bobo-ai-cli 3.0.4 → 3.0.5

package/knowledge/task-router.md

@@ -1,37 +1,46 @@
-# 任务路由决策树
-
-## 自动判断加载范围
-
-收到任务后，根据关键词决定加载哪些知识：
-
-### 关键词 → 能力域映射
-
-| 关键词 | 加载 | 说明 |
-|--------|------|------|
-| 代码/开发/写/改/修 | rules + engineering | 编码规范 + 工程方法 |
-| 错误/bug/调试/报错 | error-catalog + verification | 错误速查 + 验证 |
-| 搜索/找/查/grep | engineering.搜索策略 | 搜索方法论 |
-| 文件/读/写/编辑 | rules.代码规范 | 文件操作规范 |
-| 测试/验证/检查 | verification | 验证协议 |
-| 规划/计划/任务 | engineering.任务路由 | 任务分解方法 |
-| git/提交/推送 | rules.Git规范 | Git 流程 |
-
-### 复杂度判断
-```
-1-2 步 → 直接干
-3-5 步 → 快速 TodoList → 干
-6+  步 → 三文件模式（plan + notes + deliverable）
-```
-
-### 风险判断
-```
-可逆操作 → 直接干
-不可逆 & 低影响 → 提醒后干
-不可逆 & 高影响 → 确认后干
-```
-
-## 默认行为
-- 所有任务默认加载：system.md + rules.md
-- 中等以上复杂度追加：engineering.md
-- 调试任务追加：error-catalog.md
-- 交付前追加：verification.md
+# 任务路由决策树
+
+## 自动判断加载范围
+
+收到任务后，根据关键词决定加载哪些知识：
+
+### 关键词 → 能力域映射
+
+| 关键词 | 加载 | 说明 |
+|--------|------|------|
+| 代码/开发/写/改/修 | rules + engineering | 编码规范 + 工程方法 |
+| 错误/bug/调试/报错 | error-catalog + verification | 错误速查 + 验证 |
+| 搜索/找/查/grep | engineering.搜索策略 | 搜索方法论 |
+| 文件/读/写/编辑 | rules.代码规范 | 文件操作规范 |
+| 测试/验证/检查 | verification | 验证协议 |
+| 规划/计划/任务 | engineering.任务路由 | 任务分解方法 |
+| git/提交/推送 | rules.Git规范 | Git 流程 |
+
+### 复杂度判断
+```
+1-2 步 → 直接干
+3-5 步 → 快速 TodoList → 干
+6+  步 → 三文件模式（plan + notes + deliverable）
+```
+
+### 风险判断
+```
+可逆操作 → 直接干
+不可逆 & 低影响 → 提醒后干
+不可逆 & 高影响 → 确认后干
+```
+
+### Advisor Escalation 判断
+```
+单轮简单任务 → 不要 advisor
+3-5 步中等任务 → 开头规划 + 结尾审视 (max_uses=2)
+6+  步复杂任务 → 开头 + 卡点 + 结尾 (max_uses=3)
+调试任务     → 每次方向变化时 (max_uses=1-3)
+```
+Advisor 只给方向，不执行工具。详见 `rules/advisor-strategy.md`
+
+## 默认行为
+- 所有任务默认加载：system.md + rules.md
+- 中等以上复杂度追加：engineering.md
+- 调试任务追加：error-catalog.md
+- 交付前追加：verification.md

