[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-73403":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":16,"subscribersCount":16,"size":16,"stars1d":17,"stars7d":18,"stars30d":19,"stars90d":16,"forks30d":16,"starsTrendScore":20,"compositeScore":21,"rankGlobal":10,"rankLanguage":10,"license":22,"archived":23,"fork":23,"defaultBranch":24,"hasWiki":25,"hasPages":23,"topics":26,"createdAt":10,"pushedAt":10,"updatedAt":33,"readmeContent":34,"aiSummary":35,"trendingCount":16,"starSnapshotCount":16,"syncStatus":36,"lastSyncTime":37,"discoverSource":38},73403,"ccpm","automazeio\u002Fccpm","automazeio","Project management skill system for Agents that uses GitHub Issues and Git worktrees for parallel agent execution.","https:\u002F\u002Fautomaze.io",null,"Shell",8181,831,41,4,0,10,21,91,30,39.76,"MIT License",false,"main",true,[27,28,29,30,31,32],"ai-agents","ai-coding","claude","claude-code","project-management","vibe-coding","2026-06-12 02:03:12","# CCPM – The Project Manager Agent\n\n[![Agent Skills](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FAgent_Skills-compatible-4b3baf)](https:\u002F\u002Fagentskills.io)\n \n[![Eval Score](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Feval_score-100%25-brightgreen)](#proven-results)\n \n[![GitHub Issues](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002F+-GitHub%20Issues-1f2328)](https:\u002F\u002Fgithub.com\u002Fautomazeio\u002Fccpm)\n \n[![MIT License](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-MIT-28a745)](LICENSE)\n \n[![Follow on 𝕏](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002F𝕏-@aroussi-1c9bf0)](http:\u002F\u002Fx.com\u002Fintent\u002Ffollow?screen_name=aroussi)\n \n[![Star this repo](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fautomazeio\u002Fccpm.svg?style=social&label=Star%20this%20repo&maxAge=60)](https:\u002F\u002Fgithub.com\u002Fautomazeio\u002Fccpm)\n\n### Spec-driven development for AI agents – ship ~~faster~~ _better_ using PRDs, GitHub issues, and multiple agents running in parallel.\n\nStop losing context. Stop blocking on tasks. Stop shipping bugs. CCPM gives your AI agent a structured PM brain: turn ideas into PRDs, PRDs into epics, epics into GitHub issues, and issues into production code — with full traceability at every step.\n\n---\n\n> [!IMPORTANT]\n> 📢 **CCPM is now an AGENT SKILL!** It works with any [Agent Skills–compatible](https:\u002F\u002Fagentskills.io) harness that supports skills: **Claude Code, Codex, OpenCode, Factory, Amp, Cursor, and more.**\n\n---\n\n![CCPM](screenshot.webp)\n\n## Table of Contents\n\n- [Background](#background)\n- [The Workflow](#the-workflow)\n- [What Makes This Different](#what-makes-this-different)\n- [Why GitHub Issues](#why-github-issues)\n- [Core Principle: No Vibe Coding](#core-principle-no-vibe-coding)\n- [The Parallel Execution System](#the-parallel-execution-system)\n- [Key Features & Benefits](#key-features--benefits)\n- [Install](#install)\n- [Usage](#usage)\n- [Workflow Phases](#workflow-phases)\n- [Skill Structure](#skill-structure)\n- [Example Flow](#example-flow)\n- [Proven Results](#proven-results)\n- [Local vs Remote](#local-vs-remote)\n- [Technical Notes](#technical-notes)\n- [Who's Behind this Project](#whos-behind-this-project)\n\n---\n\n> [!NOTE]\n> Check out **[proof](https:\u002F\u002Fgithub.com\u002Fautomazeio\u002Fproof)** to get your agents capture visual proof of work of terminal output, browser interactions, and mobile simulator recordings.\n\n---\n\n## Background\n\nEvery team struggles with the same problems:\n- **Context evaporates** between sessions, forcing constant re-discovery\n- **Parallel work creates conflicts** when multiple agents touch the same code\n- **Requirements drift** as verbal decisions override written specs\n- **Progress becomes invisible** until the very end\n\nCCPM solves all of that.\n\n---\n\n## The Workflow\n\n```mermaid\ngraph LR\n    A[PRD Creation] --> B[Epic Planning]\n    B --> C[Task Decomposition]\n    C --> D[GitHub Sync]\n    D --> E[Parallel Execution]\n```\n\n### See It In Action\n\n```\n\"I want to build a notification system — where do we start?