--- name: ccw-help description: CCW command help system. Search, browse, recommend commands. Triggers "ccw-help", "ccw-issue". allowed-tools: Read, Grep, Glob, AskUserQuestion version: 7.0.0 --- # CCW-Help Skill CCW 命令帮助系统,提供命令搜索、推荐、文档查看功能。 ## Trigger Conditions - 关键词: "ccw-help", "ccw-issue", "帮助", "命令", "怎么用", "ccw 怎么用", "工作流" - 场景: 询问命令用法、搜索命令、请求下一步建议、询问任务应该用哪个工作流 ## Operation Modes ### Mode 1: Command Search **Triggers**: "搜索命令", "find command", "search" **Process**: 1. Query `command.json` commands array 2. Filter by name, description, category 3. Present top 3-5 relevant commands ### Mode 2: Smart Recommendations **Triggers**: "下一步", "what's next", "推荐" **Process**: 1. Query command's `flow.next_steps` in `command.json` 2. Explain WHY each recommendation fits ### Mode 3: Documentation **Triggers**: "怎么用", "how to use", "详情" **Process**: 1. Locate command in `command.json` 2. Read source file via `source` path 3. Provide context-specific examples ### Mode 4: Beginner Onboarding **Triggers**: "新手", "getting started", "常用命令" **Process**: 1. Query `essential_commands` array 2. Guide appropriate workflow entry point ### Mode 5: CCW Command Orchestration **Triggers**: "ccw ", "自动工作流", "自动选择工作流", "帮我规划" **Process**: 1. Analyze user intent (task type, complexity, clarity) 2. Auto-select workflow level (1-4 or Issue) 3. Build command chain based on workflow 4. Get user confirmation 5. Execute chain with TODO tracking **Supported Workflows**: - **Level 1** (Lite-Lite-Lite): Ultra-simple quick tasks - **Level 2** (Rapid/Hotfix): Bug fixes, simple features, documentation - **Level 2.5** (Rapid-to-Issue): Bridge from quick planning to issue workflow - **Level 3** (Coupled): Complex features with planning, execution, review, tests - **Level 3 Variants**: - TDD workflows (test-first development) - Test-fix workflows (debug failing tests) - Review workflows (code review and fixes) - UI design workflows - **Level 4** (Full): Exploratory tasks with brainstorming - **With-File Workflows**: Documented exploration with multi-CLI collaboration - `brainstorm-with-file`: Multi-perspective ideation - `debug-with-file`: Hypothesis-driven debugging - `analyze-with-file`: Collaborative analysis - **Issue Workflow**: Batch issue discovery, planning, queueing, execution ### Mode 6: Issue Reporting **Triggers**: "ccw-issue", "报告 bug" **Process**: 1. Use AskUserQuestion to gather context 2. Generate structured issue template ## Data Source Single source of truth: **[command.json](command.json)** | Field | Purpose | |-------|---------| | `commands[]` | Flat command list with metadata | | `commands[].flow` | Relationships (next_steps, prerequisites) | | `commands[].essential` | Essential flag for onboarding | | `agents[]` | Agent directory | | `essential_commands[]` | Core commands list | ### Source Path Format `source` 字段是相对路径(从 `skills/ccw-help/` 目录): ```json { "name": "lite-plan", "source": "../../../commands/workflow/lite-plan.md" } ``` ## Slash Commands ```bash /ccw "task description" # Auto-select workflow and execute /ccw-help # General help entry /ccw-help search # Search commands /ccw-help next # Get next step suggestions /ccw-issue # Issue reporting ``` ### CCW Command Examples ```bash /ccw "Add user authentication" # → auto-select level 2-3 /ccw "Fix memory leak in WebSocket" # → auto-select bugfix workflow /ccw "Implement with TDD" # → detect TDD, use tdd-plan → execute → tdd-verify /ccw "头脑风暴: 用户通知系统" # → detect brainstorm, use brainstorm-with-file /ccw "深度调试: 系统随机崩溃" # → detect debug-file, use debug-with-file /ccw "协作分析: 认证架构设计" # → detect analyze-file, use analyze-with-file ``` ## Maintenance ### Update Mechanism CCW-Help skill supports manual updates through user confirmation dialog. #### How to Update **Option 1: When executing the skill, user will be prompted:** ``` Would you like to update CCW-Help command index? - Yes: Run auto-update and regenerate command.json - No: Use current index ``` **Option 2: Manual update** ```bash cd D:/Claude_dms3/.claude/skills/ccw-help python scripts/auto-update.py ``` This runs `analyze_commands.py` to scan commands/ and agents/ directories and regenerate `command.json`. #### Update Scripts - **`auto-update.py`**: Simple wrapper that runs analyze_commands.py - **`analyze_commands.py`**: Scans directories and generates command index ## Statistics - **Commands**: 50+ - **Agents**: 16 - **Workflows**: 6 main levels + 3 with-file variants - **Essential**: 10 core commands ## Core Principle **智能整合,非模板复制** - 理解用户具体情况 - 整合多个来源信息 - 定制示例和说明