CCG is a command-line workflow engine that coordinates Claude, Codex, Gemini, and other models as specialized collaborators on coding tasks. It is used to analyze requests, choose a strategy, delegate work to model-specific roles, and combine their results. The catalogue entries provide the skills, commands, agents, and plugin that implement this workflow.
Borrowing it
Nothing to install: this file belongs to fengshao1227/ccg-workflow. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/fengshao1227/ccg-workflow/main/CLAUDE.mdgit clone --depth 1 https://github.com/fengshao1227/ccg-workflowWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/fengshao1227/ccg-workflow/claude-md)<a href="https://agentmods.dev/instructions/fengshao1227/ccg-workflow/claude-md"><img src="https://agentmods.dev/badge/instructions/fengshao1227/ccg-workflow/claude-md.svg" alt="Measured on agentmods" height="20"></a>What 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 5.1 | $0.18144 | $0.18144 |
| Opus 5 | $0.09072 | $0.09072 |
| Sonnet 5 | $0.03629 | $0.03629 |
| Haiku 4.5 | $0.01814 | $0.01814 |
Grade B, and why
ccg-workflow CLAUDE.md scanned grade B with 3 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 4d 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.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
- ✨ **Codex CLI 接入 APIMart**：新增 `src/utils/installer-codex-api.ts`，把 APIMart 写成 `~/.codex/config.toml` 的 `[model_providers.apimart]`。`codex-mode install` 静默注册（保持非交互），`init` 选 APIMart 时另问是否接 Codex。**注册与启用分离，启用默认否**——翻 `mo Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
- ✨ **Claude Code 安装**：支持 npm/homebrew/curl/powershell/cmd 多种方式 Runs shell commandslowCapability
Expected in a hook, worth knowing in a rule or an instructions file.
- 🐛 **`ccg doctor` 永远报 `✗ Binary Not found`**（#161，报告人已完整定位）：`dist/cli.mjs` 是 ESM 且无 `createRequire` shim，而 `execSafe()` 里是 CommonJS `require('node:child_process')` → `ReferenceError` 被 `catch` 吞掉 → 返回 `null` → 调用方读作「没装 How it starts
The opening of the file, as written. The whole thing — 831 lines — stays where its author put it; the contents beside it link to each section on GitHub.
skills-v2 (CCG Multi-Model Collaboration System)
根目录 > skills-v2
Last Updated: 2026-09-03 (v3.6.4)
⚠ 本文档主体仍停留在 v2.1.16 架构描述（v3.0 引擎重构后未全量同步）。下方变更记录保留 v3.x 修复轨迹，完整历史见 CHANGELOG.md。
变更记录 (Changelog)
完整变更历史请查看 CHANGELOG.md
2026-09-03 (v3.6.4)
- 🐛 dsh-ccg 让新版 harness 的整个 Web UI 打不开（#162）：插件配置页把
settings.plugin.item从list改声明为keyed（按卡片编辑的 settings 命名空间派发）。插槽注册缺了它那个 kind 要求的选项会抛异常，而异常吃掉的是整条 loader entry —— 不是少个卡片，是设置卡片 + 面板视图 + 团队条一起没，页面只剩Failed to load plugins。修法是一份注册同时带key和id：两种 kind 各读各的、忽略对方的。⚠ 新版 harness 已把全部 settings 命名空间发给浏览器（settings.describe({redactSecrets:true})不再过白名单）——这正是插槽改 keyed 的动机，也意味着第三方插件的卡片终于能被派发到。 - 🐛
ccg doctor永远报✗ Binary Not found（#161，报告人已完整定位）：dist/cli.mjs是 ESM 且无createRequireshim，而execSafe()里是 CommonJSrequire('node:child_process')→ReferenceError被catch吞掉 → 返回null→ 调用方读作「没装」。全平台、重装无用。实际影响 6 处调用（binary / opencode / kimi / grok 检测 +status的版本检查），比 issue 描述的更广。 - 🐛
ccg dsh install无视DSH_HOME（自测发现）：harness 自己就靠这个变量换 home，安装器却写死~/.dsh—— 设了该变量的用户会拿到一次「安装成功」，插件却躺在一个没人启动的 home 里。 - ✅ 3 组回归测试（主包 191 → 197，插件 112 → 116）。其中
esm-only.test.ts扫src/**拦 CommonJS 全局；插件侧client-slots.test.mjs用真实 slots 运行时验证同一份注册在两种 kind 声明下都成立、拆掉任一半都会抛。修复已在独立DSH_HOME+ dsh0.1.1-rc.2下实机复现并验证。
2026-08-29 (v3.6.3)
- 🐛
/ccg:codex-exec被 v3.6.0 自己的性能优化打断（魔尊追问「codex 实施模式没 MCP 怎么搞」查出）：该命令的全部前提是「Claude 不做检索，{{BACKEND_PRIMARY}} 全权执行含 MCP 搜索」，载荷里点名 ace-tool / context7 / grok-search；而 v3.6.0 让 wrapper 默认禁用子代理 MCP（-c mcp_servers={}），全仓库没有任何模板传--with-mcp。结果：子代理被要求用它看不到的工具，且不报错 —— 检索不到就照样往下写。现三处执行者调用加 flag；审核（只读 diff、禁改文件）与修正（按 file:line）不加，不白付启动开销。 - ✅ 3 例回归测试
subagent-mcp.test.ts（188 → 191），删掉 flag 会红（已验证）。⚠ 通用教训：改 wrapper 默认行为前，先 grep 模板里有没有依赖旧默认的载荷 —— 两半分别住在 Go 和 Markdown 里，除了测试没东西连得起来。
2026-08-29 (v3.6.2)
- ✨ CCG 落地 DeepSeek Harness：
dsh-ccg/进仓库并随主包一起发（files加 7 条白名单，不含 test/）。七个角色委派工具各跑各的模型；一个角色挂多模型即成模型群（同一简报、独立作答、答案在对话流里并排渲染，不投票不平均）；ccg_team把角色雇成常驻队友（跨轮次存活、独占文件、撞车直接拒绝、每次雇人先走 harness 原生userQuestions请用户批准）。另有持久所有权（存储域）+ 项目记忆（.ccg/memory.md）+ 四档分诊。全程 provider API，无外部 CLI。插件自身 112 单测。 - ✨ 入口两个：菜单「其他工具」组新增
D. DeepSeek Harness；命令ccg dsh <install|uninstall|list> [--profile x]。 - 🔄 新模块
src/utils/installer-dsh.ts+ 12 单测（总 176 → 188）。⚠ 两个易错点已在代码注释里锁死：必须把插件拷到~/.dsh/ccg/dsh-ccg（指 PACKAGE_ROOT 会在 npx 缓存被清后失效）；profile 要dependencies和dsh.profile.bundles两半才加载，而pnpm add只写前者，且 bundles 只能追加不能插入（顺序即 patch 分层）。 - 📌 不单独发 npm：
publish-dsh-ccg.yml完整可用但只留手动触发，反悔改一行，步骤见下方专章。
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.
- 4d ago Changed · +6 lines · +573 tokens per session f7fca326b2a2
- 8d ago First seen · 825 lines · 17,571 tokens per session scan B 1dafd5340786
ccg-workflow CLAUDE.md is an instructions file published in the GitHub repository fengshao1227/ccg-workflow (5,875 stars, last pushed 5d ago), licensed MIT. It adds 18,144 tokens to every session, about $0.0907 per session on Opus 5. A static security scan graded it B with 3 findings (reads agent configuration directories, makes network calls, runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
next.js AGENTS.md
AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
spec-kit AGENTS.md
AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.