agentleFS
andrewthetechieandrewthetechie/andrew-skills/skills/request-to-prd/SKILL.md

request-to-prd

Skill6 starsChanged 4 months agoMITView on GitHub

Turns a rough bug report or feature request into a self-contained, self-grilled first-draft PRD using repo docs, code exploration, recommended answers, and explicit human-only unknowns. Use when the user provides one or more paragraphs of product, bug, or feature text and asks for a PRD, first draft, self-grill, non-interactive design pass, or spec from a rough request.

The file

We split this skill into the rules it gives an agent

10 rules. About 0 are common, 7 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read EXAMPLES.md only if the expected output shape is unclear.

    line 19
  2. Very specific:

    Read relevant root guidance such as AGENTS.md, CLAUDE.md, CONTEXT-MAP.md, CONTEXT.md, and docs/adr/ when present.

    line 33
  3. Very specific:

    If CONTEXT-MAP.md exists, identify the relevant bounded context before reading glossary or ADR files.

    line 34
  4. Distinctive:

    Search code and tests for named concepts, UI text, APIs, commands, errors, workflows, and modules from the request.

    line 35
  5. Distinctive:

    Use existing domain vocabulary.

    line 36
  6. Distinctive:

    Use recommended assumptions only when they are defensible.

    line 41
  7. Very specific:

    Do not update CONTEXT.md or ADRs in this automated pass.

    line 51
  8. Distinctive:

    If the user explicitly asks to publish the PRD, discover the repo's issue tracker workflow and publish there.

    line 52
  9. Distinctive:

    If publishing is not requested or not discoverable, return the PRD draft directly.

    line 54
  10. Distinctive:

    If no relevant repo docs exist, say that and proceed from code and request evidence.

    line 62

≈ “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

Request To PRD

Core Rule

This skill is self-contained. Do not invoke or assume availability of grill-with-docs, /to-prd, to-prd, or setup skills. Recreate the needed behavior from this skill's files.

Required Reads

Before drafting, read:

  • SELF-GRILL.md for the non-interactive grilling process and question bank.
  • PRD-TEMPLATE.md for the output format, bug/feature variants, and publishing rules.

Read EXAMPLES.md only if the expected output shape is unclear.

Quick Start

Given a paragraph or more describing a bug report or feature request, perform a non-interactive grilling pass, answer each question with the best evidence-backed or recommended answer, and produce a first-draft PRD. Do not ask the user questions during the first pass unless the source request is missing or unreadable.

Workflow

  1. Intake the request
    • Classify it as bug, feature, or mixed.
    • Extract stated problem, desired outcome, actors, examples, constraints, urgency, and explicit non-goals.
    • Preserve ambiguity; do not silently convert uncertainty into fact.
  2. Discover repo context
    • Read relevant root guidance such as AGENTS.md, CLAUDE.md, CONTEXT-MAP.md, CONTEXT.md, and docs/adr/ when present.
    • If CONTEXT-MAP.md exists, identify the relevant bounded context before reading glossary or ADR files.
    • Search code and tests for named concepts, UI text, APIs, commands, errors, workflows, and modules from the request.
    • Use existing domain vocabulary. If the request conflicts with the glossary or code, call out the conflict in the PRD.
  3. Self-grill the request
    • Follow SELF-GRILL.md.
    • Answer from code/docs where possible.
    • Use recommended assumptions only when they are defensible.
    • Mark human-only unknowns as Requires human input.
  4. Draft the PRD
    • Follow PRD-TEMPLATE.md.
    • Draft for refinement, not final approval.
    • Include enough implementation shape for an engineer to reason about scope.
    • Separate facts, assumptions, contradictions, and human-only unknowns.
  5. Handle documentation and publishing
    • Do not update CONTEXT.md or ADRs in this automated pass. Recommend doc follow-ups instead.
    • If the user explicitly asks to publish the PRD, discover the repo's issue tracker workflow and publish there.
    • Apply ready-for-agent only if that label vocabulary is documented.
    • If publishing is not requested or not discoverable, return the PRD draft directly.

Quality Bar

  • Bugs include expected vs. actual behavior, reproduction or detection signals when available, suspected affected area, regression risk, and fix validation.
  • Features include actors, core workflows, edge cases, acceptance criteria, and non-goals.
  • The recommended answer is concrete when repo evidence or a defensible default exists.
  • Human-only unknowns are explicit and are not answered by invention.
  • If no relevant repo docs exist, say that and proceed from code and request evidence.

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.