PRP (Product Requirement Prompts) skill
PRPs-agentic-eng
This skill set assembles a Product Requirement Prompt: the goal, file paths, code examples and validation commands for an agent.
- 2,267 ★ on the repository
- Last commit
- MIT
Minimal access
Why this level
- The skills only assemble a text brief from codebase data, with no code execution and no access to secrets.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/prps-agentic-engDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The skills only assemble a text brief from codebase data, with no code execution and no access to secrets.
This is third-party code. Review the repository files before installing.
What it does
PRP extends an ordinary product requirement document into a format a coding agent can use directly: the goal and justification from a classic PRD stay, but exact file paths, library versions, code examples drawn from the project itself, the codebase's accepted conventions and executable commands the agent runs to verify its own work are added. The idea is a minimal packet sufficient to ship working code on the first pass, with no missing and no excess context. The project publishes a set of ready skills (`prp-core` and others in the marketplace) for planning, codebase questions and generating the PRP itself, installed as a Claude Code plugin or copied individually from the `.claude/skills/` folders.
Who it is for. Development teams on Claude Code who want to hand the agent a brief with codebase context and acceptance checks, not bare requirement text.
Good fit when
- The agent gets a vertical-slice feature and needs to ship working code on the first pass.
- Exact file paths and code examples from the project are needed instead of general task descriptions.
- The agent's work should be checked with executable commands, not just described in words.
- A team wants one agent-brief format, updated centrally through a plugin.
Not a fit when
- The task is small and does not need a dedicated brief; an ordinary agent prompt is enough.
- The codebase has not settled yet, and there are no established patterns to draw examples from.
- A product-planning tool for people is needed, not a brief format specifically for a coding agent.
Example request
Build a PRP for this feature: name the files it touches, a matching pattern from the project and the commands to verify the result.Limitations
The PRP format is built for coding agents with codebase access: without real file paths and project-drawn code examples, its value drops to that of an ordinary requirement document. The skill and template set is maintained by the author as a personal open-source project, and centralized plugin updates depend on their marketplace releases. Part of the README covers the author's paid PRP-methodology workshops, which are not required to use the skill set itself. Teams without Claude Code can still copy individual `.claude/skills/` folders, but without centralized plugin updates.
How to disable. Remove the `prp-core` plugin with the reverse Claude Code command, or delete the copied skill folders from `.claude/skills/`.
FAQ
How does PRP differ from an ordinary requirement document?
An ordinary document explains what and why, deliberately avoiding how. PRP adds exact file paths, library versions, project-drawn code examples and validation commands on top.
How do I install the whole skill set?
Run `/plugin marketplace add Wirasm/PRPs-agentic-eng`, then `/plugin install prp-core@prp-marketplace`; skills then update centrally.
Can I take just one skill without the plugin?
Yes, copy the folder you need, such as `.claude/skills/prp-plan`, into your own project's `.claude/skills/`.
Are the author's paid workshops required to use PRP?
No, the workshops are a separate commercial offer from the author; the skill set in the repository works without them.
How does the agent verify its own work against a PRP?
Through the executable commands listed directly in the PRP: the agent runs them itself and checks the result against the stated criteria.
Discussion
No questions or reviews yet.
Related
SkillsEleven skills take an agent from an idea to verified code through a PRD and RFCs kept as files in the repository
PaulRBerg's personal skill collection for Claude Code and Codex: commits, refactoring, PDF, GitHub and session reviews.
Two skills for Codex and Claude Code: prd keeps a living implementation checklist and review audits designs from the user's side
A package manager for AI agents that installs rules, skills and prompts and converts them for Cursor, Claude Code and others.
Three skills for Claude Code, Cursor and OpenCode: an interview-driven PRD, task slices and a design doc before any code
Python runtime for agent work: task boundaries, independent verification of results and a saved handoff between sessions
README in short
The author describes PRP as the minimum packet needed to ship a working vertical-slice feature on the first pass, with no excess and no missing context.