package/knowledge/verification.md

@@ -1,38 +1,38 @@
-# 验证协议
-
-## 何时必须验证
-
-以下情况完成后**必须**执行验证：
-- ≥3 个文件被编辑
-- 后端 / API 变更
-- 基础设施 / 配置变更
-- 任何"应该能工作"的直觉判断
-
-## 验证方法
-
-### 基础验证（必做）
-1. **编译检查**：代码改完能编译通过
-2. **测试运行**：有测试的跑测试
-3. **手动验证**：没测试的手动执行一遍
-4. **输出检查**：确认输出符合预期
-
-### 对抗性验证（推荐）
-至少一个对抗性探测：
-- **边界值**：空输入、超大输入、特殊字符
-- **幂等性**：同一操作执行两次，结果是否一致
-- **错误恢复**：中途失败能否正确回滚/清理
-- **并发安全**：多个请求同时执行
-
-## 验证输出格式
-
-```
-✅ PASS: [what was verified]
-❌ FAIL: [what failed + error details]
-⚠️ SKIP: [what wasn't verified + reason]
-```
-
-## 反模式（不要做）
-- ❌ 只读代码不运行就说"看起来对"
-- ❌ 假装跑了测试
-- ❌ 成功路径验证通过就收工（要测失败路径）
-- ❌ "应该没问题" — 不是验证
+# 验证协议
+
+## 何时必须验证
+
+以下情况完成后**必须**执行验证：
+- ≥3 个文件被编辑
+- 后端 / API 变更
+- 基础设施 / 配置变更
+- 任何"应该能工作"的直觉判断
+
+## 验证方法
+
+### 基础验证（必做）
+1. **编译检查**：代码改完能编译通过
+2. **测试运行**：有测试的跑测试
+3. **手动验证**：没测试的手动执行一遍
+4. **输出检查**：确认输出符合预期
+
+### 对抗性验证（推荐）
+至少一个对抗性探测：
+- **边界值**：空输入、超大输入、特殊字符
+- **幂等性**：同一操作执行两次，结果是否一致
+- **错误恢复**：中途失败能否正确回滚/清理
+- **并发安全**：多个请求同时执行
+
+## 验证输出格式
+
+```
+✅ PASS: [what was verified]
+❌ FAIL: [what failed + error details]
+⚠️ SKIP: [what wasn't verified + reason]
+```
+
+## 反模式（不要做）
+- ❌ 只读代码不运行就说"看起来对"
+- ❌ 假装跑了测试
+- ❌ 成功路径验证通过就收工（要测失败路径）
+- ❌ "应该没问题" — 不是验证

package/knowledge/worker-prompt-craft.md

@@ -0,0 +1,66 @@
+# Worker Prompt 编写指南
+
+## 核心原则
+
+> Worker 看不到你的对话。每个 prompt 必须是自包含的。
+
+## 必须包含
+
+1. **具体文件/路径/行号** — 不是"那个文件"，而是 `src/auth/validate.ts:42`
+2. **完成标准** — "做完"长什么样
+3. **目的陈述** — 帮 Worker 校准深度
+
+## 目的陈述示例
+- "这个研究用于写 PR 描述 — 聚焦用户可见的变更"
+- "我需要这个来规划实现 — 报告文件路径、行号、类型签名"
+- "这是合并前的快速检查 — 只验证 happy path"
+
+## 反模式
+
+```
+❌ "修复我们讨论的 bug"
+❌ "基于你的发现，实现修复"
+❌ "创建最近变更的 PR"
+❌ "测试好像出了问题，你看看"
+```
+
+## 正确写法
+
+### 实现类
+```
+修复 src/auth/validate.ts:42 的空指针。
+Session 过期时 user 字段为 undefined 但 token 仍在缓存。
+在 user.id 访问前加 null check — null 则返回 401 'Session expired'。
+运行相关测试和类型检查，提交并报告 hash。
+```
+
+### 研究类
+```
+调查 src/auth/ 模块。找到 session 处理和 token 验证
+可能出现空指针的位置。
+报告具体文件路径、行号和涉及的类型。
+不要修改文件。
+```
+
+### 纠正类（Continue）
+```
+你加的 null check 导致两个测试失败 —
+validate.test.ts:58 期望 'Invalid session' 但你改成了 'Session expired'。
+修复断言，提交并报告 hash。
+```
+
+### Git 操作类
+```
+从 main 创建分支 'fix/session-expiry'。
+只 cherry-pick commit abc123。
+Push 并创建 draft PR 指向 main。
+报告 PR URL。
+```
+
+## Checklist
+- [ ] 包含文件路径、行号、错误信息
+- [ ] 声明完成标准
+- [ ] 实现类：运行测试+类型检查，提交并报告 hash
+- [ ] 研究类：报告发现，不要修改文件
+- [ ] Git 操作：分支名、commit hash、draft/ready、reviewer
+- [ ] Continue 纠正：引用 Worker 做了什么，不是你讨论了什么

