function REPL({
// title silently fell through to the "Claude Code" default.
if (!titleDisabled && !sessionTitle && !agentTitle && !haikuTitleAttemptedRef.current) {
const firstUserMessage = newMessages.find(m => m.type === 'user' && !m.isMeta);
const text = firstUserMessage?.type === 'user' ? getContentText(firstUserMessage.message!.content as string | ContentBlockParam[]) : null;
const text =
firstUserMessage?.type === 'user'
? getContentText(firstUserMessage.message!.content as string | ContentBlockParam[])
: null;
// Skip synthetic breadcrumbs — slash-command output, prompt-skill
// expansions (/commit → <command-message>), local-command headers
// (/help → <command-name>), and bash-mode (!cmd → <bash-input>).
@@ -3353,10 +3362,16 @@ export function REPL({
}
if (feature('BUDDY') && typeof (globalThis as Record<string, unknown>).fireCompanionObserver === 'function') {
// eslint-disable-next-line @typescript-eslint/no-explicit-any
const _fireCompanionObserver = (globalThis as Record<string, any>).fireCompanionObserver as (msgs: unknown, cb: (r: unknown) => void) => void;
const _fireCompanionObserver = (globalThis as Record<string, unknown>).fireCompanionObserver as (
msgs: unknown,
cb: (r: unknown) => void,
) => void;
void _fireCompanionObserver(messagesRef.current, reaction =>
setAppState(prev => (prev.companionReaction === (reaction as typeof prev.companionReaction) ? prev : { ...prev, companionReaction: reaction as typeof prev.companionReaction })),
setAppState(prev =>
prev.companionReaction === (reaction as typeof prev.companionReaction)
? prev
: { ...prev, companionReaction: reaction as typeof prev.companionReaction },
),
);
}
@@ -3703,10 +3718,7 @@ export function REPL({
}
}
// Atomically: clear initial message, set permission mode and rules, and store plan for verification
const shouldStorePlanForVerification =
initialMsg.message.planContent && process.env.USER_TYPE === 'ant' && isEnvTruthy(undefined);
// Atomically: clear initial message, set permission mode and rules
setAppState(prev => {
// Build and apply permission updates (mode + allowedPrompts rules)
let updatedToolPermissionContext = initialMsg.mode
@@ -3729,13 +3741,6 @@ export function REPL({
...prev,
initialMessage: null,
toolPermissionContext: updatedToolPermissionContext,
...(shouldStorePlanForVerification ? {
pendingPlanVerification: {
plan: initialMsg.message.planContent as string,
verificationStarted: false,
verificationCompleted: false,
},
} : {}),
};
});
@@ -4299,7 +4304,7 @@ export function REPL({
});
}
} else {
injectUserMessageToTeammate(task.id, input, setAppState);
injectUserMessageToTeammate(task.id, input, undefined, setAppState);
}
setInputValue('');
helpers.setCursorOffset(0);
@@ -4804,7 +4809,7 @@ export function REPL({
// Submits incoming prompts from teammate messages or tasks mode as new turns
// Returns true if submission succeeded, false if a query is already running
const handleIncomingPrompt = useCallback(
(content: string, options?: { isMeta?: boolean }): boolean => {
(input: string | QueuedCommand, options?: { isMeta?: boolean }): boolean => {
if (queryGuard.isActive) return false;
// Defer to user-queued commands — user input always takes priority
@@ -4816,16 +4821,53 @@ export function REPL({
return false;
}
const queuedCommand =
typeof input === 'string'
? ({
value: input,
mode: 'prompt',
isMeta: options?.isMeta ? true : undefined,
} satisfies QueuedCommand)
: input;
const newAbortController = createAbortController();
setAbortController(newAbortController);
// Create a user message with the formatted content (includes XML wrapper)
const userMessage = createUserMessage({
content,
isMeta: options?.isMeta ? true : undefined,
content: queuedCommand.value as string,
isMeta: queuedCommand.isMeta ? true : undefined,
origin: queuedCommand.origin,
});
void onQuery([userMessage], newAbortController, true, [], mainLoopModel);
const autonomyRunId = queuedCommand.autonomy?.runId;
if (autonomyRunId) {
void markAutonomyRunRunning(autonomyRunId);
}
void onQuery([userMessage], newAbortController, true, [], mainLoopModel)
.then(() => {
if (autonomyRunId) {
void finalizeAutonomyRunCompleted({
runId: autonomyRunId,
currentDir: getCwd(),
priority: 'later',
}).then(nextCommands => {
for (const command of nextCommands) {
enqueue(command);
}
});
}
})
.catch((error: unknown) => {
if (autonomyRunId) {
void finalizeAutonomyRunFailed({
runId: autonomyRunId,
error: String(error),
});
}
logError(toError(error));
});
return true;
},
[onQuery, mainLoopModel, store],
@@ -4856,6 +4898,7 @@ export function REPL({
const pipeIpcState = useAppState(s => getPipeIpc(s as any));
usePipePermissionForward({ store, tools, setMessages, setToolUseConfirmQueue, getToolUseContext, mainLoopModel });
usePipeMuteSync({ setToolUseConfirmQueue });
// Pipe IPC lifecycle — extracted to usePipeIpc hook
usePipeIpc({ store, handleIncomingPrompt });
@@ -4898,8 +4941,7 @@ export function REPL({
queuedCommandsLength: queuedCommands.length,
hasActiveLocalJsxUI: isShowingLocalJSXCommand,
isInPlanMode: toolPermissionContext.mode === 'plan',
onSubmitTick: (prompt: string) => handleIncomingPrompt(prompt, { isMeta: true }),
onQueueTick: (prompt: string) => enqueue({ mode: 'prompt', value: prompt, isMeta: true }),
onQueueTick: (command: QueuedCommand) => enqueue(command),
});
// Abort the current operation when a 'now' priority message arrives
@@ -4952,16 +4994,11 @@ export function REPL({
if (!isLoading) return null;
// Find stop hook progress messages
const progressMsgs = messages.filter(
(m): m is ProgressMessage<HookProgress> => {
if (m.type !== 'progress') return false;
const data = m.data as Record<string, unknown>;
return (
data.type === 'hook_progress' &&
(data.hookEvent === 'Stop' || data.hookEvent === 'SubagentStop')
);
},
);
const progressMsgs = messages.filter((m): m is ProgressMessage<HookProgress> => {
if (m.type !== 'progress') return false;
const data = m.data as Record<string, unknown>;
return data.type === 'hook_progress' && (data.hookEvent === 'Stop' || data.hookEvent === 'SubagentStop');
});
if (progressMsgs.length === 0) return null;
// Get the most recent stop hook execution

View File

@@ -466,6 +466,7 @@ const LOCAL_GATE_DEFAULTS: Record<string, unknown> = {
tengu_birch_trellis: true, // Tree-sitter bash security analysis
tengu_collage_kaleidoscope: true, // macOS clipboard image reading
tengu_compact_cache_prefix: true, // Reuse prompt cache during compaction
tengu_kairos_assistant: true, // KAIROS assistant mode activation
tengu_kairos_cron_durable: true, // Persistent cron tasks
tengu_attribution_header: true, // API request attribution header
tengu_slate_prism: true, // Agent progress summaries
@@ -830,6 +831,16 @@ export function getFeatureValue_CACHED_MAY_BE_STALE<T>(
return localDefault !== undefined ? (localDefault as T) : defaultValue
}
// LOCAL_GATE_DEFAULTS take priority over remote values and disk cache.
// In fork/self-hosted deployments, the GrowthBook server may push false
// for gates we intentionally enable. Local defaults represent the
// project's intentional configuration and override everything except
// env/config overrides (which are explicit user intent).
const localDefault = getLocalGateDefault(feature)
if (localDefault !== undefined) {
return localDefault as T
}
// Log experiment exposure if data is available, otherwise defer until after init
if (experimentDataByFeature.has(feature)) {
logExposureForFeature(feature)
@@ -838,10 +849,6 @@ export function getFeatureValue_CACHED_MAY_BE_STALE<T>(
}
// In-memory payload is authoritative once processRemoteEvalPayload has run.
// Disk is also fresh by then (syncRemoteEvalToDisk runs synchronously inside
// init), so this is correctness-equivalent to the disk read below — but it
// skips the config JSON parse and is what onGrowthBookRefresh subscribers
// depend on to read fresh values the instant they're notified.
if (remoteEvalFeatureValues.has(feature)) {
return remoteEvalFeatureValues.get(feature) as T
}
@@ -853,14 +860,9 @@ export function getFeatureValue_CACHED_MAY_BE_STALE<T>(
return cached as T
}
} catch {
// Config not yet initialized — fall through to local gate defaults
// Config not yet initialized — fall through to defaultValue
}
// Disk cache miss (or config not initialized) — use local gate defaults
// before falling back to the caller's defaultValue. This covers:
// 1. GrowthBook "enabled" but never connected (caches empty)
// 2. Config not yet initialized (early in startup)
const localDefault = getLocalGateDefault(feature)
return localDefault !== undefined ? (localDefault as T) : defaultValue
return defaultValue
}
/**

