CCPM
Project management skill system for AI agents using GitHub Issues and Git worktrees for parallel task execution.
Turn raw ideas into PRDs, epics, GitHub issues, and code with full audit trails and zero context loss.
CCPM is an Agent Skills-compatible workflow system that brings structured project management to AI coding tools. It breaks software builds down into PRDs, technical epics, and GitHub issues, enabling multiple agents to execute tasks in parallel. Technical leads and developers using AI agents will benefit most from its strict spec-driven framework.
What it is
CCPM turns natural language prompts into structured PRDs, technical epics, and modular task files stored directly in your repository. It synchronizes these tasks with GitHub Issues and manages Git worktrees so multiple AI agents can work on independent sub-tasks concurrently. Instant status reports and standups run through local Bash scripts without burning LLM tokens.
Who it's for
It suits engineering leads, technical project managers, and software developers using AI agents like Claude Code, Cursor, or Factory who need structured tracking across sessions. It is not suitable for non-technical teams who do not use GitHub or terminal-based agent harnesses.
Key features
Spec-Driven PRD Breakdown
Converts natural language ideas into structured PRDs, technical epics, and dependency-aware task files.
Parallel Agent Execution
Runs multiple agents simultaneously on sub-tasks using Git worktrees without context overlap.
GitHub Issues Integration
Syncs epics and tasks directly to GitHub Issues as the single source of truth for project progress.
Token-Free Bash Operations
Executes status updates, standups, and task searches via local shell scripts to save LLM tokens.
Up and running in 3 steps
Our honest verdict
What's great
- Keeps complete project context persistent across agent sessions in local project files and Git.
- Runs status, standup, and search commands via Bash scripts to avoid wasting LLM API tokens.
- Integrates natively with GitHub Issues without requiring proprietary databases or GitHub Projects APIs.
- Supports any harness following the open agentskills.io standard like Claude Code, Factory, and Cursor.
Where it falls short
- Running multiple AI agents concurrently can quickly escalate your underlying LLM API costs.
- Requires local setup of Git, an authenticated GitHub CLI (`gh`), and a compatible agent harness.
- Requires strict adherence to task decomposition; unstructured vibe coding is explicitly discouraged.
Pricing & license
CCPM is completely free and open source under the MIT License. Your only costs come from the LLM API tokens consumed by the underlying AI agents and your local development environment.
Open remote AI-training roles
880 open remote AI-training roles, updated November 15, 2026; most pay $20–$175 an hour. Zekai lists them free; the application goes to the employer’s platform.
Alternatives
These reviewed AI tools do the same job instantly — from our scored directory.
Use it, or have us set it up
FAQ
Which AI agent harnesses are compatible with CCPM?
CCPM follows the agentskills.io standard and works with Claude Code, Factory, Codex, OpenCode, Amp, Cursor, and other compatible harnesses.
How does CCPM prevent agent work conflicts?
It isolates context per sub-task and uses Git worktrees to allow multiple agents to execute concurrent streams on independent branches.
Does CCPM require GitHub Projects?
No, CCPM relies strictly on standard GitHub Issues, labels, milestones, and Git branches, requiring no special Projects API setup.
