Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add commands/catlog22/claude-code-workflow/ccwgit clone --depth 1 https://github.com/catlog22/Claude-Code-WorkflowWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00020 | $0.10289 |
| Opus 5 | $0.00010 | $0.05144 |
| Sonnet 5 | $0.00004 | $0.02058 |
| Haiku 4.5 | $0.00002 | $0.01029 |
Grade A, and why
ccw scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 963 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CCW Command - Main Workflow Orchestrator
Main process orchestrator: intent analysis → workflow selection → command chain execution.
Skill 映射
命令链中的 workflow 操作通过 Skill() 调用。每个 Skill 是自包含的执行单元，内部处理完整流水线。
| Skill | 内部流水线 |
|---|---|
workflow-lite-plan |
explore → plan → confirm → handoff |
workflow-lite-execute |
task grouping → batch execution → code review → sync |
workflow-plan |
session → context → convention → gen → verify/replan |
workflow-execute |
session discovery → task processing → commit |
workflow-tdd-plan |
6-phase TDD plan → verify |
workflow-test-fix |
session → context → analysis → gen → cycle |
workflow-multi-cli-plan |
ACE context → CLI discussion → plan → execute |
review-cycle |
session/module review → fix orchestration |
brainstorm |
auto/single-role → artifacts → analysis → synthesis |
spec-generator |
product-brief → PRD → architecture → epics |
workflow:collaborative-plan-with-file |
understanding agent → parallel agents → plan-note.md |
workflow:roadmap-with-file |
strategic requirement roadmap → issue creation → execution-plan.json |
workflow:integration-test-cycle |
explore → test dev → test-fix cycle → reflection |
workflow:refactor-cycle |
tech debt discovery → prioritize → execute → validate |
team-planex |
planner + executor wave pipeline（适合大量零散 issue 或 roadmap 产出的清晰 issue） |
Core Concept: Self-Contained Skills (自包含 Skill)
Definition: 每个 Skill 内部处理完整流水线，是天然的最小执行单元。单次 Skill 调用即完成一个有意义的工作里程碑。
Why This Matters:
- Prevents Incomplete States: 每个 Skill 内部保证端到端完整性
- User Experience: User gets complete results, not intermediate artifacts requiring manual follow-up
- Simplified Orchestration: 命令链只需组合独立 Skill，无需关注内部步骤
Key Units in CCW:
| 单元类型 | Skill | 说明 |
|---|---|---|
| 轻量 Plan+Execute | workflow-lite-plan → workflow-lite-execute |
plan handoff 到 execute，分离 Skill，TodoWrite 跟踪延续 (LP-Phase → LE-Phase) |
| 标准 Planning | workflow-plan → workflow-execute |
plan 和 execute 是独立 Skill |
| TDD Planning | workflow-tdd-plan → workflow-execute |
tdd-plan 和 execute 是独立 Skill |
| 规格驱动 | spec-generator → workflow-plan → workflow-execute |
规格文档驱动完整开发 |
| 测试流水线 | workflow-test-fix |
内部完成 gen→cycle |
| 代码审查 | review-cycle |
内部完成 review→fix |
| 多CLI协作 | workflow-multi-cli-plan |
ACE context → CLI discussion → plan → Skill(lite-execute) |
| 分析→规划 | workflow:analyze-with-file → workflow-lite-plan → workflow-lite-execute |
协作分析产物自动传递给 lite-plan，Skill 调用 lite-execute |
| 头脑风暴→规划 | workflow:brainstorm-with-file → workflow-plan → workflow-execute |
头脑风暴产物自动传递给正式规划 |
| 0→1 开发(小) | workflow:brainstorm-with-file → workflow-plan → workflow-execute |
小规模从零开始，探索+正式规划+实现 |
| 0→1 开发(中/大) | workflow:brainstorm-with-file → workflow-plan → workflow-execute |
探索后正式规划+执行 |
| 协作规划 | workflow:collaborative-plan-with-file → workflow:unified-execute-with-file |
多 agent 协作规划→通用执行 |
| 需求路线图 | workflow:roadmap-with-file → team-planex |
需求拆解→issue 创建→wave pipeline 执行（需明确 roadmap 关键词） |
| 集成测试循环 | workflow:integration-test-cycle |
自迭代集成测试闭环 |
| 重构循环 | workflow:refactor-cycle |
技术债务发现→重构→验证 |
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 963 lines · 20 tokens per session scan A b91f46b05a74
ccw is a command published in the GitHub repository catlog22/Claude-Code-Workflow (2,135 stars, last pushed 2mo ago), licensed MIT. It adds 20 tokens to every session and 10,289 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other commands, from other repositories
verify-claim
Walk a claim through the SIFT method (Stop, Investigate, Find better coverage, Trace).
beat-brief
Draft a daily beat briefing from the files in sample-docs/.
cross-review
Run GitHub Copilot CLI and OpenAI Codex against the current git diff for cross-model review.
step-research
Always research before proposing a fix. The Untether bug you're chasing is often a known upstream engine quirk, a previously-fixed regression, or a documented config gotcha.
whats-next
Show current project status and suggest next steps.
ox-session-pause
belongs in the ox CLI JSON output (guidance field), not here. Skills are agent-specific wrappers; ox serves all agents (Codex, etc.). --> Suspend the current session recording. Local cache continues to receive entries, but the upload at stop time will exclude the suspended range.