View File

@@ -0,0 +1,487 @@
/**
* Tests for queryModelOpenAI in index.ts.
*
* Focused on the two bugs fixed:
* 1. stop_reason was always null in the assembled AssistantMessage because
* partialMessage (from message_start) has stop_reason: null, and the
* stop_reason captured from message_delta was never applied.
* 2. partialMessage was not reset to null after message_stop, so the safety
* fallback at the end of the loop would yield a second identical
* AssistantMessage (causing doubled content in the next API request).
*
* Strategy: mock getOpenAIClient + adaptOpenAIStreamToAnthropic so we can
* feed pre-built Anthropic events directly into queryModelOpenAI and inspect
* what it emits — without any real HTTP calls.
*/
import { describe, expect, test, mock, beforeEach, afterEach } from 'bun:test'
import type { BetaRawMessageStreamEvent } from '@anthropic-ai/sdk/resources/beta/messages/messages.mjs'
import type { AssistantMessage, StreamEvent } from '../../../../types/message.js'
// ─── helpers ─────────────────────────────────────────────────────────────────
/** Build a minimal message_start event */
function makeMessageStart(overrides: Record<string, any> = {}): BetaRawMessageStreamEvent {
return {
type: 'message_start',
message: {
id: 'msg_test',
type: 'message',
role: 'assistant',
content: [],
model: 'test-model',
stop_reason: null,
stop_sequence: null,
usage: { input_tokens: 0, output_tokens: 0, cache_creation_input_tokens: 0, cache_read_input_tokens: 0 },
...overrides,
},
} as any
}
/** Build a content_block_start event for the given block type */
function makeContentBlockStart(index: number, type: 'text' | 'tool_use' | 'thinking', extra: Record<string, any> = {}): BetaRawMessageStreamEvent {
const block =
type === 'text'
? { type: 'text', text: '' }
: type === 'tool_use'
? { type: 'tool_use', id: 'toolu_test', name: 'bash', input: {} }
: { type: 'thinking', thinking: '', signature: '' }
return { type: 'content_block_start', index, content_block: { ...block, ...extra } } as any
}
/** Build a text_delta content_block_delta event */
function makeTextDelta(index: number, text: string): BetaRawMessageStreamEvent {
return { type: 'content_block_delta', index, delta: { type: 'text_delta', text } } as any
}
/** Build an input_json_delta content_block_delta event */
function makeInputJsonDelta(index: number, json: string): BetaRawMessageStreamEvent {
return { type: 'content_block_delta', index, delta: { type: 'input_json_delta', partial_json: json } } as any
}
/** Build a thinking_delta content_block_delta event */
function makeThinkingDelta(index: number, thinking: string): BetaRawMessageStreamEvent {
return { type: 'content_block_delta', index, delta: { type: 'thinking_delta', thinking } } as any
}
/** Build a content_block_stop event */
function makeContentBlockStop(index: number): BetaRawMessageStreamEvent {
return { type: 'content_block_stop', index } as any
}
/** Build a message_delta event with stop_reason and output_tokens */
function makeMessageDelta(stopReason: string, outputTokens: number): BetaRawMessageStreamEvent {
return {
type: 'message_delta',
delta: { stop_reason: stopReason, stop_sequence: null },
usage: { output_tokens: outputTokens },
} as any
}
/** Build a message_stop event */
function makeMessageStop(): BetaRawMessageStreamEvent {
return { type: 'message_stop' } as any
}
/** Async generator from a fixed array of events */
async function* eventStream(events: BetaRawMessageStreamEvent[]) {
for (const e of events) yield e
}
/** Collect all outputs from queryModelOpenAI into typed buckets */
async function runQueryModel(
events: BetaRawMessageStreamEvent[],
envOverrides: Record<string, string | undefined> = {},
) {
// Wire events into the mocked stream adapter
_nextEvents = events
// Save + apply env overrides
const saved: Record<string, string | undefined> = {}
for (const [k, v] of Object.entries(envOverrides)) {
saved[k] = process.env[k]
if (v === undefined) delete process.env[k]
else process.env[k] = v
}
try {
// We inline mock.module inside the try block.
// Bun resolves mock.module at the call site synchronously (hoisted),
// so we register once per test file, then re-import each time.
const { queryModelOpenAI } = await import('../index.js')
const assistantMessages: AssistantMessage[] = []
const streamEvents: StreamEvent[] = []
const otherOutputs: any[] = []
const minimalOptions: any = {
model: 'test-model',
tools: [],
agents: [],
querySource: 'main_loop',
getToolPermissionContext: async () => ({
alwaysAllow: [],
alwaysDeny: [],
needsPermission: [],
mode: 'default',
isBypassingPermissions: false,
}),
}
for await (const item of queryModelOpenAI(
[],
{ type: 'text', text: '' } as any,
[],
new AbortController().signal,
minimalOptions,
)) {
if (item.type === 'assistant') {
assistantMessages.push(item as AssistantMessage)
} else if (item.type === 'stream_event') {
streamEvents.push(item as StreamEvent)
} else {
otherOutputs.push(item)
}
}
return { assistantMessages, streamEvents, otherOutputs }
} finally {
// Restore env
for (const [k, v] of Object.entries(saved)) {
if (v === undefined) delete process.env[k]
else process.env[k] = v
}
}
}
// ─── mock setup ──────────────────────────────────────────────────────────────
// We mock at module level. Bun's mock.module replaces the module for the
// entire file, so we configure the stream per-test via a shared variable.
let _nextEvents: BetaRawMessageStreamEvent[] = []
/** Captured arguments from the last chat.completions.create() call */
let _lastCreateArgs: Record<string, any> | null = null
mock.module('../client.js', () => ({
getOpenAIClient: () => ({
chat: {
completions: {
create: async (args: Record<string, any>) => {
_lastCreateArgs = args
return { [Symbol.asyncIterator]: async function* () {} }
},
},
},
}),
}))
mock.module('../streamAdapter.js', () => ({
adaptOpenAIStreamToAnthropic: (_stream: any, _model: string) => eventStream(_nextEvents),
}))
mock.module('../modelMapping.js', () => ({
resolveOpenAIModel: (m: string) => m,
}))
mock.module('../convertMessages.js', () => ({
anthropicMessagesToOpenAI: () => [],
}))
mock.module('../convertTools.js', () => ({
anthropicToolsToOpenAI: () => [],
anthropicToolChoiceToOpenAI: () => undefined,
}))
mock.module('../../../../utils/context.js', () => ({
MODEL_CONTEXT_WINDOW_DEFAULT: 200_000,
COMPACT_MAX_OUTPUT_TOKENS: 20_000,
CAPPED_DEFAULT_MAX_TOKENS: 8_000,
ESCALATED_MAX_TOKENS: 64_000,
is1mContextDisabled: () => false,
has1mContext: () => false,
modelSupports1M: () => false,
getModelMaxOutputTokens: () => ({ upperLimit: 8192, default: 8192 }),
getContextWindowForModel: () => 200_000,
getSonnet1mExpTreatmentEnabled: () => false,
calculateContextPercentages: () => ({ usedPercent: 0, remainingPercent: 100 }),
getMaxThinkingTokensForModel: () => 0,
}))
mock.module('../../../../utils/messages.js', () => ({
normalizeMessagesForAPI: (msgs: any) => msgs,
normalizeContentFromAPI: (blocks: any[]) => blocks,
createAssistantAPIErrorMessage: (opts: any) => ({
type: 'assistant',
message: { content: [{ type: 'text', text: opts.content }], apiError: opts.apiError },
uuid: 'error-uuid',
timestamp: new Date().toISOString(),
}),
}))
mock.module('../../../../utils/api.js', () => ({
toolToAPISchema: async (t: any) => t,
}))
mock.module('../../../../utils/toolSearch.js', () => ({
isToolSearchEnabled: async () => false,
extractDiscoveredToolNames: () => new Set(),
}))
mock.module('../../../../tools/ToolSearchTool/prompt.js', () => ({
isDeferredTool: () => false,
TOOL_SEARCH_TOOL_NAME: '__tool_search__',
}))
mock.module('../../../../cost-tracker.js', () => ({
addToTotalSessionCost: () => {},
}))
mock.module('../../../../utils/modelCost.js', () => ({
COST_TIER_3_15: {},
COST_TIER_15_75: {},
COST_TIER_5_25: {},
COST_TIER_30_150: {},
COST_HAIKU_35: {},
COST_HAIKU_45: {},
getOpus46CostTier: () => ({}),
MODEL_COSTS: {},
getModelCosts: () => ({}),
calculateUSDCost: () => 0,
calculateCostFromTokens: () => 0,
formatModelPricing: () => '',
getModelPricingString: () => undefined,
}))
mock.module('../../../../utils/debug.js', () => ({
logForDebugging: () => {},
logAntError: () => {},
isDebugMode: () => false,
isDebugToStdErr: () => false,
getDebugFilePath: () => null,
getDebugLogPath: () => '',
getDebugFilter: () => null,
getMinDebugLogLevel: () => 'debug',
enableDebugLogging: () => false,
setHasFormattedOutput: () => {},
getHasFormattedOutput: () => false,
flushDebugLogs: async () => {},
}))
// ─── tests ───────────────────────────────────────────────────────────────────
describe('queryModelOpenAI — stop_reason propagation', () => {
test('assembled AssistantMessage has stop_reason end_turn (not null)', async () => {
_nextEvents = [
makeMessageStart(),
makeContentBlockStart(0, 'text'),
makeTextDelta(0, 'Hello'),
makeContentBlockStop(0),
makeMessageDelta('end_turn', 10),
makeMessageStop(),
]
const { assistantMessages } = await runQueryModel(_nextEvents)
expect(assistantMessages).toHaveLength(1)
expect(assistantMessages[0]!.message.stop_reason).toBe('end_turn')
})
test('assembled AssistantMessage has stop_reason tool_use', async () => {
_nextEvents = [
makeMessageStart(),
makeContentBlockStart(0, 'tool_use'),
makeInputJsonDelta(0, '{"cmd":"ls"}'),
makeContentBlockStop(0),
makeMessageDelta('tool_use', 20),
makeMessageStop(),
]
const { assistantMessages } = await runQueryModel(_nextEvents)
expect(assistantMessages).toHaveLength(1)
expect(assistantMessages[0]!.message.stop_reason).toBe('tool_use')
})
test('assembled AssistantMessage has stop_reason max_tokens', async () => {
_nextEvents = [
makeMessageStart(),
makeContentBlockStart(0, 'text'),
makeTextDelta(0, 'truncated'),
makeContentBlockStop(0),
makeMessageDelta('max_tokens', 8192),
makeMessageStop(),
]
const { assistantMessages } = await runQueryModel(_nextEvents)
// Two assistant-typed items: the content message + the max_output_tokens error signal.
// The error signal is emitted as a synthetic assistant message by createAssistantAPIErrorMessage.
expect(assistantMessages).toHaveLength(2)
const contentMsg = assistantMessages[0]!
expect(contentMsg.message.stop_reason).toBe('max_tokens')
// Second item is the error signal (has apiError set)
const errorMsg = assistantMessages[1]!.message as any
expect(errorMsg.apiError).toBe('max_output_tokens')
})
test('stop_reason is null when no message_delta was received (safety fallback path)', async () => {
// Stream ends without message_stop — triggers the safety fallback branch.
// stop_reason stays null since no message_delta was ever seen.
_nextEvents = [
makeMessageStart(),
makeContentBlockStart(0, 'text'),
makeTextDelta(0, 'partial'),
makeContentBlockStop(0),
// No message_delta / message_stop
]
const { assistantMessages } = await runQueryModel(_nextEvents)
// Safety fallback should yield the partial content
expect(assistantMessages).toHaveLength(1)
expect(assistantMessages[0]!.message.stop_reason).toBeNull()
})
})
describe('queryModelOpenAI — usage accumulation', () => {
test('usage in assembled message reflects all four fields from message_delta', async () => {
// message_start has all fields=0 (trailing-chunk pattern: usage not yet available).
// message_delta carries the real values after stream ends.
// The spread in the message_delta handler must override all zeros from message_start,
// including cache_read_input_tokens which was previously missing from message_delta.
_nextEvents = [
makeMessageStart({ usage: { input_tokens: 0, output_tokens: 0, cache_creation_input_tokens: 0, cache_read_input_tokens: 0 } }),
makeContentBlockStart(0, 'text'),
makeTextDelta(0, 'response'),
makeContentBlockStop(0),
// message_delta carries all four Anthropic usage fields (as emitted by the fixed streamAdapter)
{
type: 'message_delta',
delta: { stop_reason: 'end_turn', stop_sequence: null },
usage: { input_tokens: 30011, output_tokens: 190, cache_read_input_tokens: 19904, cache_creation_input_tokens: 0 },
} as any,
makeMessageStop(),
]
const { assistantMessages } = await runQueryModel(_nextEvents)
expect(assistantMessages).toHaveLength(1)
const usage = assistantMessages[0]!.message.usage as any
expect(usage.input_tokens).toBe(30011)
expect(usage.output_tokens).toBe(190)
// cache_read_input_tokens from message_delta overrides the 0 from message_start
expect(usage.cache_read_input_tokens).toBe(19904)
expect(usage.cache_creation_input_tokens).toBe(0)
})
test('usage is zero when no usage events arrive (prevents false autocompact)', async () => {
// If usage stays 0, tokenCountWithEstimation will undercount — so at least
// verify the field exists and is numeric (to detect regressions).
_nextEvents = [
makeMessageStart(),
makeContentBlockStart(0, 'text'),
makeTextDelta(0, 'hi'),
makeContentBlockStop(0),
makeMessageDelta('end_turn', 0),
makeMessageStop(),
]
const { assistantMessages } = await runQueryModel(_nextEvents)
const usage = assistantMessages[0]!.message.usage as any
expect(typeof usage.input_tokens).toBe('number')
expect(typeof usage.output_tokens).toBe('number')
})
})
describe('queryModelOpenAI — no duplicate AssistantMessage (partialMessage reset)', () => {
test('yields exactly one AssistantMessage per message_stop when content is present', async () => {
_nextEvents = [
makeMessageStart(),
makeContentBlockStart(0, 'text'),
makeTextDelta(0, 'only once'),
makeContentBlockStop(0),
makeMessageDelta('end_turn', 5),
makeMessageStop(),
]
const { assistantMessages } = await runQueryModel(_nextEvents)
// Before the fix, partialMessage was not reset to null, so the safety
// fallback at the end of the loop would yield a second message with the
// same message.id — causing mergeAssistantMessages to concatenate content.
expect(assistantMessages).toHaveLength(1)
})
test('thinking + text response yields exactly one AssistantMessage', async () => {
_nextEvents = [
makeMessageStart(),
makeContentBlockStart(0, 'thinking'),
makeThinkingDelta(0, 'let me think'),
makeContentBlockStop(0),
makeContentBlockStart(1, 'text'),
makeTextDelta(1, 'answer'),
makeContentBlockStop(1),
makeMessageDelta('end_turn', 30),
makeMessageStop(),
]
const { assistantMessages } = await runQueryModel(_nextEvents)
expect(assistantMessages).toHaveLength(1)
})
test('safety fallback path still yields message when stream ends without message_stop', async () => {
// Simulates a stream that cuts off without the normal termination sequence.
_nextEvents = [
makeMessageStart(),
makeContentBlockStart(0, 'text'),
makeTextDelta(0, 'abrupt end'),
// No content_block_stop, no message_delta, no message_stop
]
const { assistantMessages } = await runQueryModel(_nextEvents)
expect(assistantMessages).toHaveLength(1)
})
})
describe('queryModelOpenAI — stream_events forwarded', () => {
test('every adapted event is also yielded as stream_event for real-time display', async () => {
_nextEvents = [
makeMessageStart(),
makeContentBlockStart(0, 'text'),
makeTextDelta(0, 'hello'),
makeContentBlockStop(0),
makeMessageDelta('end_turn', 5),
makeMessageStop(),
]
const { streamEvents } = await runQueryModel(_nextEvents)
const eventTypes = streamEvents.map(e => (e as any).event?.type)
expect(eventTypes).toContain('message_start')
expect(eventTypes).toContain('content_block_start')
expect(eventTypes).toContain('content_block_delta')
expect(eventTypes).toContain('content_block_stop')
expect(eventTypes).toContain('message_delta')
expect(eventTypes).toContain('message_stop')
})
})
describe('queryModelOpenAI — max_tokens forwarded to request', () => {
test('buildOpenAIRequestBody includes max_tokens in the request payload', async () => {
_nextEvents = [
makeMessageStart(),
makeContentBlockStart(0, 'text'),
makeTextDelta(0, 'hi'),
makeContentBlockStop(0),
makeMessageDelta('end_turn', 5),
makeMessageStop(),
]
await runQueryModel(_nextEvents)
expect(_lastCreateArgs).not.toBeNull()
expect(_lastCreateArgs!.max_tokens).toBe(8192)
})
})

