agentleFS
sainzssainzs/santiagosainz-skills/skills/core/requirements-review/SKILL.md

requirements-review

Skill1 starsChanged 3 months agoMITView on GitHub

Review requirements before implementation by turning fuzzy intent into decisions, constraints, risks, and a first executable step. Use when the user asks to clarify scope, review requirements, resolve ambiguity, identify constraints, or use align-change.

4 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

12 rules. 0 are common, 12 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 12

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Do one job: turn unclear intent into the next executable step.

    line 12
  2. Distinctive:

    Do not hide capability gaps.

    line 15
  3. Distinctive:

    Stop discovery when the entry point and first proof are clear.

    line 16
  4. Distinctive:

    Use for feature ideas, refactors, product decisions, workflow changes, migrations, and plans that are not yet concrete enough to edit …

    line 20
  5. Distinctive:

    Do not use when the request is already a clear edit, a failing command, a review, or a pure explanation.

    line 24
  6. Distinctive:

    Read the current repo or config surface that will be changed.

    line 28
  7. Distinctive:

    Ask only the decisions that block a safe first step.

    line 30
  8. Distinctive:

    If enough context exists, choose conservative defaults and state them.

    line 31
  9. Distinctive:

    Prefer repo truth over stale docs.

    line 36
  10. Distinctive:

    Prefer the smallest user-visible behavior that can be verified end to end.

    line 37
  11. Distinctive:

    Prefer preserving existing interfaces unless the interface is the problem.

    line 38
  12. Distinctive:

    If the user asked for execution, continue into the first step after the review.

    line 51

“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “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

requirements-review

Resolve ambiguity quickly enough that implementation can start.

Unix/Pi Contract

  • Do one job: turn unclear intent into the next executable step.
  • Consume repo facts and user intent; emit decisions, blockers, next step, and proof command.
  • Stay terse. Ask only the question that changes the next action.
  • Do not hide capability gaps. If evidence, tools, credentials, or runtime access are missing, say exactly what is missing.
  • Stop discovery when the entry point and first proof are clear.

Use

Use for feature ideas, refactors, product decisions, workflow changes, migrations, and plans that are not yet concrete enough to edit safely.

Do Not Use

Do not use when the request is already a clear edit, a failing command, a review, or a pure explanation.

Workflow

  1. Read the current repo or config surface that will be changed.
  2. Extract the user goal, non-goals, constraints, existing contracts, likely risks, and missing choices.
  3. Ask only the decisions that block a safe first step. Prefer one concise question over a broad interview.
  4. If enough context exists, choose conservative defaults and state them.
  5. End with a first executable step and the verification command that proves it.

Decision Quality

  • Prefer repo truth over stale docs.
  • Prefer the smallest user-visible behavior that can be verified end to end.
  • Prefer preserving existing interfaces unless the interface is the problem.
  • Name irreversible, external, security, or data-loss risks explicitly.

Output

Return:

  1. Goal in one sentence.
  2. Decisions already resolved.
  3. Open blocker, if any.
  4. First step to implement.
  5. Verification command.

If the user asked for execution, continue into the first step after the review.

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.