\"\n→ Guided brainstorming + PRD creation\n\n\"break down the notification-system epic\"\n→ Parallelizable task files with dependencies\n\n\"sync the notification-system epic to GitHub\"\n→ Epic issue + sub-issues + worktree\n\n\"start working on issue 42\"\n→ Parallel stream analysis + multiple agents launched\n\n\"what's our standup for today?\"\n→ Instant report from project files\n```\n\n---\n\n## What Makes This Different\n\n| Traditional AI Development | CCPM |\n|---|---|\n| Context lost between sessions | **Persistent context** across all work |\n| One agent, one task | **Parallel agents** on independent streams |\n| Vibe coding from memory | **Spec-driven** with full traceability |\n| Progress hidden in chat logs | **Transparent audit trail** in GitHub |\n| Scattered status updates | **Structured standup, blocked, next** |\n\n---\n\n## Why GitHub Issues\n\nMost AI coding workflows operate in isolation — a single session with no shared state. CCPM uses GitHub Issues as the source of truth, which unlocks something fundamentally different:\n\n**Team collaboration** — multiple agents (or humans) work on the same project simultaneously. Progress is visible in real-time through issue comments.\n\n**Seamless handoffs** — an agent can start a task, a human can finish it, or vice versa. No \"what did the AI do?\" meetings.\n\n**Single source of truth** — no separate databases or project management tools. Issue state is project state. Comments are the audit trail.\n\n**Works with what you have** — no dependency on GitHub Projects. Integrates with existing labels, milestones, and PR workflows.\n\n---\n\n## Core Principle: No Vibe Coding\n\n> **Every line of code must trace back to a specification.**\n\nCCPM enforces a strict 5-phase discipline:\n\n1. **🧠 Brainstorm** — think deeper than comfortable\n2. **📝 Document** — write specs that leave nothing to interpretation\n3. **📐 Plan** — architect with explicit technical decisions\n4. **⚡ Execute** — build exactly what was specified\n5. **📊 Track** — maintain transparent progress at every step\n\nNo shortcuts. No assumptions. No regrets.\n\n---\n\n## The Parallel Execution System\n\n### Issues Aren't Atomic\n\nTraditional thinking: **one issue = one agent = one task**\n\nReality: a single \"Implement user authentication\" issue is actually:\n\n- **Agent 1**: Database tables and migrations\n- **Agent 2**: Service layer and business logic\n- **Agent 3**: API endpoints and middleware\n- **Agent 4**: UI components and forms\n- **Agent 5**: Test suites and documentation\n\nAll running **simultaneously** in the same worktree.\n\n### The Math of Velocity\n\n| Approach | Agents working | Wall time |\n|---|---|---|\n| Traditional (serial) | 1 | 5x |\n| CCPM (parallel streams) | 5 | 1x |\n\n### Context Stays Clean\n\nEach agent handles its own context in isolation. Your main conversation becomes the conductor — it never drowns in implementation details. Agents read from `.claude\u002Fepics\u002F` and commit progress back through Git.\n\n---\n\n## Key Features & Benefits\n\n**🧠 Context preservation** — project state lives in files, not in your head or chat history. Start a session anywhere, any time.\n\n**⚡ Parallel execution** — tasks marked `parallel: true` run concurrently across multiple agents without conflicts.\n\n**🔗 GitHub native** — works with tools your team already uses. No dependency on the Projects API.\n\n**📊 Full traceability** — every decision documented. PRD → Epic → Task → Issue → Code → Commit.\n\n**🤖 Deterministic ops run as scripts** — status, standup, search, validate all run as bash scripts: fast, consistent, no LLM token cost.\n\n**🌐 Harness-agnostic** — follows the [agentskills.io](https:\u002F\u002Fagentskills.io) open standard. Works with Factory, Claude Code, Amp, OpenCode, Codex, Cursor, and more.\n\n---\n\n## Install\n\nCCPM is a standard [Agent Skill](https:\u002F\u002Fagentskills.io). Point your harness at `skill\u002Fccpm\u002F` — that's it.\n\n### Clone the repo\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Fautomazeio\u002Fccpm.git\n```\n\n### Factory \u002F Droid\n\n```bash\n# Symlink into your skills directory\nln -s \u002Fpath\u002Fto\u002Fccpm\u002Fskill\u002Fccpm ~\u002F.factory\u002Fskills\u002Fccpm\n```\n\n### Claude Code\n\nIn your project root, add a `skills\u002F` directory and symlink or copy the skill:\n\n```bash\nln -s \u002Fpath\u002Fto\u002Fccpm\u002Fskill\u002Fccpm .claude\u002Fskills\u002Fccpm\n```\n\n### Any other Agent Skills–compatible harness\n\nPoint it at `skill\u002Fccpm\u002F`. It follows the [agentskills.io](https:\u002F\u002Fagentskills.io) standard and works out of the box.