View File

@@ -1,454 +1,40 @@
/**
* Tests for queryModelOpenAI in index.ts.
*
* Focused on the two bugs fixed:
* 1. stop_reason was always null in the assembled AssistantMessage because
* partialMessage (from message_start) has stop_reason: null, and the
* stop_reason captured from message_delta was never applied.
* 2. partialMessage was not reset to null after message_stop, so the safety
* fallback at the end of the loop would yield a second identical
* AssistantMessage (causing doubled content in the next API request).
*
* Strategy: mock getOpenAIClient + adaptOpenAIStreamToAnthropic so we can
* feed pre-built Anthropic events directly into queryModelOpenAI and inspect
* what it emits — without any real HTTP calls.
*/
import { describe, expect, test, mock, beforeEach, afterEach } from 'bun:test'
import type { BetaRawMessageStreamEvent } from '@anthropic-ai/sdk/resources/beta/messages/messages.mjs'
import type { AssistantMessage, StreamEvent } from '../../../../types/message.js'
import { describe, expect, test } from 'bun:test'
import { fileURLToPath } from 'node:url'
// ─── helpers ─────────────────────────────────────────────────────────────────
const isolatedPath = fileURLToPath(
new URL('./queryModelOpenAI.isolated.ts', import.meta.url),
)
/** Build a minimal message_start event */
function makeMessageStart(overrides: Record<string, any> = {}): BetaRawMessageStreamEvent {
return {
type: 'message_start',
message: {
id: 'msg_test',
type: 'message',
role: 'assistant',
content: [],
model: 'test-model',
stop_reason: null,
stop_sequence: null,
usage: { input_tokens: 0, output_tokens: 0, cache_creation_input_tokens: 0, cache_read_input_tokens: 0 },
...overrides,
},
} as any
}
describe('queryModelOpenAI', () => {
test('passes in isolated subprocess', async () => {
const proc = Bun.spawn({
cmd: [process.execPath, 'test', isolatedPath],
cwd: process.cwd(),
stdout: 'pipe',
stderr: 'pipe',
env: process.env,
})
/** Build a content_block_start event for the given block type */
function makeContentBlockStart(index: number, type: 'text' | 'tool_use' | 'thinking', extra: Record<string, any> = {}): BetaRawMessageStreamEvent {
const block =
type === 'text'
? { type: 'text', text: '' }
: type === 'tool_use'
? { type: 'tool_use', id: 'toolu_test', name: 'bash', input: {} }
: { type: 'thinking', thinking: '', signature: '' }
return { type: 'content_block_start', index, content_block: { ...block, ...extra } } as any
}
const [stdout, stderr, exitCode] = await Promise.all([
new Response(proc.stdout).text(),
new Response(proc.stderr).text(),
proc.exited,
])
/** Build a text_delta content_block_delta event */
function makeTextDelta(index: number, text: string): BetaRawMessageStreamEvent {
return { type: 'content_block_delta', index, delta: { type: 'text_delta', text } } as any
}
/** Build an input_json_delta content_block_delta event */
function makeInputJsonDelta(index: number, json: string): BetaRawMessageStreamEvent {
return { type: 'content_block_delta', index, delta: { type: 'input_json_delta', partial_json: json } } as any
}
/** Build a thinking_delta content_block_delta event */
function makeThinkingDelta(index: number, thinking: string): BetaRawMessageStreamEvent {
return { type: 'content_block_delta', index, delta: { type: 'thinking_delta', thinking } } as any
}
/** Build a content_block_stop event */
function makeContentBlockStop(index: number): BetaRawMessageStreamEvent {
return { type: 'content_block_stop', index } as any
}
/** Build a message_delta event with stop_reason and output_tokens */
function makeMessageDelta(stopReason: string, outputTokens: number): BetaRawMessageStreamEvent {
return {
type: 'message_delta',
delta: { stop_reason: stopReason, stop_sequence: null },
usage: { output_tokens: outputTokens },
} as any
}
/** Build a message_stop event */
function makeMessageStop(): BetaRawMessageStreamEvent {
return { type: 'message_stop' } as any
}
/** Async generator from a fixed array of events */
async function* eventStream(events: BetaRawMessageStreamEvent[]) {
for (const e of events) yield e
}
/** Collect all outputs from queryModelOpenAI into typed buckets */
async function runQueryModel(
events: BetaRawMessageStreamEvent[],
envOverrides: Record<string, string | undefined> = {},
) {
// Wire events into the mocked stream adapter
_nextEvents = events
// Save + apply env overrides
const saved: Record<string, string | undefined> = {}
for (const [k, v] of Object.entries(envOverrides)) {
saved[k] = process.env[k]
if (v === undefined) delete process.env[k]
else process.env[k] = v
}
try {
// We inline mock.module inside the try block.
// Bun resolves mock.module at the call site synchronously (hoisted),
// so we register once per test file, then re-import each time.
const { queryModelOpenAI } = await import('../index.js')
const assistantMessages: AssistantMessage[] = []
const streamEvents: StreamEvent[] = []
const otherOutputs: any[] = []
const minimalOptions: any = {
model: 'test-model',
tools: [],
agents: [],
querySource: 'main_loop',
getToolPermissionContext: async () => ({
alwaysAllow: [],
alwaysDeny: [],
needsPermission: [],
mode: 'default',
isBypassingPermissions: false,
}),
if (exitCode !== 0) {
throw new Error(
[
`isolated queryModelOpenAI test failed with exit code ${exitCode}`,
'',
'STDOUT:',
stdout,
'',
'STDERR:',
stderr,
].join('\n'),
)
}
for await (const item of queryModelOpenAI(
[],
{ type: 'text', text: '' } as any,
[],
new AbortController().signal,
minimalOptions,
)) {
if (item.type === 'assistant') {
assistantMessages.push(item as AssistantMessage)
} else if (item.type === 'stream_event') {
streamEvents.push(item as StreamEvent)
} else {
otherOutputs.push(item)
}
}
return { assistantMessages, streamEvents, otherOutputs }
} finally {
// Restore env
for (const [k, v] of Object.entries(saved)) {
if (v === undefined) delete process.env[k]
else process.env[k] = v
}
}
}
// ─── mock setup ──────────────────────────────────────────────────────────────
// We mock at module level. Bun's mock.module replaces the module for the
// entire file, so we configure the stream per-test via a shared variable.
let _nextEvents: BetaRawMessageStreamEvent[] = []
/** Captured arguments from the last chat.completions.create() call */
let _lastCreateArgs: Record<string, any> | null = null
mock.module('../client.js', () => ({
getOpenAIClient: () => ({
chat: {
completions: {
create: async (args: Record<string, any>) => {
_lastCreateArgs = args
return { [Symbol.asyncIterator]: async function* () {} }
},
},
},
}),
}))
mock.module('../streamAdapter.js', () => ({
adaptOpenAIStreamToAnthropic: (_stream: any, _model: string) => eventStream(_nextEvents),
}))
mock.module('../modelMapping.js', () => ({
resolveOpenAIModel: (m: string) => m,
}))
mock.module('../convertMessages.js', () => ({
anthropicMessagesToOpenAI: () => [],
}))
mock.module('../convertTools.js', () => ({
anthropicToolsToOpenAI: () => [],
anthropicToolChoiceToOpenAI: () => undefined,
}))
mock.module('../../../../utils/context.js', () => ({
getModelMaxOutputTokens: () => ({ upperLimit: 8192, default: 8192 }),
getContextWindowForModel: () => 200_000,
}))
mock.module('../../../../utils/messages.js', () => ({
normalizeMessagesForAPI: (msgs: any) => msgs,
normalizeContentFromAPI: (blocks: any[]) => blocks,
createAssistantAPIErrorMessage: (opts: any) => ({
type: 'assistant',
message: { content: [{ type: 'text', text: opts.content }], apiError: opts.apiError },
uuid: 'error-uuid',
timestamp: new Date().toISOString(),
}),
}))
mock.module('../../../../utils/api.js', () => ({
toolToAPISchema: async (t: any) => t,
}))
mock.module('../../../../utils/toolSearch.js', () => ({
isToolSearchEnabled: async () => false,
extractDiscoveredToolNames: () => new Set(),
}))
mock.module('../../../../tools/ToolSearchTool/prompt.js', () => ({
isDeferredTool: () => false,
TOOL_SEARCH_TOOL_NAME: '__tool_search__',
}))
mock.module('../../../../cost-tracker.js', () => ({
addToTotalSessionCost: () => {},
}))
mock.module('../../../../utils/modelCost.js', () => ({
calculateUSDCost: () => 0,
}))
mock.module('../../../../utils/debug.js', () => ({
logForDebugging: () => {},
}))
// ─── tests ───────────────────────────────────────────────────────────────────
describe('queryModelOpenAI — stop_reason propagation', () => {
test('assembled AssistantMessage has stop_reason end_turn (not null)', async () => {
_nextEvents = [
makeMessageStart(),
makeContentBlockStart(0, 'text'),
makeTextDelta(0, 'Hello'),
makeContentBlockStop(0),
makeMessageDelta('end_turn', 10),
makeMessageStop(),
]
const { assistantMessages } = await runQueryModel(_nextEvents)
expect(assistantMessages).toHaveLength(1)
expect(assistantMessages[0]!.message.stop_reason).toBe('end_turn')
})
test('assembled AssistantMessage has stop_reason tool_use', async () => {
_nextEvents = [
makeMessageStart(),
makeContentBlockStart(0, 'tool_use'),
makeInputJsonDelta(0, '{"cmd":"ls"}'),
makeContentBlockStop(0),
makeMessageDelta('tool_use', 20),
makeMessageStop(),
]
const { assistantMessages } = await runQueryModel(_nextEvents)
expect(assistantMessages).toHaveLength(1)
expect(assistantMessages[0]!.message.stop_reason).toBe('tool_use')
})
test('assembled AssistantMessage has stop_reason max_tokens', async () => {
_nextEvents = [
makeMessageStart(),
makeContentBlockStart(0, 'text'),
makeTextDelta(0, 'truncated'),
makeContentBlockStop(0),
makeMessageDelta('max_tokens', 8192),
makeMessageStop(),
]
const { assistantMessages } = await runQueryModel(_nextEvents)
// Two assistant-typed items: the content message + the max_output_tokens error signal.
// The error signal is emitted as a synthetic assistant message by createAssistantAPIErrorMessage.
expect(assistantMessages).toHaveLength(2)
const contentMsg = assistantMessages[0]!
expect(contentMsg.message.stop_reason).toBe('max_tokens')
// Second item is the error signal (has apiError set)
const errorMsg = assistantMessages[1]!.message as any
expect(errorMsg.apiError).toBe('max_output_tokens')
})
test('stop_reason is null when no message_delta was received (safety fallback path)', async () => {
// Stream ends without message_stop — triggers the safety fallback branch.
// stop_reason stays null since no message_delta was ever seen.
_nextEvents = [
makeMessageStart(),
makeContentBlockStart(0, 'text'),
makeTextDelta(0, 'partial'),
makeContentBlockStop(0),
// No message_delta / message_stop
]
const { assistantMessages } = await runQueryModel(_nextEvents)
// Safety fallback should yield the partial content
expect(assistantMessages).toHaveLength(1)
expect(assistantMessages[0]!.message.stop_reason).toBeNull()
})
})
describe('queryModelOpenAI — usage accumulation', () => {
test('usage in assembled message reflects all four fields from message_delta', async () => {
// message_start has all fields=0 (trailing-chunk pattern: usage not yet available).
// message_delta carries the real values after stream ends.
// The spread in the message_delta handler must override all zeros from message_start,
// including cache_read_input_tokens which was previously missing from message_delta.
_nextEvents = [
makeMessageStart({ usage: { input_tokens: 0, output_tokens: 0, cache_creation_input_tokens: 0, cache_read_input_tokens: 0 } }),
makeContentBlockStart(0, 'text'),
makeTextDelta(0, 'response'),
makeContentBlockStop(0),
// message_delta carries all four Anthropic usage fields (as emitted by the fixed streamAdapter)
{
type: 'message_delta',
delta: { stop_reason: 'end_turn', stop_sequence: null },
usage: { input_tokens: 30011, output_tokens: 190, cache_read_input_tokens: 19904, cache_creation_input_tokens: 0 },
} as any,
makeMessageStop(),
]
const { assistantMessages } = await runQueryModel(_nextEvents)
expect(assistantMessages).toHaveLength(1)
const usage = assistantMessages[0]!.message.usage as any
expect(usage.input_tokens).toBe(30011)
expect(usage.output_tokens).toBe(190)
// cache_read_input_tokens from message_delta overrides the 0 from message_start
expect(usage.cache_read_input_tokens).toBe(19904)
expect(usage.cache_creation_input_tokens).toBe(0)
})
test('usage is zero when no usage events arrive (prevents false autocompact)', async () => {
// If usage stays 0, tokenCountWithEstimation will undercount — so at least
// verify the field exists and is numeric (to detect regressions).
_nextEvents = [
makeMessageStart(),
makeContentBlockStart(0, 'text'),
makeTextDelta(0, 'hi'),
makeContentBlockStop(0),
makeMessageDelta('end_turn', 0),
makeMessageStop(),
]
const { assistantMessages } = await runQueryModel(_nextEvents)
const usage = assistantMessages[0]!.message.usage as any
expect(typeof usage.input_tokens).toBe('number')
expect(typeof usage.output_tokens).toBe('number')
})
})
describe('queryModelOpenAI — no duplicate AssistantMessage (partialMessage reset)', () => {
test('yields exactly one AssistantMessage per message_stop when content is present', async () => {
_nextEvents = [
makeMessageStart(),
makeContentBlockStart(0, 'text'),
makeTextDelta(0, 'only once'),
makeContentBlockStop(0),
makeMessageDelta('end_turn', 5),
makeMessageStop(),
]
const { assistantMessages } = await runQueryModel(_nextEvents)
// Before the fix, partialMessage was not reset to null, so the safety
// fallback at the end of the loop would yield a second message with the
// same message.id — causing mergeAssistantMessages to concatenate content.
expect(assistantMessages).toHaveLength(1)
})
test('thinking + text response yields exactly one AssistantMessage', async () => {
_nextEvents = [
makeMessageStart(),
makeContentBlockStart(0, 'thinking'),
makeThinkingDelta(0, 'let me think'),
makeContentBlockStop(0),
makeContentBlockStart(1, 'text'),
makeTextDelta(1, 'answer'),
makeContentBlockStop(1),
makeMessageDelta('end_turn', 30),
makeMessageStop(),
]
const { assistantMessages } = await runQueryModel(_nextEvents)
expect(assistantMessages).toHaveLength(1)
})
test('safety fallback path still yields message when stream ends without message_stop', async () => {
// Simulates a stream that cuts off without the normal termination sequence.
_nextEvents = [
makeMessageStart(),
makeContentBlockStart(0, 'text'),
makeTextDelta(0, 'abrupt end'),
// No content_block_stop, no message_delta, no message_stop
]
const { assistantMessages } = await runQueryModel(_nextEvents)
expect(assistantMessages).toHaveLength(1)
})
})
describe('queryModelOpenAI — stream_events forwarded', () => {
test('every adapted event is also yielded as stream_event for real-time display', async () => {
_nextEvents = [
makeMessageStart(),
makeContentBlockStart(0, 'text'),
makeTextDelta(0, 'hello'),
makeContentBlockStop(0),
makeMessageDelta('end_turn', 5),
makeMessageStop(),
]
const { streamEvents } = await runQueryModel(_nextEvents)
const eventTypes = streamEvents.map(e => (e as any).event?.type)
expect(eventTypes).toContain('message_start')
expect(eventTypes).toContain('content_block_start')
expect(eventTypes).toContain('content_block_delta')
expect(eventTypes).toContain('content_block_stop')
expect(eventTypes).toContain('message_delta')
expect(eventTypes).toContain('message_stop')
})
})
describe('queryModelOpenAI — max_tokens forwarded to request', () => {
test('buildOpenAIRequestBody includes max_tokens in the request payload', async () => {
_nextEvents = [
makeMessageStart(),
makeContentBlockStart(0, 'text'),
makeTextDelta(0, 'hi'),
makeContentBlockStop(0),
makeMessageDelta('end_turn', 5),
makeMessageStop(),
]
await runQueryModel(_nextEvents)
expect(_lastCreateArgs).not.toBeNull()
expect(_lastCreateArgs!.max_tokens).toBe(8192)
expect(exitCode).toBe(0)
})
})

