Skip to content
Get packs

Guide · Agent skill

How to use the CCPM Project Manager agent skill

Updated · written with AI from the project's own documentation and checked against our catalogue data

CCPM is an agent skill that teaches an AI agent a spec-driven workflow for shipping software features: PRD, then technical epic, then GitHub issues, then parallel agents and shipped code. This guide explains what an agent skill is, how to install the ccpm skill, how to trigger it with plain requests, and what it needs before you start. By the end you will be able to ask your agent to plan a feature, check project status, and work from tracked issues.

CCPM Project Managerautomazeio/ccpm · MIT

Quick start

  1. 1Download the ZIP. Get the ccpm ZIP from the skill's page on this directory.
  2. 2Unzip the skill. Unzip it into ~/.claude/skills/ for all projects, or into a project's .claude/skills/ folder for one project.
  3. 3Or upload in Claude apps. In the Claude apps, open Settings, go to Capabilities, then Skills, and upload the ZIP.
  4. 4Confirm requirements. Make sure you have a bash shell, a Git repository, and a GitHub account and repository if you plan to sync issues.
  5. 5Plan a feature. Ask the agent to write a PRD for a feature, answering its brainstorming questions.
  6. 6Turn the PRD into an epic. Ask the agent to turn the PRD into an epic, then break the epic into tasks.
  7. 7Check status. Ask 'what's next' or 'standup' to run the status scripts against your local epic files.

What is an agent skill and what does ccpm do?

An agent skill is a folder of instructions and optional scripts that an AI agent loads when a request matches the skill's description. The ccpm skill is described as a spec-driven project manager, and its name in the source is CCPM, short for Claude Code Project Manager.

The skill moves a feature through five phases: Plan (write a PRD through guided brainstorming and convert it to an epic), Structure (break the epic into numbered task files), Sync (push epics and tasks to GitHub issues), Execute (analyse parallel work streams and launch agents on issues), and Track (status, standup, next and blocked items).

  • Writes requirement documents to .claude/prds/<name>.md
  • Turns a PRD into a technical epic and decomposes it into tasks
  • Syncs epics and tasks to GitHub issues and posts progress comments
  • Runs read-only status scripts that report on local epic files under .claude/epics

Install the ccpm skill with a ZIP file

Download the ccpm ZIP from the skill's page on this directory. Then choose one of two install locations, depending on whether you want the skill available everywhere or only in one project.

For a personal install across all projects, unzip the file into ~/.claude/skills/. For a single project, unzip it into the project's .claude/skills/ folder. Claude apps also accept an upload: open Settings, go to Capabilities, then Skills, and upload the ZIP there.

  • Personal (all projects): unzip into ~/.claude/skills/
  • Single project: unzip into .claude/skills/ inside the project
  • Claude apps: Settings → Capabilities → Skills, then upload the ZIP

Check what you need before you start

The skill expects a bash shell and a Git repository. The sync and issue steps depend on GitHub, so you need a GitHub account and a repository for issue sync. The source says the agent needs GitHub CLI or equivalent credentials to create and update issues; this is inferred from the sync workflow, so check references/conventions.md for the exact requirements.

The skill runs bash scripts and uses the network, so it is not a purely offline tool once you reach the GitHub steps.

  • A bash shell
  • A Git repository
  • A GitHub account and repository for issue sync
  • GitHub CLI or equivalent credentials (confirm in references/conventions.md)

Trigger the skill with natural requests

You do not need to type the skill name. The description lists phrases that should activate it, and it also applies when you talk about shipping a feature or tracking progress even if you never say 'ccpm' or 'PRD'. Describe what you want in plain language.

The quick reference in the skill maps common requests to actions. Examples include planning a feature, turning a PRD into an epic, pushing an epic to GitHub, starting an issue, checking status, and merging an epic.