\n\n### Prerequisites\n\n- `git` and `gh` CLI (authenticated: `gh auth login`)\n- A GitHub repository for your project\n\n---\n\n## Usage\n\nCCPM activates automatically when your agent detects PM intent. Just talk naturally — no special syntax needed.\n\n### Natural language triggers\n\n| What you say | What happens |\n|---|---|\n| \"I want to build X\" \u002F \"let's plan X\" | Brainstorming + PRD creation |\n| \"parse the X PRD\" \u002F \"create an epic for X\" | PRD → technical epic |\n| \"break down the X epic\" | Epic decomposition into tasks |\n| \"sync the X epic to GitHub\" | Issues created, worktree set up |\n| \"start working on issue N\" | Analysis + parallel agents launched |\n| \"standup\" \u002F \"what's our status\" | Bash script runs instantly |\n| \"what's next\" \u002F \"what's blocked\" | Priority queue from project files |\n| \"close issue N\" | Local + GitHub updated |\n| \"merge the X epic\" | Tests, merge, cleanup |\n\n---\n\n## Workflow Phases\n\n### 1. Plan — Capture requirements\n\n```\n\"I want to build a notification system — push, email, and in-app\"\n```\n\nCCPM conducts guided brainstorming before writing anything. It asks about the problem, users, success criteria, constraints, and what's out of scope — then creates a structured PRD at `.claude\u002Fprds\u002F\u003Cname>.md`.\n\nWhen ready: \"parse the notification-system PRD\" → produces a technical epic at `.claude\u002Fepics\u002Fnotification-system\u002Fepic.md` with architecture decisions, technical approach, and task preview.\n\n### 2. Structure — Break it down\n\n```\n\"break down the notification-system epic into tasks\"\n```\n\nEach task gets a file with acceptance criteria, effort estimate, `depends_on`, `parallel`, and `conflicts_with` metadata. Tasks are intelligently batched for parallel creation. ≤10 tasks per epic by default.\n\n### 3. Sync — Push to GitHub\n\n```\n\"sync the notification-system epic to GitHub\"\n```\n\nCreates an epic issue, creates sub-issues for each task, renames local files to match GitHub issue numbers, sets up a dedicated worktree (`..\u002Fepic-notification-system\u002F`), and creates a mapping file for reference.\n\n### 4. Execute — Start building\n\n```\n\"start working on issue 42\"\n```\n\nAnalyzes the issue for independent work streams, launches parallel agents scoped to their own files, and sets up progress tracking. Each agent commits with `Issue #N: description` and coordinates through Git.\n\n### 5. Track — Know where things stand\n\n```\n\"standup\" \u002F \"what's blocked\" \u002F \"what's next\"\n```\n\nAll tracking operations run as bash scripts — instant output, no LLM overhead. The scripts scan `.claude\u002Fepics\u002F` and report what's in progress, what's next, and what's blocked.\n\n---\n\n## Skill Structure\n\n```\nskill\u002Fccpm\u002F\n├── SKILL.md                  # Entry point — detects intent, routes to reference\n└── references\u002F\n    ├── plan.md               # PRD writing + parsing to epic\n    ├── structure.md          # Epic decomposition into tasks\n    ├── sync.md               # GitHub sync, progress comments, close, merge\n    ├── execute.md            # Issue analysis + parallel agent launch\n    ├── track.md              # Status, standup, search, next, blocked\n    ├── conventions.md        # File formats, frontmatter schemas, git rules\n    └── scripts\u002F              # Bash scripts for deterministic operations\n        ├── status.sh\n        ├── standup.sh\n        ├── epic-list.sh\n        ├── search.sh\n        └── ...               # 14 scripts total\n```\n\nYour project files live in `.claude\u002F` in your project root:\n\n```\n.claude\u002F\n├── prds\u002F                     # Product requirement documents\n├── epics\u002F\n│   └── \u003Cfeature>\u002F\n│       ├── epic.md           # Technical epic\n│       ├── \u003CN>.md            # Task files (named by GitHub issue number after sync)\n│       ├── \u003CN>-analysis.md   # Parallel work stream analysis\n│       └── updates\u002F          # Agent progress tracking\n└── (archived epics)\n```\n\nFiles are the source of truth — plain markdown that lives in your repo, no external services.