View File

@@ -1,6 +1,21 @@
import { describe, expect, test } from 'bun:test'
import { adaptOpenAIStreamToAnthropic } from '../streamAdapter.js'
import type { ChatCompletionChunk } from 'openai/resources/chat/completions/completions.mjs'
import { join, dirname } from 'path'
import { fileURLToPath } from 'url'
import { readFileSync, writeFileSync, mkdirSync } from 'fs'
import { tmpdir } from 'os'
// Guard against mock pollution from queryModelOpenAI.test.ts which replaces
// ../streamAdapter.js process-wide via mock.module (bun has no un-mock API).
// We copy the source to a unique temp path so the import bypasses bun's
// module mock cache completely.
const _testDir = dirname(fileURLToPath(import.meta.url))
const _realSource = readFileSync(join(_testDir, '..', 'streamAdapter.ts'), 'utf-8')
const _tempDir = join(tmpdir(), `stream-adapter-test-${Date.now()}`)
mkdirSync(_tempDir, { recursive: true })
const _tempFile = join(_tempDir, 'streamAdapter.ts')
writeFileSync(_tempFile, _realSource, 'utf-8')
const { adaptOpenAIStreamToAnthropic } = await import(_tempFile)
/** Helper to create a mock async iterable from chunk array */
function mockStream(chunks: ChatCompletionChunk[]): AsyncIterable<ChatCompletionChunk> {
@@ -31,6 +46,11 @@ function makeChunk(overrides: Partial<ChatCompletionChunk> & any = {}): ChatComp
/** Collect all emitted Anthropic events from the stream adapter for assertion */
async function collectEvents(chunks: ChatCompletionChunk[]) {
const realModuleUrl = new URL(
`../streamAdapter.js?real=${Date.now()}-${Math.random().toString(36).slice(2)}`,
import.meta.url,
).href
const { adaptOpenAIStreamToAnthropic } = await import(realModuleUrl)
const events: any[] = []
for await (const event of adaptOpenAIStreamToAnthropic(mockStream(chunks), 'gpt-4o')) {
events.push(event)

View File

@@ -8,6 +8,7 @@ import {
} from '../utils/messages.js'
import { getSmallFastModel } from '../utils/model/model.js'
import { asSystemPrompt } from '../utils/systemPromptType.js'
import { getResolvedLanguage } from '../utils/language.js'
import { queryModelWithoutStreaming } from './api/claude.js'
import { getSessionMemoryContent } from './SessionMemory/sessionMemoryUtils.js'
@@ -15,11 +16,18 @@ import { getSessionMemoryContent } from './SessionMemory/sessionMemoryUtils.js'
// large sessions. 30 messages ≈ ~15 exchanges, plenty for "where we left off."
const RECENT_MESSAGE_WINDOW = 30
const PROMPT_EN =
'The user stepped away and is coming back. Write exactly 1-3 short sentences. Start by stating the high-level task — what they are building or debugging, not implementation details. Next: the concrete next step. Skip status reports and commit recaps.'
const PROMPT_ZH =
'用户离开后回来了。用中文写 1-3 句话。先说明用户在做什么(高层目标,不是实现细节),然后说明下一步具体操作。不要写状态报告或提交总结。'
function buildAwaySummaryPrompt(memory: string | null): string {
const memoryBlock = memory
? `Session memory (broader context):\n${memory}\n\n`
: ''
return `${memoryBlock}The user stepped away and is coming back. Write exactly 1-3 short sentences. Start by stating the high-level task — what they are building or debugging, not implementation details. Next: the concrete next step. Skip status reports and commit recaps.`
const prompt = getResolvedLanguage() === 'zh' ? PROMPT_ZH : PROMPT_EN
return `${memoryBlock}${prompt}`
}
/**

