How to use the breakdown-epic-pm Epic PRD skill

What is an agent skill and what does this one do?

An agent skill is a folder containing a SKILL.md file with instructions that an AI agent loads when a request matches the skill. This one comes from the github/awesome-copilot repository and is released under the MIT licence.

Once the skill is active, the agent acts as a product manager for a large-scale SaaS platform. It turns your high-level idea into a Markdown document with eight parts: epic name, goal (problem, solution and impact), user personas, high-level user journeys, business requirements (functional and non-functional), success metrics, out of scope, and business value.

• The output is saved as Markdown to /docs/ways-of-work/plan/{epic-name}/epic.md.
• If the idea is vague, the agent is instructed to ask clarifying questions before writing.
• The document is meant to feed a later technical architecture specification.

How do you install the breakdown-epic-pm skill?

Download the ZIP file from this page and unzip it. You can install the skill in one of two places, depending on how you work.

For a personal setup available across your projects, unzip the skill folder into ~/.claude/skills/. For a single project, unzip it into the .claude/skills/ folder inside that project.

If you use the Claude apps rather than a local folder, open Settings, go to Capabilities, then Skills, and upload the skill there.

• Personal: ~/.claude/skills/
• Project only: .claude/skills/ inside your project folder
• Claude apps: Settings → Capabilities → Skills, then upload

How do you trigger the skill with a natural request?

You do not need a special command. Describe the epic in plain language and say you want a product requirements document or PRD. The skill's description covers creating a PRD for a new epic, so requests that mention an epic or a large feature are the most likely to match.

The skill also works best when you include a short description of the target users if you have one. The original template has an optional field for this, so you can mention it in your request.

• Write a PRD for an epic that lets small clinics book and reschedule appointments online.
• Draft an epic product requirements document for adding single sign-on to our admin console. Target users are IT administrators.
• Create an epic PRD for a team-wide export feature, then ask me the questions you need first.

What should you expect in the output?

The agent produces a complete PRD in Markdown with the structure listed above. Expect the goal section to describe the problem in three to five sentences, followed by functional requirements as a bulleted list and non-functional requirements such as performance, security, accessibility and data privacy.

Treat the result as a first draft. The skill does not research facts, so any success metrics or business value estimates it gives are text generated from your prompt and need checking against real data before you rely on them.

• Read the Out of Scope section first to catch scope creep early.
• Replace any generic KPI with a figure your team can actually measure.
• Check that user journeys match how your real users work.

What does the skill need to run?

The skill lists no required tools or dependencies. It contains a single file, SKILL.md, and does not run scripts or use the network. It works purely from the text of your request and the instructions in that file.

Because it has no scripts, there is no code to review beyond the Markdown instructions themselves, which you can open and read before installing.

• Requires: nothing listed
• Runs scripts: no
• Uses network: no
• Files: SKILL.md

What are the limits and tips for better results?

The template is built for large, cross-cutting epics. For a small single-feature task it will feel heavyweight, so a short ticket is a better fit. The skill also does not verify market data, so treat any numbers as placeholders.

To get a stronger document, describe the user problem, who is affected and what success looks like. The more context you give, the fewer clarifying questions the agent needs to ask.

• Give the problem and the people affected, not just a feature name.
• Answer the clarifying questions; vague answers produce vague requirements.
• Use the Markdown output as input for an architecture spec, not as the final word.