package/knowledge/workflows/3d-viz.md

@@ -1,47 +1,47 @@
----
-id: 3d-viz
-title: "3D Visualization Layout"
-type: 3d-viz
-triggers: [Three.js, R3F, 3D scene, layout algorithm]
-checklist: [Node distribution balanced, Surface offsets applied, Global layout verified, Local connections clean]
----
-
-# 3D Visualization Layout
-
-## Triggers
-- Three.js
-- R3F
-- 3D scene
-- layout algorithm
-
-## Steps
-### Step 1: Understand data scale
-Analyze node count and category distribution
-- Action: `analyze`
-- Depends on: none
-
-### Step 2: Weight-based sector allocation
-Allocate sector angles by node count (not equal)
-- Action: `allocate`
-- Depends on: analyze
-
-### Step 3: Apply surface offsets
-Offset connection points to geometry surface (sphere=1.0, cube=1.2, torus=1.4)
-- Action: `offset`
-- Depends on: allocate
-
-### Step 4: Handle overflow
-Enable multi-ring spiral if nodes exceed sector capacity
-- Action: `overflow`
-- Depends on: allocate
-
-### Step 5: Visual verification
-Start dev server and screenshot verify layout + connections
-- Action: `verify`
-- Depends on: offset, overflow
-
-## Checklist
-- [ ] Node distribution balanced
-- [ ] Surface offsets applied
-- [ ] Global layout verified
-- [ ] Local connections clean
+---
+id: 3d-viz
+title: "3D Visualization Layout"
+type: 3d-viz
+triggers: [Three.js, R3F, 3D scene, layout algorithm]
+checklist: [Node distribution balanced, Surface offsets applied, Global layout verified, Local connections clean]
+---
+
+# 3D Visualization Layout
+
+## Triggers
+- Three.js
+- R3F
+- 3D scene
+- layout algorithm
+
+## Steps
+### Step 1: Understand data scale
+Analyze node count and category distribution
+- Action: `analyze`
+- Depends on: none
+
+### Step 2: Weight-based sector allocation
+Allocate sector angles by node count (not equal)
+- Action: `allocate`
+- Depends on: analyze
+
+### Step 3: Apply surface offsets
+Offset connection points to geometry surface (sphere=1.0, cube=1.2, torus=1.4)
+- Action: `offset`
+- Depends on: allocate
+
+### Step 4: Handle overflow
+Enable multi-ring spiral if nodes exceed sector capacity
+- Action: `overflow`
+- Depends on: allocate
+
+### Step 5: Visual verification
+Start dev server and screenshot verify layout + connections
+- Action: `verify`
+- Depends on: offset, overflow
+
+## Checklist
+- [ ] Node distribution balanced
+- [ ] Surface offsets applied
+- [ ] Global layout verified
+- [ ] Local connections clean

package/knowledge/workflows/data-pipeline.md

@@ -1,47 +1,47 @@
----
-id: data-pipeline
-title: "Large Data Pipeline"
-type: data-pipeline
-triggers: [batch query, big dataset, frontend rendering, search]
-checklist: [Data batched, Deduplication applied, Lazy loading enabled, Multi-field search supported]
----
-
-# Large Data Pipeline
-
-## Triggers
-- batch query
-- big dataset
-- frontend rendering
-- search
-
-## Steps
-### Step 1: Batch data fetch
-Query in controlled batch sizes
-- Action: `fetch`
-- Depends on: none
-
-### Step 2: Process data
-Dedup + classify + aggregate
-- Action: `process`
-- Depends on: fetch
-
-### Step 3: Optimize frontend
-Lazy load images, video thumbnails with #t=0.1, object-fit: cover
-- Action: `frontend`
-- Depends on: process
-
-### Step 4: Add search
-Support multi-field search (name + ID)
-- Action: `search`
-- Depends on: frontend
-
-### Step 5: Deduplicate display
-Use Set to deduplicate marquee/list display names
-- Action: `dedup`
-- Depends on: process
-
-## Checklist
-- [ ] Data batched
-- [ ] Deduplication applied
-- [ ] Lazy loading enabled
-- [ ] Multi-field search supported
+---
+id: data-pipeline
+title: "Large Data Pipeline"
+type: data-pipeline
+triggers: [batch query, big dataset, frontend rendering, search]
+checklist: [Data batched, Deduplication applied, Lazy loading enabled, Multi-field search supported]
+---
+
+# Large Data Pipeline
+
+## Triggers
+- batch query
+- big dataset
+- frontend rendering
+- search
+
+## Steps
+### Step 1: Batch data fetch
+Query in controlled batch sizes
+- Action: `fetch`
+- Depends on: none
+
+### Step 2: Process data
+Dedup + classify + aggregate
+- Action: `process`
+- Depends on: fetch
+
+### Step 3: Optimize frontend
+Lazy load images, video thumbnails with #t=0.1, object-fit: cover
+- Action: `frontend`
+- Depends on: process
+
+### Step 4: Add search
+Support multi-field search (name + ID)
+- Action: `search`
+- Depends on: frontend
+
+### Step 5: Deduplicate display
+Use Set to deduplicate marquee/list display names
+- Action: `dedup`
+- Depends on: process
+
+## Checklist
+- [ ] Data batched
+- [ ] Deduplication applied
+- [ ] Lazy loading enabled
+- [ ] Multi-field search supported