View File

@@ -0,0 +1,702 @@
import { mock, describe, test, expect, beforeEach } from 'bun:test'
// Mock @langfuse/otel before any imports
const mockForceFlush = mock(() => Promise.resolve())
const mockShutdown = mock(() => Promise.resolve())
mock.module('@langfuse/otel', () => ({
LangfuseSpanProcessor: class MockLangfuseSpanProcessor {
forceFlush = mockForceFlush
shutdown = mockShutdown
onStart = mock(() => {})
onEnd = mock(() => {})
},
}))
// Mock @opentelemetry/sdk-trace-base
mock.module('@opentelemetry/sdk-trace-base', () => ({
BasicTracerProvider: class MockBasicTracerProvider {
constructor(_opts?: unknown) {}
},
}))
// Mock @langfuse/tracing
const mockChildUpdate = mock(() => {})
const mockChildEnd = mock(() => {})
const mockRootUpdate = mock(() => {})
const mockRootEnd = mock(() => {})
// Mock LangfuseOtelSpanAttributes (re-exported from @langfuse/core)
const mockLangfuseOtelSpanAttributes: Record<string, string> = {
TRACE_SESSION_ID: 'session.id',
TRACE_USER_ID: 'user.id',
OBSERVATION_TYPE: 'observation.type',
OBSERVATION_INPUT: 'observation.input',
OBSERVATION_OUTPUT: 'observation.output',
OBSERVATION_MODEL: 'observation.model',
OBSERVATION_COMPLETION_START_TIME: 'observation.completionStartTime',
OBSERVATION_USAGE_DETAILS: 'observation.usageDetails',
}
const mockSpanContext = {
traceId: 'test-trace-id',
spanId: 'test-span-id',
traceFlags: 1,
}
const mockSetAttribute = mock(() => {})
// Child observation mock (returned by startObservation for tools/generations)
const mockStartObservation = mock(() => ({
id: 'test-span-id',
traceId: 'test-trace-id',
type: 'span',
otelSpan: {
spanContext: () => mockSpanContext,
setAttribute: mockSetAttribute,
},
update: mockRootUpdate,
end: mockRootEnd,
}))
const mockSetLangfuseTracerProvider = mock(() => {})
mock.module('@langfuse/tracing', () => ({
startObservation: mockStartObservation,
LangfuseOtelSpanAttributes: mockLangfuseOtelSpanAttributes,
propagateAttributes: mock((_params: unknown, fn?: () => void) => fn?.()),
setLangfuseTracerProvider: mockSetLangfuseTracerProvider,
}))
// Mock debug logger
mock.module('src/utils/debug.js', () => ({
logForDebugging: mock(() => {}),
logAntError: mock(() => {}),
isDebugToStdErr: () => false,
isDebugMode: () => false,
getDebugLogPath: () => '/tmp/debug.log',
}))
// Mock user module to avoid heavy dependency chain (execa, config, cwd, env, etc.)
mock.module('src/utils/user.js', () => ({
getCoreUserData: () => ({
email: 'test@example.com',
deviceId: 'test-device',
}),
getUserDataForLogging: () => ({}),
}))
describe('Langfuse integration', () => {
beforeEach(() => {
// Reset env
process.env.HOME = '/Users/testuser'
delete process.env.LANGFUSE_PUBLIC_KEY
delete process.env.LANGFUSE_SECRET_KEY
delete process.env.LANGFUSE_BASE_URL
delete process.env.LANGFUSE_USER_ID
mockStartObservation.mockClear()
mockRootUpdate.mockClear()
mockRootEnd.mockClear()
mockForceFlush.mockClear()
mockShutdown.mockClear()
mockSetAttribute.mockClear()
})
// ── sanitize tests ──────────────────────────────────────────────────────────
describe('sanitizeToolInput', () => {
test('replaces home dir in file_path', async () => {
const { sanitizeToolInput } = await import('../sanitize.js')
const home = process.env.HOME ?? '/Users/testuser'
const result = sanitizeToolInput('FileReadTool', {
file_path: `${home}/project/file.ts`,
}) as Record<string, string>
expect(result.file_path).toBe('~/project/file.ts')
})
test('redacts sensitive keys', async () => {
const { sanitizeToolInput } = await import('../sanitize.js')
const result = sanitizeToolInput('MCPTool', {
api_key: 'secret123',
token: 'abc',
}) as Record<string, string>
expect(result.api_key).toBe('[REDACTED]')
expect(result.token).toBe('[REDACTED]')
})
test('returns non-object input unchanged', async () => {
const { sanitizeToolInput } = await import('../sanitize.js')
expect(sanitizeToolInput('BashTool', 'raw string')).toBe('raw string')
expect(sanitizeToolInput('BashTool', null)).toBe(null)
})
})
describe('sanitizeToolOutput', () => {
test('redacts FileReadTool output', async () => {
const { sanitizeToolOutput } = await import('../sanitize.js')
const result = sanitizeToolOutput('FileReadTool', 'file content here')
expect(result).toBe('[file content redacted, 17 chars]')
})
test('redacts FileWriteTool output', async () => {
const { sanitizeToolOutput } = await import('../sanitize.js')
const result = sanitizeToolOutput('FileWriteTool', 'written content')
expect(result).toBe('[file content redacted, 15 chars]')
})
test('truncates BashTool output over 500 chars', async () => {
const { sanitizeToolOutput } = await import('../sanitize.js')
const longOutput = 'x'.repeat(600)
const result = sanitizeToolOutput('BashTool', longOutput)
expect(result).toContain('[truncated]')
expect(result.length).toBeLessThan(600)
})
test('does not truncate BashTool output under 500 chars', async () => {
const { sanitizeToolOutput } = await import('../sanitize.js')
const shortOutput = 'hello world'
expect(sanitizeToolOutput('BashTool', shortOutput)).toBe('hello world')
})
test('redacts ConfigTool output', async () => {
const { sanitizeToolOutput } = await import('../sanitize.js')
const result = sanitizeToolOutput('ConfigTool', 'config data')
expect(result).toBe('[ConfigTool output redacted, 11 chars]')
})
test('redacts MCPTool output', async () => {
const { sanitizeToolOutput } = await import('../sanitize.js')
const result = sanitizeToolOutput('MCPTool', 'mcp data')
expect(result).toBe('[MCPTool output redacted, 8 chars]')
})
})
describe('sanitizeGlobal', () => {
test('replaces home dir in strings', async () => {
const { sanitizeGlobal } = await import('../sanitize.js')
const home = process.env.HOME ?? '/Users/testuser'
expect(sanitizeGlobal(`path: ${home}/file`)).toBe('path: ~/file')
})
test('recursively sanitizes nested objects', async () => {
const { sanitizeGlobal } = await import('../sanitize.js')
const result = sanitizeGlobal({
nested: { api_key: 'secret', name: 'test' },
}) as Record<string, Record<string, string>>
expect(result.nested.api_key).toBe('[REDACTED]')
expect(result.nested.name).toBe('test')
})
test('returns non-string/object values unchanged', async () => {
const { sanitizeGlobal } = await import('../sanitize.js')
expect(sanitizeGlobal(42)).toBe(42)
expect(sanitizeGlobal(true)).toBe(true)
})
})
// ── client tests ────────────────────────────────────────────────────────────
describe('isLangfuseEnabled', () => {
test('returns false when keys not configured', async () => {
const { isLangfuseEnabled } = await import('../client.js')
expect(isLangfuseEnabled()).toBe(false)
})
test('returns true when both keys are set', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { isLangfuseEnabled } = await import('../client.js')
expect(isLangfuseEnabled()).toBe(true)
})
})
describe('initLangfuse', () => {
test('returns false when keys not configured', async () => {
const { initLangfuse } = await import('../client.js')
expect(initLangfuse()).toBe(false)
})
test('returns true when keys are configured', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { isLangfuseEnabled } = await import('../client.js')
expect(isLangfuseEnabled()).toBe(true)
})
test('is idempotent — multiple calls do not re-initialize', async () => {
const { initLangfuse } = await import('../client.js')
expect(() => {
initLangfuse()
initLangfuse()
}).not.toThrow()
})
})
describe('shutdownLangfuse', () => {
test('calls forceFlush and shutdown on processor', async () => {
const { shutdownLangfuse } = await import('../client.js')
await expect(shutdownLangfuse()).resolves.toBeUndefined()
})
})
// ── tracing tests ───────────────────────────────────────────────────────────
describe('createTrace', () => {
test('returns null when langfuse not enabled', async () => {
const { createTrace } = await import('../tracing.js')
const span = createTrace({
sessionId: 's1',
model: 'claude-3',
provider: 'firstParty',
})
expect(span).toBeNull()
})
test('creates root span when enabled', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createTrace } = await import('../tracing.js')
const span = createTrace({
sessionId: 's1',
model: 'claude-3',
provider: 'firstParty',
input: [],
})
expect(span).not.toBeNull()
expect(mockStartObservation).toHaveBeenCalledWith(
'agent-run',
expect.objectContaining({
metadata: expect.objectContaining({
provider: 'firstParty',
model: 'claude-3',
agentType: 'main',
}),
}),
{ asType: 'agent' },
)
// Should set session.id attribute
expect(mockSetAttribute).toHaveBeenCalledWith('session.id', 's1')
})
})
describe('recordLLMObservation', () => {
test('no-ops when rootSpan is null', async () => {
const { recordLLMObservation } = await import('../tracing.js')
recordLLMObservation(null, {
model: 'm',
provider: 'firstParty',
input: [],
output: [],
usage: { input_tokens: 10, output_tokens: 5 },
})
expect(mockStartObservation).toHaveBeenCalledTimes(0)
})
test('records generation child observation via global startObservation', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createTrace, recordLLMObservation } = await import(
'../tracing.js'
)
const span = createTrace({
sessionId: 's1',
model: 'claude-3',
provider: 'firstParty',
})
mockStartObservation.mockClear()
mockRootUpdate.mockClear()
mockRootEnd.mockClear()
recordLLMObservation(span, {
model: 'claude-3',
provider: 'firstParty',
input: [{ role: 'user', content: 'hello' }],
output: [{ role: 'assistant', content: 'hi' }],
usage: { input_tokens: 10, output_tokens: 5 },
})
// Should call the global startObservation with asType: 'generation' and parentSpanContext
expect(mockStartObservation).toHaveBeenCalledWith(
'ChatAnthropic',
expect.objectContaining({
model: 'claude-3',
}),
expect.objectContaining({
asType: 'generation',
parentSpanContext: mockSpanContext,
}),
)
expect(mockRootUpdate).toHaveBeenCalledWith(
expect.objectContaining({
usageDetails: { input: 10, output: 5 },
}),
)
expect(mockRootEnd).toHaveBeenCalled()
})
})
describe('recordToolObservation', () => {
test('no-ops when rootSpan is null', async () => {
const { recordToolObservation } = await import('../tracing.js')
recordToolObservation(null, {
toolName: 'BashTool',
toolUseId: 'id1',
input: {},
output: 'out',
})
})
test('records tool child observation via global startObservation', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createTrace, recordToolObservation } = await import(
'../tracing.js'
)
const span = createTrace({
sessionId: 's1',
model: 'claude-3',
provider: 'firstParty',
})
mockStartObservation.mockClear()
mockRootUpdate.mockClear()
mockRootEnd.mockClear()
recordToolObservation(span, {
toolName: 'BashTool',
toolUseId: 'tu-1',
input: { command: 'ls' },
output: 'file.ts',
})
// Should call the global startObservation with asType: 'tool' and parentSpanContext
expect(mockStartObservation).toHaveBeenCalledWith(
'BashTool',
expect.objectContaining({
input: expect.any(Object),
}),
expect.objectContaining({
asType: 'tool',
parentSpanContext: mockSpanContext,
}),
)
expect(mockRootUpdate).toHaveBeenCalled()
expect(mockRootEnd).toHaveBeenCalled()
})
test('passes startTime to global startObservation', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createTrace, recordToolObservation } = await import(
'../tracing.js'
)
const span = createTrace({
sessionId: 's1',
model: 'claude-3',
provider: 'firstParty',
})
mockStartObservation.mockClear()
const startTime = new Date('2026-01-01T00:00:00Z')
recordToolObservation(span, {
toolName: 'BashTool',
toolUseId: 'tu-2',
input: {},
output: 'out',
startTime,
})
expect(mockStartObservation).toHaveBeenCalledWith(
'BashTool',
expect.any(Object),
expect.objectContaining({
startTime,
parentSpanContext: mockSpanContext,
}),
)
})
test('sanitizes FileReadTool output', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createTrace, recordToolObservation } = await import(
'../tracing.js'
)
const span = createTrace({
sessionId: 's1',
model: 'claude-3',
provider: 'firstParty',
})
mockRootUpdate.mockClear()
recordToolObservation(span, {
toolName: 'FileReadTool',
toolUseId: 'tu-2',
input: { file_path: '/tmp/file.ts' },
output: 'file content here',
})
expect(mockRootUpdate).toHaveBeenCalledWith(
expect.objectContaining({
output: '[file content redacted, 17 chars]',
}),
)
})
test('sets ERROR level for error observations', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createTrace, recordToolObservation } = await import(
'../tracing.js'
)
const span = createTrace({
sessionId: 's1',
model: 'claude-3',
provider: 'firstParty',
})
mockRootUpdate.mockClear()
recordToolObservation(span, {
toolName: 'BashTool',
toolUseId: 'tu-3',
input: {},
output: 'error occurred',
isError: true,
})
expect(mockRootUpdate).toHaveBeenCalledWith(
expect.objectContaining({ level: 'ERROR' }),
)
})
})
describe('endTrace', () => {
test('no-ops when rootSpan is null', async () => {
const { endTrace } = await import('../tracing.js')
endTrace(null)
expect(mockRootEnd).not.toHaveBeenCalled()
})
test('calls span.end()', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createTrace, endTrace } = await import('../tracing.js')
const span = createTrace({
sessionId: 's1',
model: 'claude-3',
provider: 'firstParty',
})
mockRootEnd.mockClear()
endTrace(span)
expect(mockRootEnd).toHaveBeenCalled()
})
test('calls span.update() with output when provided', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createTrace, endTrace } = await import('../tracing.js')
const span = createTrace({
sessionId: 's1',
model: 'claude-3',
provider: 'firstParty',
})
mockRootUpdate.mockClear()
mockRootEnd.mockClear()
endTrace(span, 'final output')
expect(mockRootUpdate).toHaveBeenCalledWith(
expect.objectContaining({ output: 'final output' }),
)
expect(mockRootEnd).toHaveBeenCalled()
})
})
describe('createSubagentTrace', () => {
test('returns null when langfuse not enabled', async () => {
const { createSubagentTrace } = await import('../tracing.js')
const span = createSubagentTrace({
sessionId: 's1',
agentType: 'Explore',
agentId: 'agent-1',
model: 'claude-3',
provider: 'firstParty',
})
expect(span).toBeNull()
})
test('creates trace with agentType and agentId metadata', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createSubagentTrace } = await import('../tracing.js')
const span = createSubagentTrace({
sessionId: 's1',
agentType: 'Explore',
agentId: 'agent-1',
model: 'claude-3',
provider: 'firstParty',
input: [{ role: 'user', content: 'search for X' }],
})
expect(span).not.toBeNull()
expect(mockStartObservation).toHaveBeenCalledWith(
'agent:Explore',
expect.objectContaining({
metadata: expect.objectContaining({
agentType: 'Explore',
agentId: 'agent-1',
provider: 'firstParty',
model: 'claude-3',
}),
}),
{ asType: 'agent' },
)
// Verify session.id attribute is set
expect(mockSetAttribute).toHaveBeenCalledWith('session.id', 's1')
})
test('returns null on SDK error', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
mockStartObservation.mockImplementationOnce(() => {
throw new Error('SDK error')
})
const { createSubagentTrace } = await import('../tracing.js')
const span = createSubagentTrace({
sessionId: 's1',
agentType: 'Plan',
agentId: 'agent-2',
model: 'claude-3',
provider: 'firstParty',
})
expect(span).toBeNull()
})
})
describe('createTrace with querySource', () => {
test('includes querySource in metadata', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createTrace } = await import('../tracing.js')
const span = createTrace({
sessionId: 's1',
model: 'claude-3',
provider: 'firstParty',
querySource: 'user',
})
expect(span).not.toBeNull()
expect(mockStartObservation).toHaveBeenCalledWith(
'agent-run:user',
expect.objectContaining({
metadata: expect.objectContaining({
agentType: 'main',
querySource: 'user',
}),
}),
{ asType: 'agent' },
)
})
test('omits querySource when not provided', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
mockStartObservation.mockClear()
const { createTrace } = await import('../tracing.js')
createTrace({
sessionId: 's1',
model: 'claude-3',
provider: 'firstParty',
})
const calls = mockStartObservation.mock.calls as unknown[][]
const secondArg = calls[0]?.[1] as Record<string, unknown> | undefined
const metadata = (secondArg?.metadata ?? {}) as Record<string, unknown>
expect(metadata).not.toHaveProperty('querySource')
})
})
describe('nested agent scenario', () => {
test('sub-agent trace shares sessionId with parent', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createTrace, createSubagentTrace } = await import('../tracing.js')
mockSetAttribute.mockClear()
// Create parent trace
const parentSpan = createTrace({
sessionId: 'shared-session',
model: 'claude-3',
provider: 'firstParty',
})
// Create sub-agent trace with same sessionId
const subSpan = createSubagentTrace({
sessionId: 'shared-session',
agentType: 'Explore',
agentId: 'agent-explore-1',
model: 'claude-3',
provider: 'firstParty',
})
expect(parentSpan).not.toBeNull()
expect(subSpan).not.toBeNull()
// Both should have set session.id attribute
const sessionAttributeCalls = mockSetAttribute.mock.calls.filter(
(call: unknown[]) =>
Array.isArray(call) &&
call[0] === 'session.id' &&
call[1] === 'shared-session',
)
expect(sessionAttributeCalls.length).toBeGreaterThanOrEqual(2)
})
test('query reuses passed langfuseTrace instead of creating new one', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createSubagentTrace } = await import('../tracing.js')
const subTrace = createSubagentTrace({
sessionId: 's1',
agentType: 'Explore',
agentId: 'agent-1',
model: 'claude-3',
provider: 'firstParty',
})
expect(subTrace).not.toBeNull()
// Simulate query.ts logic: if langfuseTrace already set, don't create new one
const ownsTrace = false
const langfuseTrace = subTrace
expect(ownsTrace).toBe(false)
expect(langfuseTrace).toBe(subTrace)
})
})
describe('SDK exceptions do not affect main flow', () => {
test('createTrace returns null on SDK error', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
mockStartObservation.mockImplementationOnce(() => {
throw new Error('SDK error')
})
const { createTrace } = await import('../tracing.js')
const span = createTrace({
sessionId: 's1',
model: 'claude-3',
provider: 'firstParty',
})
expect(span).toBeNull()
})
test('recordLLMObservation silently fails on SDK error', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createTrace, recordLLMObservation } = await import(
'../tracing.js'
)
const span = createTrace({
sessionId: 's1',
model: 'claude-3',
provider: 'firstParty',
})
// The next call to startObservation (for the generation) will throw
mockStartObservation.mockImplementationOnce(() => {
throw new Error('SDK error')
})
expect(() =>
recordLLMObservation(span, {
model: 'm',
provider: 'firstParty',
input: [],
output: [],
usage: { input_tokens: 1, output_tokens: 1 },
}),
).not.toThrow()
})
})
})

