Product Requirements Document skill for Claude and Codex
A PRD with the problem and evidence first, requirements written as observable behaviour with edge cases, a firm out-of-scope list, and open questions with…
Install it once and Claude or Codex runs it for you. When your request matches, your AI loads the skill, asks for anything it is missing, and works through the template itself. Prefer to fill it in by hand? See the full prompt template.
Download the skill (.zip) · View SKILL.md
When it is used
Use when a product manager needs a requirements document precise enough that two engineers would build the same thing. Do not use for a one-page pitch or sprint-level user stories. Related topics: Product, Planning, Writing.
What it will ask you for
- What is wrong today (longer free text)
- Evidence (longer free text)
- Who is affected and roughly how many (longer free text)
- What we would build (longer free text)
- User-visible behaviour (longer free text)
- Edge cases and error states (longer free text)
- Explicitly out of scope (longer free text)
- Constraints (longer free text)
- Deadline and what drives it (free text, e.g. "Date and why")
- Success metric and target (free text, e.g. "Metric and number")
- Date to check results (free text, e.g. "Date")
How it works
- Write the problem and evidence section first and mark plainly where the evidence is thin.
- Describe the solution as observable behaviour step by step, done when an engineer could test each line.
- Add edge cases and error states, including what happens to existing data.
- Write a prominent out-of-scope section naming what people will otherwise assume.
- Close with success metric and check date, constraints, and open questions each with an owner and a date.
What you get
A PRD with sections for problem and evidence, solution and behaviour, edge cases, out of scope, constraints, success, and open questions with owners.
What it watches out for
- Describing the solution before the problem, which invites debate about the design instead of the need.
- Leaving edge cases and existing data unaddressed, so engineers invent the answer.
- Setting a success metric with no target or date, so nobody can call it a win or a miss.
How to install
Claude.ai and the Claude desktop app
- Download the skill .zip.
- Open Settings, then Capabilities, and find Skills.
- Choose Upload skill and select the .zip.
- Ask Claude for help with the task. It loads the skill when your request matches.
Claude Code
- Download the skill .zip.
- Unzip it into ~/.claude/skills/ (or .claude/skills/ inside a project).
- Start a new session. Claude Code loads the skill when your request matches.
Codex
- Download the skill .zip.
- Unzip it into ~/.agents/skills/ (or .agents/skills/ inside a project).
- Codex picks it up on its own. Type $ and the skill name to use it, or just ask for the task.
Cursor
- Download the Cursor rule (.mdc).
- Save it in .cursor/rules/ inside your project.
- Cursor's agent applies the rule when your request matches its description.
AGENTS.md (Codex, Cursor, and other coding agents)
- Copy the AGENTS.md section.
- Paste it into AGENTS.md at the root of your repository, creating the file if needed.
- Coding agents that read AGENTS.md follow it when your request fits.