package/knowledge/workflows/db-migration.md

@@ -1,51 +1,51 @@
----
-id: db-migration
-title: "Database Migration Orchestration"
-type: db-migration
-triggers: [schema change, data migration, multi-table operation]
-checklist: [Dependencies analyzed, Rollback script ready, Integrity verified, Compatible views built]
----
-
-# Database Migration Orchestration
-
-## Triggers
-- schema change
-- data migration
-- multi-table operation
-
-## Steps
-### Step 1: Analyze dependencies
-Map all tables, foreign keys, and dependencies
-- Action: `analyze`
-- Depends on: none
-
-### Step 2: Write migration script
-Create forward migration in single transaction
-- Action: `write-migration`
-- Depends on: analyze
-
-### Step 3: Write rollback script
-Create rollback script for failure recovery
-- Action: `write-rollback`
-- Depends on: analyze
-
-### Step 4: Build compatibility views
-Create backward-compatible views if needed
-- Action: `compat-views`
-- Depends on: write-migration
-
-### Step 5: Test in staging
-Execute migration in test environment
-- Action: `test`
-- Depends on: write-migration, write-rollback
-
-### Step 6: Verify integrity
-Run verification queries for data consistency
-- Action: `verify`
-- Depends on: test
-
-## Checklist
-- [ ] Dependencies analyzed
-- [ ] Rollback script ready
-- [ ] Integrity verified
-- [ ] Compatible views built
+---
+id: db-migration
+title: "Database Migration Orchestration"
+type: db-migration
+triggers: [schema change, data migration, multi-table operation]
+checklist: [Dependencies analyzed, Rollback script ready, Integrity verified, Compatible views built]
+---
+
+# Database Migration Orchestration
+
+## Triggers
+- schema change
+- data migration
+- multi-table operation
+
+## Steps
+### Step 1: Analyze dependencies
+Map all tables, foreign keys, and dependencies
+- Action: `analyze`
+- Depends on: none
+
+### Step 2: Write migration script
+Create forward migration in single transaction
+- Action: `write-migration`
+- Depends on: analyze
+
+### Step 3: Write rollback script
+Create rollback script for failure recovery
+- Action: `write-rollback`
+- Depends on: analyze
+
+### Step 4: Build compatibility views
+Create backward-compatible views if needed
+- Action: `compat-views`
+- Depends on: write-migration
+
+### Step 5: Test in staging
+Execute migration in test environment
+- Action: `test`
+- Depends on: write-migration, write-rollback
+
+### Step 6: Verify integrity
+Run verification queries for data consistency
+- Action: `verify`
+- Depends on: test
+
+## Checklist
+- [ ] Dependencies analyzed
+- [ ] Rollback script ready
+- [ ] Integrity verified
+- [ ] Compatible views built

package/knowledge/workflows/feature-dev.md

@@ -1,41 +1,41 @@
----
-id: feature-dev
-title: "Feature Development Workflow"
-type: feature-dev
-triggers: [new feature, complex feature, multi-file change]
-checklist: [Plan approved, Dependencies identified, Tests written, Code reviewed]
----
-
-# Feature Development Workflow
-
-## Triggers
-- new feature
-- complex feature
-- multi-file change
-
-## Steps
-### Step 1: Structured planning
-Break feature into clear steps with dependencies
-- Action: `plan`
-- Depends on: none
-
-### Step 2: Identify dependencies
-Map which steps depend on which
-- Action: `deps`
-- Depends on: plan
-
-### Step 3: Implement step by step
-Execute in dependency order, verify each step
-- Action: `implement`
-- Depends on: deps
-
-### Step 4: Verify each step
-Test and verify before moving to next step
-- Action: `verify`
-- Depends on: implement
-
-## Checklist
-- [ ] Plan approved
-- [ ] Dependencies identified
-- [ ] Tests written
-- [ ] Code reviewed
+---
+id: feature-dev
+title: "Feature Development Workflow"
+type: feature-dev
+triggers: [new feature, complex feature, multi-file change]
+checklist: [Plan approved, Dependencies identified, Tests written, Code reviewed]
+---
+
+# Feature Development Workflow
+
+## Triggers
+- new feature
+- complex feature
+- multi-file change
+
+## Steps
+### Step 1: Structured planning
+Break feature into clear steps with dependencies
+- Action: `plan`
+- Depends on: none
+
+### Step 2: Identify dependencies
+Map which steps depend on which
+- Action: `deps`
+- Depends on: plan
+
+### Step 3: Implement step by step
+Execute in dependency order, verify each step
+- Action: `implement`
+- Depends on: deps
+
+### Step 4: Verify each step
+Test and verify before moving to next step
+- Action: `verify`
+- Depends on: implement
+
+## Checklist
+- [ ] Plan approved
+- [ ] Dependencies identified
+- [ ] Tests written
+- [ ] Code reviewed