View File

@@ -1,683 +1,40 @@
import { mock, describe, test, expect, beforeEach } from 'bun:test'
import { describe, expect, test } from 'bun:test'
import { fileURLToPath } from 'node:url'
// Mock @langfuse/otel before any imports
const mockForceFlush = mock(() => Promise.resolve())
const mockShutdown = mock(() => Promise.resolve())
mock.module('@langfuse/otel', () => ({
LangfuseSpanProcessor: class MockLangfuseSpanProcessor {
forceFlush = mockForceFlush
shutdown = mockShutdown
onStart = mock(() => {})
onEnd = mock(() => {})
},
}))
// Mock @opentelemetry/sdk-trace-base
mock.module('@opentelemetry/sdk-trace-base', () => ({
BasicTracerProvider: class MockBasicTracerProvider {
constructor(_opts?: unknown) {}
},
}))
// Mock @langfuse/tracing
const mockChildUpdate = mock(() => {})
const mockChildEnd = mock(() => {})
const mockRootUpdate = mock(() => {})
const mockRootEnd = mock(() => {})
// Mock LangfuseOtelSpanAttributes (re-exported from @langfuse/core)
const mockLangfuseOtelSpanAttributes: Record<string, string> = {
TRACE_SESSION_ID: 'session.id',
TRACE_USER_ID: 'user.id',
OBSERVATION_TYPE: 'observation.type',
OBSERVATION_INPUT: 'observation.input',
OBSERVATION_OUTPUT: 'observation.output',
OBSERVATION_MODEL: 'observation.model',
OBSERVATION_COMPLETION_START_TIME: 'observation.completionStartTime',
OBSERVATION_USAGE_DETAILS: 'observation.usageDetails',
}
const mockSpanContext = { traceId: 'test-trace-id', spanId: 'test-span-id', traceFlags: 1 }
const mockSetAttribute = mock(() => {})
// Child observation mock (returned by rootSpan.startObservation for tools)
const mockChildStartObservation = mock(() => ({
id: 'child-id',
update: mockChildUpdate,
end: mockChildEnd,
}))
const mockStartObservation = mock(() => ({
id: 'test-span-id',
traceId: 'test-trace-id',
type: 'span',
otelSpan: {
spanContext: () => mockSpanContext,
setAttribute: mockSetAttribute,
},
update: mockRootUpdate,
end: mockRootEnd,
// Instance method — used by recordToolObservation
startObservation: mockChildStartObservation,
}))
const mockSetLangfuseTracerProvider = mock(() => {})
mock.module('@langfuse/tracing', () => ({
startObservation: mockStartObservation,
LangfuseOtelSpanAttributes: mockLangfuseOtelSpanAttributes,
propagateAttributes: mock((_params: unknown, fn?: () => void) => fn?.()),
setLangfuseTracerProvider: mockSetLangfuseTracerProvider,
}))
// Mock debug logger
mock.module('src/utils/debug.js', () => ({
logForDebugging: mock(() => {}),
}))
// Mock user data — resolveLangfuseUserId uses getCoreUserData().email and .deviceId
mock.module('src/utils/user.js', () => ({
getCoreUserData: mock(() => ({
email: 'test-device-id',
deviceId: 'test-device-id',
})),
}))
const isolatedPath = fileURLToPath(
new URL('./langfuse.isolated.ts', import.meta.url),
)
describe('Langfuse integration', () => {
beforeEach(() => {
// Reset env
delete process.env.LANGFUSE_PUBLIC_KEY
delete process.env.LANGFUSE_SECRET_KEY
delete process.env.LANGFUSE_BASE_URL
mockStartObservation.mockClear()
mockChildStartObservation.mockClear()
mockChildUpdate.mockClear()
mockChildEnd.mockClear()
mockRootUpdate.mockClear()
mockRootEnd.mockClear()
mockForceFlush.mockClear()
mockShutdown.mockClear()
mockSetAttribute.mockClear()
})
// ── sanitize tests ──────────────────────────────────────────────────────────
describe('sanitizeToolInput', () => {
test('replaces home dir in file_path', async () => {
const { sanitizeToolInput } = await import('../sanitize.js')
const home = process.env.HOME ?? '/Users/testuser'
const result = sanitizeToolInput('FileReadTool', { file_path: `${home}/project/file.ts` }) as Record<string, string>
expect(result.file_path).toBe('~/project/file.ts')
test('passes in isolated subprocess', async () => {
const proc = Bun.spawn({
cmd: [process.execPath, 'test', isolatedPath],
cwd: process.cwd(),
stdout: 'pipe',
stderr: 'pipe',
env: process.env,
})
test('redacts sensitive keys', async () => {
const { sanitizeToolInput } = await import('../sanitize.js')
const result = sanitizeToolInput('MCPTool', { api_key: 'secret123', token: 'abc' }) as Record<string, string>
expect(result.api_key).toBe('[REDACTED]')
expect(result.token).toBe('[REDACTED]')
})
const [stdout, stderr, exitCode] = await Promise.all([
new Response(proc.stdout).text(),
new Response(proc.stderr).text(),
proc.exited,
])
test('returns non-object input unchanged', async () => {
const { sanitizeToolInput } = await import('../sanitize.js')
expect(sanitizeToolInput('BashTool', 'raw string')).toBe('raw string')
expect(sanitizeToolInput('BashTool', null)).toBe(null)
})
})
describe('sanitizeToolOutput', () => {
test('redacts FileReadTool output', async () => {
const { sanitizeToolOutput } = await import('../sanitize.js')
const result = sanitizeToolOutput('FileReadTool', 'file content here')
expect(result).toBe('[file content redacted, 17 chars]')
})
test('redacts FileWriteTool output', async () => {
const { sanitizeToolOutput } = await import('../sanitize.js')
const result = sanitizeToolOutput('FileWriteTool', 'written content')
expect(result).toBe('[file content redacted, 15 chars]')
})
test('truncates BashTool output over 500 chars', async () => {
const { sanitizeToolOutput } = await import('../sanitize.js')
const longOutput = 'x'.repeat(600)
const result = sanitizeToolOutput('BashTool', longOutput)
expect(result).toContain('[truncated]')
expect(result.length).toBeLessThan(600)
})
test('does not truncate BashTool output under 500 chars', async () => {
const { sanitizeToolOutput } = await import('../sanitize.js')
const shortOutput = 'hello world'
expect(sanitizeToolOutput('BashTool', shortOutput)).toBe('hello world')
})
test('redacts ConfigTool output', async () => {
const { sanitizeToolOutput } = await import('../sanitize.js')
const result = sanitizeToolOutput('ConfigTool', 'config data')
expect(result).toBe('[ConfigTool output redacted, 11 chars]')
})
test('redacts MCPTool output', async () => {
const { sanitizeToolOutput } = await import('../sanitize.js')
const result = sanitizeToolOutput('MCPTool', 'mcp data')
expect(result).toBe('[MCPTool output redacted, 8 chars]')
})
})
describe('sanitizeGlobal', () => {
test('replaces home dir in strings', async () => {
const { sanitizeGlobal } = await import('../sanitize.js')
const home = process.env.HOME ?? '/Users/testuser'
expect(sanitizeGlobal(`path: ${home}/file`)).toBe('path: ~/file')
})
test('recursively sanitizes nested objects', async () => {
const { sanitizeGlobal } = await import('../sanitize.js')
const result = sanitizeGlobal({ nested: { api_key: 'secret', name: 'test' } }) as Record<string, Record<string, string>>
expect(result.nested.api_key).toBe('[REDACTED]')
expect(result.nested.name).toBe('test')
})
test('returns non-string/object values unchanged', async () => {
const { sanitizeGlobal } = await import('../sanitize.js')
expect(sanitizeGlobal(42)).toBe(42)
expect(sanitizeGlobal(true)).toBe(true)
})
})
// ── client tests ────────────────────────────────────────────────────────────
describe('isLangfuseEnabled', () => {
test('returns false when keys not configured', async () => {
const { isLangfuseEnabled } = await import('../client.js')
expect(isLangfuseEnabled()).toBe(false)
})
test('returns true when both keys are set', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { isLangfuseEnabled } = await import('../client.js')
expect(isLangfuseEnabled()).toBe(true)
})
})
describe('initLangfuse', () => {
test('returns false when keys not configured', async () => {
const { initLangfuse } = await import('../client.js')
expect(initLangfuse()).toBe(false)
})
test('returns true when keys are configured', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
// client.js is a singleton — test via isLangfuseEnabled which reads env directly
const { isLangfuseEnabled } = await import('../client.js')
expect(isLangfuseEnabled()).toBe(true)
})
test('is idempotent — multiple calls do not re-initialize', async () => {
// client.js singleton: once processor is set, initLangfuse returns true immediately
// We verify this by checking that calling it multiple times doesn't throw
const { initLangfuse } = await import('../client.js')
expect(() => { initLangfuse(); initLangfuse() }).not.toThrow()
})
})
describe('shutdownLangfuse', () => {
test('calls forceFlush and shutdown on processor', async () => {
// Verify shutdown is callable without error even when no processor is set
const { shutdownLangfuse } = await import('../client.js')
await expect(shutdownLangfuse()).resolves.toBeUndefined()
})
})
// ── tracing tests ───────────────────────────────────────────────────────────
describe('createTrace', () => {
test('returns null when langfuse not enabled', async () => {
const { createTrace } = await import('../tracing.js')
const span = createTrace({ sessionId: 's1', model: 'claude-3', provider: 'firstParty' })
expect(span).toBeNull()
})
test('creates root span when enabled', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createTrace } = await import('../tracing.js')
const span = createTrace({ sessionId: 's1', model: 'claude-3', provider: 'firstParty', input: [] })
expect(span).not.toBeNull()
expect(mockStartObservation).toHaveBeenCalledWith('agent-run', expect.objectContaining({
metadata: expect.objectContaining({ provider: 'firstParty', model: 'claude-3' }),
}), { asType: 'agent' })
})
})
describe('recordLLMObservation', () => {
test('no-ops when rootSpan is null', async () => {
const { recordLLMObservation } = await import('../tracing.js')
recordLLMObservation(null, { model: 'm', provider: 'firstParty', input: [], output: [], usage: { input_tokens: 10, output_tokens: 5 } })
expect(mockStartObservation).toHaveBeenCalledTimes(0)
})
test('records generation child observation via global startObservation', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createTrace, recordLLMObservation } = await import('../tracing.js')
const span = createTrace({ sessionId: 's1', model: 'claude-3', provider: 'firstParty' })
mockStartObservation.mockClear()
recordLLMObservation(span, {
model: 'claude-3',
provider: 'firstParty',
input: [{ role: 'user', content: 'hello' }],
output: [{ role: 'assistant', content: 'hi' }],
usage: { input_tokens: 10, output_tokens: 5 },
})
// Should call the global startObservation with asType: 'generation' and parentSpanContext
expect(mockStartObservation).toHaveBeenCalledWith('ChatAnthropic', expect.objectContaining({
model: 'claude-3',
}), expect.objectContaining({
asType: 'generation',
parentSpanContext: mockSpanContext,
}))
expect(mockRootUpdate).toHaveBeenCalledWith(expect.objectContaining({
usageDetails: { input: 10, output: 5 },
}))
expect(mockRootEnd).toHaveBeenCalled()
})
test('includes cache tokens in usageDetails when provided', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createTrace, recordLLMObservation } = await import('../tracing.js')
const span = createTrace({ sessionId: 's1', model: 'claude-3', provider: 'firstParty' })
mockStartObservation.mockClear()
mockRootUpdate.mockClear()
recordLLMObservation(span, {
model: 'claude-3',
provider: 'firstParty',
input: [],
output: [],
usage: { input_tokens: 10000, output_tokens: 50, cache_creation_input_tokens: 2000, cache_read_input_tokens: 7000 },
})
expect(mockRootUpdate).toHaveBeenCalledWith(expect.objectContaining({
usageDetails: {
input: 19000, // 10000 + 2000 + 7000
output: 50,
cache_read: 7000,
cache_creation: 2000,
},
}))
})
test('omits cache fields when not provided', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createTrace, recordLLMObservation } = await import('../tracing.js')
const span = createTrace({ sessionId: 's1', model: 'claude-3', provider: 'firstParty' })
mockRootUpdate.mockClear()
recordLLMObservation(span, {
model: 'claude-3',
provider: 'firstParty',
input: [],
output: [],
usage: { input_tokens: 100, output_tokens: 20 },
