CCW command help system, providing command search, recommendations, documentation viewing, and Skill/Team browsing features.
Trigger Conditions
- Keywords: "ccw-help", "ccw-issue", "help", "command", "how to use", "how to use ccw", "workflow", "skill", "team"
- Scenarios: asking about command usage, searching commands, requesting next-step suggestions, asking which workflow to use for a task, browsing the Skill/Team directory
Operation Modes
Mode 1: Command Search
Triggers: "search commands", "find command", "search"
Process:
- Query
command.jsoncommands array - Filter by name, description, category
- Present top 3-5 relevant commands
Mode 2: Smart Recommendations
Triggers: "next step", "what's next", "recommend"
Process:
- Query command's
flow.next_stepsincommand.json - Explain WHY each recommendation fits
Mode 3: Documentation
Triggers: "how to use", "how to use", "details"
Process:
- Locate command in
command.json - Read source file via
sourcepath - Provide context-specific examples
Mode 4: Beginner Onboarding
Triggers: "beginner", "getting started", "common commands"
Process:
- Query
essential_commandsarray - Guide appropriate workflow entry point
Mode 5: CCW Command Orchestration
Triggers: "ccw ", "automatic workflow", "auto-select workflow", "help me plan"
Process:
- Analyze user intent (task type, complexity, clarity)
- Auto-select workflow level (1-4 or Issue)
- Build command chain based on workflow
- Get user confirmation
- Execute chain with TODO tracking
Supported Workflows (reference ccw.md):
- 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
- Multi-CLI collaborative workflows
- Cycle workflows (integration-test, refactor)
- Level 4 (Full): Exploratory tasks with brainstorming
- With-File Workflows: Documented exploration with multi-CLI collaboration
brainstorm-with-file: Multi-perspective ideation → workflow-plan → workflow-executedebug-with-file: Hypothesis-driven debugging (standalone)analyze-with-file: Collaborative analysis → workflow-lite-plancollaborative-plan-with-file: Multi-agent planning → unified-executeroadmap-with-file: Strategic requirement roadmap → team-planex
- Issue Workflow: Batch issue discovery, planning, queueing, execution
- Team Workflow: team-planex wave pipeline for parallel execution
Mode 6: Issue Reporting
Triggers: "ccw-issue", "report bug"
Process:
- Use AskUserQuestion to gather context
- Generate structured issue template
Mode 7: Skill & Team Browsing
Triggers: "skill", "team", "skill", "team", "what skills are there", "how to use team"
Process:
- Query
command.jsonskills array - Filter by category: workflow / team / review / meta / utility / standalone
- Present categorized skill list with descriptions
- For team skills, explain team architecture and usage patterns
Data Source
Single source of truth: command.json
| Field | Purpose |
|---|---|
commands[] |
Flat command list with metadata |
commands[].flow |
Relationships (next_steps, prerequisites) |
agents[] |
Agent directory |
skills[] |
Skill directory with categories |
skills[].is_team |
Whether skill uses team architecture |
essential_commands[] |
Core commands list |
Source Path Format
source The field is a relative path (from the skills/ccw-help/ directory):
{
"name": "lite-plan",
"source": "../../../commands/workflow/lite-plan.md"
}
Skill Catalog
Workflow Skills (Core Workflows)
| Skill | Internal pipeline | Trigger words |
|---|---|---|
workflow-lite-plan |
explore → plan → confirm → execute | "lite-plan", quick tasks |
workflow-plan |
session → context → convention → gen → verify | "workflow-plan", formal planning |
workflow-execute |
session discovery → task processing → commit | "workflow-execute", execution |
workflow-tdd-plan |
6-phase TDD plan → verify | "tdd-plan", TDD development |
workflow-test-fix |
session → context → analysis → gen → cycle | "test-fix", test fix |
workflow-multi-cli-plan |
ACE context → CLI discussion → plan → execute | "multi-cli", multi-CLI collaboration |
workflow-skill-designer |
Meta-skill for designing workflow skills | "skill-designer" |
Team Skills (Team Collaboration)
Team Skills use team-worker agent architecture, Coordinator orchestrates the pipeline, Workers are team-worker agents。
| Skill | Purpose | Architecture |
|---|---|---|
team-planex |
Planning + execution wave pipeline | planner + executor, suitable for clear issue/roadmap |
team-lifecycle |
Full lifecycle (spec/impl/test) | team-worker agents with role-specs |
team-lifecycle-v4 |
Optimized lifecycle | Optimized pipeline |
team-lifecycle-v3 |
Basic lifecycle | All roles invoke unified skill |
team-coordinate |
General dynamic team coordination | Dynamically generate role-specs at runtime |
team-coordinate |