  • "I want to build X" or "create a PRD for X" to plan a feature
  • "turn the X PRD into an epic" to create a technical epic
  • "break down the X epic into tasks" to decompose it
  • "push the X epic to GitHub" to sync issues
  • "start working on issue 42" to begin execution
  • "what should I work on next" or "standup" for status
  • "merge the X epic" to finish an epic

Use the status scripts

For reporting questions, the skill tells the agent to run the bash script directly rather than working it out by hand. These scripts read local files and report back. The skill's table lists scripts such as epic-list.sh, epic-show.sh, epic-status.sh, prd-list.sh, search.sh, in-progress.sh, next.sh and blocked.sh.

You can also ask the agent to run them. For example, to list all epics, the skill uses the epic-list script.

  • Epic list: references/scripts/epic-list.sh
  • Epic details: references/scripts/epic-show.sh <name>
  • Next work: references/scripts/next.sh
  • Blocked work: references/scripts/blocked.sh
bash references/scripts/epic-list.sh

Safety notes on scripts and network use

The skill runs scripts and uses the network. The status scripts described in the source are read-only and work from local epic files, but the GitHub sync, issue-creation and closing steps change state on GitHub. Review what the agent proposes before it creates, updates or closes issues in your repository.

The skill says it should not be used for debugging code, writing tests, reviewing pull requests, or raw GitHub issue and PR operations with no delivery context. If you see it triggered for those tasks, rephrase the request or ask the agent to proceed without the skill.

  • Status scripts are described as read-only and reading local files
  • Sync, issue updates and closing do change GitHub, so check them first
  • Use a test repository if you are unsure how the sync behaves

Tips for getting good results

The skill's core idea is that requirements live in files rather than in people's heads. Answer the brainstorming questions about the problem, users, success criteria, out-of-scope items and constraints, because the PRD is only as good as those answers. Feature names must be kebab-case, such as user-auth or payment-v2.

Use the status scripts for quick checks and the agent for reasoning tasks such as writing PRDs and analysing parallel work.

  • Use kebab-case feature names such as user-auth or payment-v2
  • Answer the PRD questions fully; the agent asks about scope and constraints
  • Keep the repository's .claude folder under version control if your team relies on it
  • Use the status scripts for quick reports and the agent for planning

What people use it for

  • Writing a PRD for a new feature through guided questions, saved as a file under .claude/prds/
  • Turning a PRD into a technical epic and splitting it into numbered task files
  • Pushing an epic and its tasks to GitHub issues so each task is tracked
  • Asking what is in progress, next or blocked across a small team's epics
  • Starting work on an issue and analysing which parallel work streams can run at once

Frequently asked questions

How do I install the ccpm skill?

Download the ccpm ZIP from the skill's page on this directory and unzip it into ~/.claude/skills/ for all projects, or into a project's .claude/skills/ folder. In the Claude apps you can instead upload the ZIP under Settings, Capabilities, Skills.

What do I need to use ccpm?

You need a bash shell and a Git repository. For issue sync you also need a GitHub account and repository, plus GitHub CLI or equivalent credentials, which the source infers from its sync workflow and says to check in references/conventions.md.

Is ccpm free to use?

The skill is published under the MIT licence. The source does not list prices, so check the licence file and the project's repository for terms.

Does ccpm work without GitHub?

Not fully. The planning and local status features work on files, but the sync and issue steps depend on GitHub, so users without GitHub or a Git repository will not get the full workflow.

Why did the agent trigger ccpm for a code debugging task?

The skill's description says it should not be used for debugging code, writing tests, reviewing pull requests, or raw GitHub operations. Vague wording can cause a wrong trigger, so make the request specific about the delivery task.

How does ccpm compare with other project management tools?

The source does not compare ccpm with other tools. Its distinctive approach is keeping PRDs, epics and task files in a local .claude folder and syncing them to GitHub issues, so choose it if that workflow suits you.

New agent skills, repos and AI tools — every week

The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.

Our verdict

ccpm suits developers who plan features in files, want a tracked GitHub issue for each task, and run parallel work streams on one repository. It is a poor fit for debugging, writing tests, reviewing pull requests, or teams without GitHub, which the skill itself excludes.

Download the skill