})
expect(mockRootUpdate).toHaveBeenCalledWith(expect.objectContaining({
usageDetails: { input: 100, output: 20 },
}))
})
})
describe('recordToolObservation', () => {
test('no-ops when rootSpan is null', async () => {
const { recordToolObservation } = await import('../tracing.js')
recordToolObservation(null, { toolName: 'BashTool', toolUseId: 'id1', input: {}, output: 'out' })
// startObservation should not be called beyond the initial trace creation (none here)
})
test('records tool child observation via global startObservation', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createTrace, recordToolObservation } = await import('../tracing.js')
const span = createTrace({ sessionId: 's1', model: 'claude-3', provider: 'firstParty' })
mockStartObservation.mockClear()
mockRootUpdate.mockClear()
mockRootEnd.mockClear()
recordToolObservation(span, {
toolName: 'BashTool',
toolUseId: 'tu-1',
input: { command: 'ls' },
output: 'file.ts',
})
// Should call the global startObservation with asType: 'tool' and parentSpanContext
expect(mockStartObservation).toHaveBeenCalledWith('BashTool', expect.objectContaining({
input: expect.any(Object),
}), expect.objectContaining({
asType: 'tool',
parentSpanContext: mockSpanContext,
}))
expect(mockRootUpdate).toHaveBeenCalled()
expect(mockRootEnd).toHaveBeenCalled()
})
test('passes startTime to global startObservation', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createTrace, recordToolObservation } = await import('../tracing.js')
const span = createTrace({ sessionId: 's1', model: 'claude-3', provider: 'firstParty' })
mockStartObservation.mockClear()
const startTime = new Date('2026-01-01T00:00:00Z')
recordToolObservation(span, {
toolName: 'BashTool',
toolUseId: 'tu-2',
input: {},
output: 'out',
startTime,
})
expect(mockStartObservation).toHaveBeenCalledWith('BashTool', expect.any(Object), expect.objectContaining({
startTime,
parentSpanContext: mockSpanContext,
}))
})
test('sanitizes FileReadTool output', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createTrace, recordToolObservation } = await import('../tracing.js')
const span = createTrace({ sessionId: 's1', model: 'claude-3', provider: 'firstParty' })
mockRootUpdate.mockClear()
recordToolObservation(span, {
toolName: 'FileReadTool',
toolUseId: 'tu-2',
input: { file_path: '/tmp/file.ts' },
output: 'file content here',
})
expect(mockRootUpdate).toHaveBeenCalledWith(expect.objectContaining({
output: '[file content redacted, 17 chars]',
}))
})
test('sets ERROR level for error observations', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createTrace, recordToolObservation } = await import('../tracing.js')
const span = createTrace({ sessionId: 's1', model: 'claude-3', provider: 'firstParty' })
mockRootUpdate.mockClear()
recordToolObservation(span, {
toolName: 'BashTool',
toolUseId: 'tu-3',
input: {},
output: 'error occurred',
isError: true,
})
expect(mockRootUpdate).toHaveBeenCalledWith(expect.objectContaining({ level: 'ERROR' }))
})
})
describe('endTrace', () => {
test('no-ops when rootSpan is null', async () => {
const { endTrace } = await import('../tracing.js')
endTrace(null)
expect(mockRootEnd).not.toHaveBeenCalled()
})
test('calls span.end()', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createTrace, endTrace } = await import('../tracing.js')
const span = createTrace({ sessionId: 's1', model: 'claude-3', provider: 'firstParty' })
endTrace(span)
expect(mockRootEnd).toHaveBeenCalled()
})
test('calls span.update() with output when provided', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createTrace, endTrace } = await import('../tracing.js')
const span = createTrace({ sessionId: 's1', model: 'claude-3', provider: 'firstParty' })
endTrace(span, 'final output')
expect(mockRootUpdate).toHaveBeenCalledWith({ output: 'final output' })
expect(mockRootEnd).toHaveBeenCalled()
})
})
describe('createSubagentTrace', () => {
test('returns null when langfuse not enabled', async () => {
const { createSubagentTrace } = await import('../tracing.js')
const span = createSubagentTrace({
sessionId: 's1',
agentType: 'Explore',
agentId: 'agent-1',
model: 'claude-3',
provider: 'firstParty',
})
expect(span).toBeNull()
})
test('creates trace with agentType and agentId metadata', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createSubagentTrace } = await import('../tracing.js')
const span = createSubagentTrace({
sessionId: 's1',
agentType: 'Explore',
agentId: 'agent-1',
model: 'claude-3',
provider: 'firstParty',
input: [{ role: 'user', content: 'search for X' }],
})
expect(span).not.toBeNull()
expect(mockStartObservation).toHaveBeenCalledWith('agent:Explore', expect.objectContaining({
metadata: expect.objectContaining({
agentType: 'Explore',
agentId: 'agent-1',
provider: 'firstParty',
model: 'claude-3',
}),
}), { asType: 'agent' })
// Verify session.id attribute is set
expect(mockSetAttribute).toHaveBeenCalledWith('session.id', 's1')
})
test('returns null on SDK error', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
mockStartObservation.mockImplementationOnce(() => { throw new Error('SDK error') })
const { createSubagentTrace } = await import('../tracing.js')
const span = createSubagentTrace({
sessionId: 's1',
agentType: 'Plan',
agentId: 'agent-2',
model: 'claude-3',
provider: 'firstParty',
})
expect(span).toBeNull()
})
})
describe('createTrace with querySource', () => {
test('includes querySource in metadata', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createTrace } = await import('../tracing.js')
const span = createTrace({
sessionId: 's1',
model: 'claude-3',
provider: 'firstParty',
querySource: 'user',
})
expect(span).not.toBeNull()
expect(mockStartObservation).toHaveBeenCalledWith('agent-run:user', expect.objectContaining({
metadata: expect.objectContaining({
agentType: 'main',
querySource: 'user',
}),
}), { asType: 'agent' })
})
test('omits querySource when not provided', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
mockStartObservation.mockClear()
const { createTrace } = await import('../tracing.js')
createTrace({ sessionId: 's1', model: 'claude-3', provider: 'firstParty' })
const calls = mockStartObservation.mock.calls as unknown[][]
const secondArg = calls[0]?.[1] as Record<string, unknown> | undefined
const metadata = (secondArg?.metadata ?? {}) as Record<string, unknown>
expect(metadata).not.toHaveProperty('querySource')
})
})
describe('createTrace with username', () => {
test('sets user.id attribute when username is provided', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
mockSetAttribute.mockClear()
const { createTrace } = await import('../tracing.js')
const span = createTrace({
sessionId: 's1',
model: 'claude-3',
provider: 'firstParty',
username: 'user@example.com',
})
expect(span).not.toBeNull()
expect(mockSetAttribute).toHaveBeenCalledWith('user.id', 'user@example.com')
})
test('falls back to LANGFUSE_USER_ID env when username not provided', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
process.env.LANGFUSE_USER_ID = 'env-user@test.com'
mockSetAttribute.mockClear()
const { createTrace } = await import('../tracing.js')
const span = createTrace({
sessionId: 's1',
model: 'claude-3',
provider: 'firstParty',
})
expect(span).not.toBeNull()
expect(mockSetAttribute).toHaveBeenCalledWith('user.id', 'env-user@test.com')
delete process.env.LANGFUSE_USER_ID
})
test('falls back to deviceId when neither username nor env is provided', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
delete process.env.LANGFUSE_USER_ID
mockSetAttribute.mockClear()
const { createTrace } = await import('../tracing.js')
createTrace({ sessionId: 's1', model: 'claude-3', provider: 'firstParty' })
// Falls back to getCoreUserData().deviceId (mocked as 'test-device-id')
expect(mockSetAttribute).toHaveBeenCalledWith('user.id', 'test-device-id')
})
test('username takes precedence over LANGFUSE_USER_ID env', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
process.env.LANGFUSE_USER_ID = 'env-user@test.com'
mockSetAttribute.mockClear()
const { createTrace } = await import('../tracing.js')
createTrace({
sessionId: 's1',
model: 'claude-3',
provider: 'firstParty',
username: 'param-user@test.com',
})
const userIdCalls = mockSetAttribute.mock.calls.filter(
(call: unknown[]) => Array.isArray(call) && call[0] === 'user.id',
if (exitCode !== 0) {
throw new Error(
[
`isolated langfuse test failed with exit code ${exitCode}`,
'',
'STDOUT:',
stdout,
'',
'STDERR:',
stderr,
].join('\n'),
)
expect(userIdCalls.length).toBe(1)
expect((userIdCalls[0] as unknown[])[1]).toBe('param-user@test.com')
delete process.env.LANGFUSE_USER_ID
})
})
}
describe('nested agent scenario', () => {
test('sub-agent trace shares sessionId with parent', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createTrace, createSubagentTrace } = await import('../tracing.js')
mockSetAttribute.mockClear()
// Create parent trace
const parentSpan = createTrace({
sessionId: 'shared-session',
model: 'claude-3',
provider: 'firstParty',
})
// Create sub-agent trace with same sessionId
const subSpan = createSubagentTrace({
sessionId: 'shared-session',
agentType: 'Explore',
agentId: 'agent-explore-1',
model: 'claude-3',
provider: 'firstParty',
})
expect(parentSpan).not.toBeNull()
expect(subSpan).not.toBeNull()
// Both should have set session.id attribute
const sessionAttributeCalls = mockSetAttribute.mock.calls.filter(
(call: unknown[]) => Array.isArray(call) && call[0] === 'session.id' && call[1] === 'shared-session',
)
expect(sessionAttributeCalls.length).toBeGreaterThanOrEqual(2)
})
test('query reuses passed langfuseTrace instead of creating new one', async () => {
// This validates the pattern used in query.ts:
// const ownsTrace = !params.toolUseContext.langfuseTrace
// const langfuseTrace = params.toolUseContext.langfuseTrace ?? createTrace(...)
// When langfuseTrace is already set, createTrace should NOT be called
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
const { createSubagentTrace } = await import('../tracing.js')
// Simulate what runAgent does: create subTrace, then pass it as langfuseTrace
const subTrace = createSubagentTrace({
sessionId: 's1',
agentType: 'Explore',
agentId: 'agent-1',
model: 'claude-3',
provider: 'firstParty',
})
expect(subTrace).not.toBeNull()
// Simulate query.ts logic: if langfuseTrace already set, don't create new one
const ownsTrace = false // Would be: !params.toolUseContext.langfuseTrace
const langfuseTrace = subTrace // Would be: params.toolUseContext.langfuseTrace ?? createTrace(...)
expect(ownsTrace).toBe(false)
expect(langfuseTrace).toBe(subTrace)
})
})
describe('SDK exceptions do not affect main flow', () => {
test('createTrace returns null on SDK error', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
mockStartObservation.mockImplementationOnce(() => { throw new Error('SDK error') })
const { createTrace } = await import('../tracing.js')
const span = createTrace({ sessionId: 's1', model: 'claude-3', provider: 'firstParty' })
expect(span).toBeNull()
})
test('recordLLMObservation silently fails on SDK error', async () => {
process.env.LANGFUSE_PUBLIC_KEY = 'pk-test'
process.env.LANGFUSE_SECRET_KEY = 'sk-test'
mockStartObservation.mockImplementationOnce(() => { throw new Error('SDK error') })
const { createTrace, recordLLMObservation } = await import('../tracing.js')
const span = createTrace({ sessionId: 's1', model: 'claude-3', provider: 'firstParty' })
// The second call to startObservation (for the generation) will throw
mockStartObservation.mockImplementationOnce(() => { throw new Error('SDK error') })
expect(() => recordLLMObservation(span, {
model: 'm',
provider: 'firstParty',
input: [],
output: [],
usage: { input_tokens: 1, output_tokens: 1 },
})).not.toThrow()
})
expect(exitCode).toBe(0)
})
})