General team coordination v1 | Dynamic role generation |
team-brainstorm |
Team brainstorming | Multi-perspective analysis |
team-frontend |
Frontend development team | Frontend specialists |
team-issue |
Issue resolution team | Issue resolution pipeline |
team-iterdev |
Iterative development team | Iterative development |
team-review |
Code scanning/vulnerability review | Scanning + vulnerability review |
team-roadmap-dev |
Roadmap-driven development | Requirement → implementation |
team-tech-debt |
Technical debt cleanup | Debt identification + cleanup |
team-testing |
Testing team | Test planning + execution |
team-quality-assurance |
QA team | Quality assurance pipeline |
team-uidesign |
UI design team | Design system + prototyping |
team-ultra-analyze |
In-depth collaborative analysis | Deep collaborative analysis |
team-executor |
Lightweight execution (resume session) | Resume existing sessions |
team-executor |
Lightweight execution v2 | Improved session resumption |
Standalone Skills
| Skill | Purpose |
|---|---|
brainstorm |
Dual-mode brainstorming (auto pipeline / single role) |
review-code |
Multi-dimensional code review |
review-cycle |
Review + automated fix orchestration |
spec-generator |
6-stage specification document chain (product-brief → PRD → architecture → epics) |
issue-manage |
Interactive Issue management (CRUD) |
memory-capture |
Unified memory capture (session compact / quick tip) |
memory-manage |
Unified memory management (CLAUDE.md + documentation) |
command-generator |
Command file generator |
skill-generator |
Meta-skill: create a new Skill |
skill-tuning |
Skill diagnosis and optimization |
Workflow Mapping (CCW Auto-Route)
CCW automatically selects the workflow level based on task intent (reference ccw.md):
| Input example | Type | Level | Pipeline |
|---|---|---|---|
| "Add API endpoint" | feature (low) | 2 | workflow-lite-plan → workflow-test-fix |
| "Fix login timeout" | bugfix | 2 | workflow-lite-plan → workflow-test-fix |
| "Collaborative analysis: Authentication architecture" | analyze-file | 3 | analyze-with-file → workflow-lite-plan |
| "Refactor auth module" | refactor | 3 | workflow:refactor-cycle |
| "multi-cli: API design" | multi-cli | 3 | workflow-multi-cli-plan → workflow-test-fix |
| "Brainstorm: Notification system" | brainstorm | 4 | brainstorm-with-file → workflow-plan → workflow-execute |
| "roadmap: OAuth + 2FA" | roadmap | 4 | roadmap-with-file → team-planex |
| "specification: User system" | spec-driven | 4 | spec-generator → workflow-plan → workflow-execute |
| "team planex: User system" | team-planex | Team | team-planex |
Slash Commands
/ccw "task description" # Auto-select workflow and execute
/ccw-help # General help entry
/ccw-help search <keyword> # Search commands
/ccw-help next <command> # Get next step suggestions
/ccw-help skills # Browse skill catalog
/ccw-help teams # Browse team skills
/ccw-issue # Issue reporting
CCW Command Examples
/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
/ccw "roadmap: OAuth + 2FA 路线图" # → roadmap-with-file → team-planex
/ccw "集成测试: 支付流程" # → integration-test-cycle
/ccw "重构 auth 模块" # → refactor-cycle
Maintenance
Update Mechanism
CCW-Help skill supports manual updates through user confirmation dialog.
Script scans commands/, agents/, and skills/ directories to regenerate all indexes.
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
cd D:/Claude_dms3/.claude/skills/ccw-help
python scripts/auto-update.py
This runs analyze_commands.py to scan commands/, agents/, and skills/ directories and regenerate command.json + all index files.
Update Scripts
auto-update.py: Simple wrapper that runs analyze_commands.pyanalyze_commands.py: Scans directories and generates command/agent/skill indexes
Generated Index Files
| File | Content |
|---|---|
command.json |
Master index: commands + agents + skills |
index/all-commands.json |
Flat command list |
index/all-agents.json |
Agent directory |
index/all-skills.json |
Skill directory with metadata |
index/skills-by-category.json |
Skills grouped by category |
index/by-category.json |
Commands by category |
index/by-use-case.json |
Commands by usage scenario |
index/essential-commands.json |
Core commands for onboarding |
index/command-relationships.json |
Command flow relationships |
Statistics
- Commands: 50+
- Agents: 22
- Skills: 36+ (7 workflow, 19 team, 10+ standalone/utility)
- Workflows: 6 main levels + 5 with-file variants + 2 cycle variants
- Essential: 10 core commands
Core Principle
Intelligent synthesis, not template copying
- Understand the user's specific situation
- Integrate information from multiple sources
- Customize examples and explanations