\n\n---\n\n## Example Flow\n\n```\nYou: \"I want to build a payment integration with Stripe — subscriptions and one-time charges\"\n\nCCPM: Asks 5 clarifying questions about scope, users, success criteria...\n\nYou: [answers]\n\nCCPM: ✅ PRD created: .claude\u002Fprds\u002Fpayment-integration.md\n      Ready to create the technical epic?\n\nYou: \"yes, parse it\"\n\nCCPM: ✅ Epic created: .claude\u002Fepics\u002Fpayment-integration\u002Fepic.md\n      8 task categories identified. Ready to decompose?\n\nYou: \"break it down\"\n\nCCPM: ✅ Created 7 tasks — 5 parallel, 2 sequential\n      Ready to push to GitHub?\n\nYou: \"sync it\"\n\nCCPM: ✅ Epic #1234 created\n      ✅ 7 sub-issues created (#1235–#1241)\n      ✅ Worktree: ..\u002Fepic-payment-integration\u002F\n\nYou: \"start working on issue 1235\"\n\nCCPM: Analyzed 3 parallel streams:\n      Stream A: Stripe client setup ✓ Started\n      Stream B: Webhook handler ✓ Started\n      Stream C: Database models ⏸ Waiting on A\n\nYou: \"what's our standup?\"\n\nCCPM: [runs standup.sh instantly]\n      📅 Daily Standup — 2026-03-18\n      🔄 In Progress: Issue #1235 (payment-integration) — 60%\n      ⏭️ Next: Issue #1236 — Subscription billing logic\n      📊 Tasks: 2 in progress, 5 open, 0 closed\n```\n\n---\n\n## Proven Results\n\nTeams using this system report:\n- **89% less time** lost to context switching\n- **5–8 parallel tasks** vs 1 previously\n- **75% reduction** in bug rates — due to detailed task breakdown before coding\n- **Up to 3× faster** feature delivery\n\n### Benchmark\n\nIn structured evals comparing CCPM-equipped agents vs baseline (no skill):\n\n| Scenario | With CCPM | Without |\n|---|---|---|\n| PRD creation (brainstorm-first, correct paths) | ✅ 4\u002F4 | ❌ 2\u002F4 |\n| Issue execution (analysis + worktree checks) | ✅ 4\u002F4 | ❌ 0\u002F4 |\n| Standup (runs script, real data) | ✅ 3\u002F3 | ❌ 1\u002F3 |\n| **Overall** | **100%** | **27.7%** |\n\n---\n\n## Local vs Remote\n\n| Operation | Local | GitHub |\n|---|---|---|\n| PRD creation | ✅ | — |\n| Implementation planning | ✅ | — |\n| Task breakdown | ✅ | ✅ (on sync) |\n| Execution | ✅ | — |\n| Progress updates | ✅ | ✅ (on sync) |\n| Final deliverables | — | ✅ |\n\n---\n\n## Technical Notes\n\n**GitHub integration** — uses `gh-sub-issue` extension for proper parent-child relationships. Falls back to task lists if not installed. Install with: `gh extension install yahsan2\u002Fgh-sub-issue`\n\n**File naming** — tasks start as `001.md`, `002.md` during decomposition. After GitHub sync, renamed to `{issue-id}.md` (e.g. `1234.md`). Issue #1234 = file `1234.md`.\n\n**Design decisions** — intentionally avoids GitHub Projects API complexity. All operations work on local files first for speed. GitHub sync is explicit and controlled. Worktrees provide clean git isolation for parallel work.\n\n**Looking for v1?** — The original `\u002Fpm:*` Claude Code slash command system is preserved on the [`v1` branch](https:\u002F\u002Fgithub.com\u002Fautomazeio\u002Fccpm\u002Ftree\u002Fv1).\n\n---\n\n## Who's behind this project\n\nCCPM was developed at [Automaze](https:\u002F\u002Fautomaze.io) **for developers who ship, by developers who ship**.\n\nIf CCPM helps your team ship better software:\n\n- ⭐ **[Star this repository](https:\u002F\u002Fgithub.com\u002Fautomazeio\u002Fccpm)** to show your support\n- 🐦 **[Follow @aroussi on X](https:\u002F\u002Fx.com\u002Faroussi)** for updates and tips\n\n---\n\n> [!TIP]\n> **Ship faster with Automaze.** We partner with founders to bring their vision to life, scale their business, and optimize for success.\n> **[Visit Automaze to book a call ›](https:\u002F\u002Fautomaze.io)**\n\n---\n\n![Star History Chart](https:\u002F\u002Fapi.star-history.com\u002Fsvg?repos=automazeio\u002Fccpm)\n","CCPM 是一个面向AI代理的项目管理系统，利用GitHub Issues和Git工作树实现并行代理执行。其核心功能包括将想法转化为产品需求文档（PRD），再分解为史诗级任务，并进一步细化成可追踪的GitHub问题，最终生成生产代码。该系统通过结构化的项目管理方式，确保开发过程中的上下文不丢失、减少任务冲突及需求漂移，同时提高透明度与可追溯性。适用于需要高效管理和协调多个AI代理进行软件开发的场景，特别是在强调规范驱动开发和避免“氛围编码”的团队中尤为适用。",2,"2026-06-11 03:45:24","high_star"]