View File

@@ -9,19 +9,14 @@
* 4. Can be idle (waiting for work) or active (processing)
*/
import {
isTerminalTaskStatus,
type SetAppState,
type Task,
type TaskStateBase,
} from '../../Task.js'
import type { Message } from '../../types/message.js'
import { logForDebugging } from '../../utils/debug.js'
import { createUserMessage } from '../../utils/messages.js'
import { killInProcessTeammate } from '../../utils/swarm/spawnInProcess.js'
import { updateTaskState } from '../../utils/task/framework.js'
import type { InProcessTeammateTaskState } from './types.js'
import { appendCappedMessage, isInProcessTeammateTask } from './types.js'
import { isTerminalTaskStatus, type SetAppState, type Task, type TaskStateBase } from '../../Task.js';
import type { Message, MessageOrigin } from '../../types/message.js';
import { logForDebugging } from '../../utils/debug.js';
import { createUserMessage } from '../../utils/messages.js';
import { killInProcessTeammate } from '../../utils/swarm/spawnInProcess.js';
import { updateTaskState } from '../../utils/task/framework.js';
import type { InProcessTeammateTaskState, PendingTeammateUserMessage } from './types.js';
import { appendCappedMessage, isInProcessTeammateTask } from './types.js';
/**
* InProcessTeammateTask - Handles in-process teammate execution.
@@ -30,48 +25,41 @@ export const InProcessTeammateTask: Task = {
name: 'InProcessTeammateTask',
type: 'in_process_teammate',
async kill(taskId, setAppState) {
killInProcessTeammate(taskId, setAppState)
killInProcessTeammate(taskId, setAppState);
},
}
};
/**
* Request shutdown for a teammate.
*/
export function requestTeammateShutdown(
taskId: string,
setAppState: SetAppState,
): void {
export function requestTeammateShutdown(taskId: string, setAppState: SetAppState): void {
updateTaskState<InProcessTeammateTaskState>(taskId, setAppState, task => {
if (task.status !== 'running' || task.shutdownRequested) {
return task
return task;
}
return {
...task,
shutdownRequested: true,
}
})
};
});
}
/**
* Append a message to a teammate's conversation history.
* Used for zoomed view to show the teammate's conversation.
*/
export function appendTeammateMessage(
taskId: string,
message: Message,
setAppState: SetAppState,
): void {
export function appendTeammateMessage(taskId: string, message: Message, setAppState: SetAppState): void {
updateTaskState<InProcessTeammateTaskState>(taskId, setAppState, task => {
if (task.status !== 'running') {
return task
return task;
}
return {
...task,
messages: appendCappedMessage(task.messages, message),
}
})
};
});
}
/**
@@ -82,27 +70,47 @@ export function appendTeammateMessage(
export function injectUserMessageToTeammate(
taskId: string,
message: string,
options:
| {
autonomyRunId?: string;
origin?: MessageOrigin;
}
| undefined,
setAppState: SetAppState,
): void {
): boolean {
let injected = false;
updateTaskState<InProcessTeammateTaskState>(taskId, setAppState, task => {
// Allow message injection when teammate is running or idle (waiting for input)
// Only reject if teammate is in a terminal state
if (isTerminalTaskStatus(task.status)) {
logForDebugging(
`Dropping message for teammate task ${taskId}: task status is "${task.status}"`,
)
return task
logForDebugging(`Dropping message for teammate task ${taskId}: task status is "${task.status}"`);
return task;
}
injected = true;
const pendingMessage: PendingTeammateUserMessage = { message };
if (options?.autonomyRunId !== undefined) {
pendingMessage.autonomyRunId = options.autonomyRunId;
}
if (options?.origin !== undefined) {
pendingMessage.origin = options.origin;
}
const userMessageArgs: Parameters<typeof createUserMessage>[0] = {
content: message,
};
if (options?.origin !== undefined) {
userMessageArgs.origin = options.origin;
}
return {
...task,
pendingUserMessages: [...task.pendingUserMessages, message],
messages: appendCappedMessage(
task.messages,
createUserMessage({ content: message }),
),
}
})
pendingUserMessages: [...task.pendingUserMessages, pendingMessage],
messages: appendCappedMessage(task.messages, createUserMessage(userMessageArgs)),
};
});
return injected;
}
/**
@@ -115,30 +123,28 @@ export function findTeammateTaskByAgentId(
agentId: string,
tasks: Record<string, TaskStateBase>,
): InProcessTeammateTaskState | undefined {
let fallback: InProcessTeammateTaskState | undefined
let fallback: InProcessTeammateTaskState | undefined;
for (const task of Object.values(tasks)) {
if (isInProcessTeammateTask(task) && task.identity.agentId === agentId) {
// Prefer running tasks in case old killed tasks still exist in AppState
// alongside new running ones with the same agentId
if (task.status === 'running') {
return task
return task;
}
// Keep first match as fallback in case no running task exists
if (!fallback) {
fallback = task
fallback = task;
}
}
}
return fallback
return fallback;
}
/**
* Get all in-process teammate tasks from AppState.
*/
export function getAllInProcessTeammateTasks(
tasks: Record<string, TaskStateBase>,
): InProcessTeammateTaskState[] {
return Object.values(tasks).filter(isInProcessTeammateTask)
export function getAllInProcessTeammateTasks(tasks: Record<string, TaskStateBase>): InProcessTeammateTaskState[] {
return Object.values(tasks).filter(isInProcessTeammateTask);
}
/**
@@ -147,10 +153,8 @@ export function getAllInProcessTeammateTasks(
* and useBackgroundTaskNavigation — selectedIPAgentIndex maps into this
* array, so all three must agree on sort order.
*/
export function getRunningTeammatesSorted(
tasks: Record<string, TaskStateBase>,
): InProcessTeammateTaskState[] {
export function getRunningTeammatesSorted(tasks: Record<string, TaskStateBase>): InProcessTeammateTaskState[] {
return getAllInProcessTeammateTasks(tasks)
.filter(t => t.status === 'running')
.sort((a, b) => a.identity.agentName.localeCompare(b.identity.agentName))
.sort((a, b) => a.identity.agentName.localeCompare(b.identity.agentName));
}

View File

@@ -1,7 +1,7 @@
import type { TaskStateBase } from '../../Task.js'
import type { AgentToolResult } from '@claude-code-best/builtin-tools/tools/AgentTool/agentToolUtils.js'
import type { AgentDefinition } from '@claude-code-best/builtin-tools/tools/AgentTool/loadAgentsDir.js'
import type { Message } from '../../types/message.js'
import type { Message, MessageOrigin } from '../../types/message.js'
import type { PermissionMode } from '../../utils/permissions/PermissionMode.js'
import type { AgentProgress } from '../LocalAgentTask/LocalAgentTask.js'
@@ -19,6 +19,12 @@ export type TeammateIdentity = {
parentSessionId: string // Leader's session ID
}
export type PendingTeammateUserMessage = {
message: string
autonomyRunId?: string
origin?: MessageOrigin
}
export type InProcessTeammateTaskState = TaskStateBase & {
type: 'in_process_teammate'
@@ -56,7 +62,7 @@ export type InProcessTeammateTaskState = TaskStateBase & {
inProgressToolUseIDs?: Set<string>
// Queue of user messages to deliver when viewing teammate transcript
pendingUserMessages: string[]
pendingUserMessages: PendingTeammateUserMessage[]
// UI: random spinner verbs (stable across re-renders, shared between components)
spinnerVerb?: string

View File

@@ -355,6 +355,19 @@ export type QueuedCommand = {
* unified the queue but lost the isolation the dual-queue accidentally had).
*/
agentId?: AgentId
/**
* Autonomy-run provenance for system-generated automatic turns.
* Used by the autonomy ledger to track queue → execution lifecycle.
*/
autonomy?: {
runId: string
trigger: 'scheduled-task' | 'proactive-tick' | 'managed-flow-step'
sourceId?: string
sourceLabel?: string
flowId?: string
flowStepId?: string
flowStepName?: string
}
}
/**

View File

@@ -0,0 +1,241 @@
import { afterEach, beforeEach, describe, expect, test } from 'bun:test'
import { join } from 'node:path'
import {
AUTONOMY_AGENTS_PATH_POSIX,
AUTONOMY_DIR,
buildAutonomyTurnPrompt,
loadAutonomyAuthority,
resetAutonomyAuthorityForTests,
} from '../autonomyAuthority'
import {
cleanupTempDir,
createTempDir,
createTempSubdir,
writeTempFile,
} from '../../../tests/mocks/file-system'
const AGENTS_REL = join(AUTONOMY_DIR, 'AGENTS.md')
const HEARTBEAT_REL = join(AUTONOMY_DIR, 'HEARTBEAT.md')
let tempDir = ''
beforeEach(async () => {
tempDir = await createTempDir('autonomy-authority-')
})
afterEach(async () => {
resetAutonomyAuthorityForTests()
if (tempDir) {
await cleanupTempDir(tempDir)
}
})
describe('autonomyAuthority', () => {
test('loadAutonomyAuthority merges AGENTS.md files from root to current directory', async () => {
const nestedDir = await createTempSubdir(tempDir, 'packages/app')
await writeTempFile(tempDir, AGENTS_REL, 'root authority')
await writeTempFile(nestedDir, AGENTS_REL, 'nested authority')
await writeTempFile(
tempDir,
HEARTBEAT_REL,
[
'# Heartbeat',
'tasks:',
' - name: inbox',
' interval: 30m',
' prompt: "Check inbox"',
].join('\n'),
)
const snapshot = await loadAutonomyAuthority({
rootDir: tempDir,
currentDir: nestedDir,
})
expect(snapshot.agentsFiles.map(file => file.relativePath)).toEqual([
AUTONOMY_AGENTS_PATH_POSIX,
`packages/app/${AUTONOMY_AGENTS_PATH_POSIX}`,
])
expect(snapshot.agentsContent).toContain('root authority')
expect(snapshot.agentsContent).toContain('nested authority')
expect(snapshot.heartbeatContent).toContain('# Heartbeat')
expect(snapshot.heartbeatTasks).toEqual([
{
name: 'inbox',
interval: '30m',
prompt: 'Check inbox',
steps: [],
},
])
})
test('loadAutonomyAuthority reads HEARTBEAT.md only from the workspace root', async () => {
const nestedDir = await createTempSubdir(tempDir, 'child')
await writeTempFile(
tempDir,
HEARTBEAT_REL,
'# Root heartbeat\nRemember the root task',
)
await writeTempFile(
nestedDir,
HEARTBEAT_REL,
'# Nested heartbeat\nThis should not be used',
)
const snapshot = await loadAutonomyAuthority({
rootDir: tempDir,
currentDir: nestedDir,
})
expect(snapshot.heartbeatFile?.path).toBe(join(tempDir, HEARTBEAT_REL))
expect(snapshot.heartbeatContent).toContain('Root heartbeat')
expect(snapshot.heartbeatContent).not.toContain('Nested heartbeat')
})
test('buildAutonomyTurnPrompt returns the original prompt when no authority files exist', async () => {
const prompt = await buildAutonomyTurnPrompt({
basePrompt: 'Run the scheduled task.',
trigger: 'scheduled-task',
rootDir: tempDir,
currentDir: tempDir,
})
expect(prompt).toBe('Run the scheduled task.')
})
test('buildAutonomyTurnPrompt injects AGENTS.md and HEARTBEAT.md for automated turns', async () => {
const nestedDir = await createTempSubdir(tempDir, 'nested')
await writeTempFile(tempDir, AGENTS_REL, 'root rules')
await writeTempFile(nestedDir, AGENTS_REL, 'nested rules')
await writeTempFile(tempDir, HEARTBEAT_REL, 'Check heartbeat directives')
const scheduledPrompt = await buildAutonomyTurnPrompt({
basePrompt: 'Review the nightly report.',
trigger: 'scheduled-task',
rootDir: tempDir,
currentDir: nestedDir,
})
const tickPrompt = await buildAutonomyTurnPrompt({
basePrompt: '<tick>12:00:00</tick>',
trigger: 'proactive-tick',
rootDir: tempDir,
currentDir: nestedDir,
})
expect(scheduledPrompt).toContain(
'This prompt was generated automatically. Follow the workspace authority below before acting.',
)
expect(scheduledPrompt).toContain('<autonomy_authority>')
expect(scheduledPrompt).toContain('root rules')
expect(scheduledPrompt).toContain('nested rules')
expect(scheduledPrompt).toContain('Check heartbeat directives')
expect(scheduledPrompt).toContain('Review the nightly report.')
expect(tickPrompt).toContain(
'This is an autonomous proactive turn. Follow the workspace authority below before acting.',
)
expect(tickPrompt).toContain('<tick>12:00:00</tick>')
})
test('proactive prompts surface due HEARTBEAT.md tasks only when their interval elapses', async () => {
await writeTempFile(
tempDir,
HEARTBEAT_REL,
[
'tasks:',
' - name: inbox',
' interval: 30m',
' prompt: "Check inbox"',
].join('\n'),
)
const first = await buildAutonomyTurnPrompt({
basePrompt: '<tick>12:00:00</tick>',
trigger: 'proactive-tick',
rootDir: tempDir,
currentDir: tempDir,
nowMs: 0,
})
const second = await buildAutonomyTurnPrompt({
basePrompt: '<tick>12:10:00</tick>',
trigger: 'proactive-tick',
rootDir: tempDir,
currentDir: tempDir,
nowMs: 10 * 60_000,
})
const third = await buildAutonomyTurnPrompt({
basePrompt: '<tick>12:31:00</tick>',
trigger: 'proactive-tick',
rootDir: tempDir,
currentDir: tempDir,
nowMs: 31 * 60_000,
})
expect(first).toContain('Due HEARTBEAT.md tasks:')
expect(first).toContain('- inbox (30m): Check inbox')
expect(second).not.toContain('Due HEARTBEAT.md tasks:')
expect(third).toContain('Due HEARTBEAT.md tasks:')
})
test('managed HEARTBEAT.md tasks parse nested steps and are not duplicated into the inline due-task section', async () => {
await writeTempFile(
tempDir,
HEARTBEAT_REL,
[
'tasks:',
' - name: inbox',
' interval: 30m',
' prompt: "Check inbox"',
' - name: weekly-report',
' interval: 7d',
' prompt: "Ship the weekly report"',
' steps:',
' - name: gather',
' prompt: "Gather weekly inputs"',
' - name: draft',
' prompt: "Draft the weekly report"',
' wait_for: manual',
].join('\n'),
)
const snapshot = await loadAutonomyAuthority({
rootDir: tempDir,
currentDir: tempDir,
})
const prompt = await buildAutonomyTurnPrompt({
basePrompt: '<tick>12:00:00</tick>',
trigger: 'proactive-tick',
rootDir: tempDir,
currentDir: tempDir,
nowMs: 0,
})
expect(snapshot.heartbeatTasks).toEqual([
{
name: 'inbox',
interval: '30m',
prompt: 'Check inbox',
steps: [],
},
{
name: 'weekly-report',
interval: '7d',
prompt: 'Ship the weekly report',
steps: [
{
name: 'gather',
prompt: 'Gather weekly inputs',
},
{
name: 'draft',
prompt: 'Draft the weekly report',
waitFor: 'manual',
},
],
},
])
expect(prompt).toContain('- inbox (30m): Check inbox')
expect(prompt).not.toContain('- weekly-report (7d): Ship the weekly report')
expect(prompt).not.toContain('- gather (')
})
})

File diff suppressed because it is too large Load Diff

Some files were not shown because too many files have changed in this diff Show More