agentleFS
Sign inStart free
elfrostelfrost/ai-patchlab/.agents/skills/execute-prp/SKILL.md

execute-prp

Skill1 starsChanged 30 days agoMITView on GitHub

Implement a PRP end to end with validation and mandatory documentation sync. Use when the user asks to execute a PRP, build a planned feature, implement a spec from PRPs/, or carry a planned change through coding, tests, and housekeeping.

The file

Show its rules and text

We split this skill into the rules it gives an agent

13 rules. About 0 are common, 11 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈11

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read the PRP completely.

    line 11
  2. Distinctive:

    every file referenced by the PRP

    line 16
  3. Distinctive:

    After each major task:

    line 23
  4. Distinctive:

    run the task validation command

    line 24
  5. Distinctive:

    After all tasks, run the full validation loop from the PRP.

    line 26
  6. Very specific:

    update DECISIONS.md if architecture changed

    line 29
  7. Distinctive:

    update CLAUDE.md if present

    line 31
  8. Distinctive:

    update README.md if setup or usage changed

    line 32
  9. Distinctive:

    Do not skip validation.

    line 48
  10. Distinctive:

    Do not leave docs stale after implementation.

    line 49
  11. Distinctive:

    Keep commits focused if you commit, but do not force commits when the user did not ask.

    line 50
  12. Distinctive:

    Treat AGENTS.md and CLAUDE.md as sibling runtime documents when both exist.

    line 51
  13. Very specific:

    NEVER run the retrospective skill in the same context that did the implementation — recommend it only.

    line 52

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Execute PRP

Implement the PRP exactly, validate each stage, then sync the shared project documents.

Workflow

  1. Read the PRP completely.
  2. Read:
    • AGENTS.md
    • CLAUDE.md if present
    • DECISIONS.md
    • every file referenced by the PRP
    • examples/ or template/examples/
  3. Confirm pre-flight checks:
    • referenced files exist
    • each task has validation
    • required dependent PRPs are complete
  4. Break the work into concrete tasks and execute them in order.
  5. After each major task:
    • run the task validation command
    • fix failures before continuing
  6. After all tasks, run the full validation loop from the PRP.
  7. Perform mandatory housekeeping:
    • update ROADMAP.md
    • update DECISIONS.md if architecture changed
    • update AGENTS.md
    • update CLAUDE.md if present
    • update README.md if setup or usage changed
    • archive the PRP into PRPs/done/ with a date prefix
    • remove scratch files
  8. Output a self-healing handoff message and STOP. Do NOT run a retrospective in this same conversation:
    PRP [name] complete. Housekeeping done.

Recommended next step in a SEPARATE session:
  retrospective last      (or: retrospective <prp-name>)

Reason: the implementer is biased about its own work. A fresh context
surfaces AI-layer drift (CLAUDE.md, AGENTS.md, examples, skills, agents)
and proposes concrete edits to prevent the same friction next time.

Rules

  • Do not skip validation.
  • Do not leave docs stale after implementation.
  • Keep commits focused if you commit, but do not force commits when the user did not ask.
  • Treat AGENTS.md and CLAUDE.md as sibling runtime documents when both exist.
  • NEVER run the retrospective skill in the same context that did the implementation — recommend it only.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free