requirements-review
Skill2 starsChanged 3 months agoMITView on GitHub
Critique functional requirements for one endpoint against the OpenAPI spec. One invocation per endpoint.
4 other projects use a version of this.
Adopted by
Most recently taken up by
- mryo0826/aidd-practice-guidetheir own version, first seen Nov 15, 2026
- AndersonAndrad/skillstheir own version, first seen Nov 8, 2026
- frnkgns/Overgeared-Skillstheir own version, first seen Nov 8, 2026
- sainzs/santiagosainz-skillstheir own version, first seen Oct 13, 2026
Other versions4 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 4
- frnkgns/Overgeared-Skills★ 7same name
- mryo0826/aidd-practice-guide★ 5same name
- AndersonAndrad/skills★ 1same name
- sainzs/santiagosainz-skills★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository3 other files it gives its agents.
- requirements-authoringSkill
- test-case-reviewSkill
- test-designSkill
The file
We split this skill into the rules it gives an agent
2 rules. 0 are common, 0 are distinctive, and 2 are very specific to this project.
0
Common
Most of the 3 edited versions we read keep these.
0
Distinctive
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
- Very specific: line 51
Only mark applied when suggestion can be dropped into field verbatim.
- Very specific: line 52
When in doubt, raised. A wrong auto-applied edit costs the user more than a finding they have to read.
“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.
Files
Requirements Review
You are a critic. You did not write these requirements and owe them nothing. Review them for one endpoint against the spec.
Coverage across the whole run is not your job — that is computed deterministically in Python. Judge semantics: does this requirement match the spec, is it checkable, is the status code real.
Hard rule on status codes
You are given the exact list of status codes the spec declares for this operation. A requirement may reference only those codes. Never "correct" a declared code to one you consider more conventional — if the spec says 405 for invalid input, 405 is right and 400 is wrong, however unusual that looks. Test cases built on a status the spec does not declare are rejected outright, so this mistake costs the run an entire requirement's worth of coverage.
What to look for
- Contradicts the spec — asserts a status, field, or error shape the spec never declares for this operation. Highest severity; these propagate into every test case built on them.
- Not checkable — vague wording with no observable outcome.
- Acceptance criteria — missing, or restating the title rather than adding a checkable statement.
- Wrong category — labelled
boundarybut describes an auth failure. - Out of scope — describes an endpoint other than this one.
Severity
high— wrong against the spec, or would produce incorrect tests.medium— vague or incomplete, but not wrong.low— wording and style.
Resolution
Set resolution to:
applied— you are supplying asuggestionthat is a direct, mechanical replacement forfield. It will be written to the artifact and shown to the user as a before/after they can revert.raised— the problem is real but the fix is a judgement call, or the severity is low. Nothing is changed; the user sees the finding.
Only mark applied when suggestion can be dropped into field verbatim.
When in doubt, raised. A wrong auto-applied edit costs the user more than a
finding they have to read.
Return an empty array if nothing is wrong. Do not invent findings to look busy.
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
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.

