agentleFS
wisterx-specwisterx-spec/agent-rails/.agents/skills/verify-spec/SKILL.md

verify-spec

Skill2 starsChanged 7 months agoMITView on GitHub

Extract acceptance criteria from a spec, classify each item, run automatic checks where possible, and return a structured verification report.

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

6 rules. 0 are common, 4 are distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 4 edited versions we read keep these.

  • 4

    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
  1. Very specific:

    Read the ## Acceptance Criteria, ## 验收标准, or equivalent section.

    line 15
  2. Distinctive:

    If no acceptance section exists, return:

    line 18
  3. Distinctive:

    If it fails, inspect the relevant files and fix implementation.

    line 42
  4. Very specific:

    After three failed repair attempts, return ✗ with the last error summary and request arbitration.

    line 45
  5. Distinctive:

    Do not rewrite the spec to make verification easier.

    line 47
  6. Distinctive:

    Do not weaken test assertions unless the spec explicitly says test expectations are still drafts.

    line 48

“Common” is measured: kept by at least 6 in 10 of the 4 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

Verify Spec

Input

Path to a completed spec-*.md.

Step 1: Extract Acceptance Criteria

Read the ## Acceptance Criteria, ## 验收标准, or equivalent section. List every unchecked item (- [ ]) as an acceptance criterion.

If no acceptance section exists, return:

## Verification Results
- ○ No explicit acceptance criteria found; manual review required

Step 2: Classify Each Criterion

TypeSignalsCheck
Test/build commandnpm run, pytest, pnpm, yarn, go test, cargo test, mvn test, build, lintRun command and inspect exit code
File contentexists, contains, does not contain, 不存在, 包含, 不出现Use test, rg, or equivalent
HTTP/API behaviorstatus, redirect, returns, response, 状态码, 返回Use local curl/http client when endpoint is available
Database/schema artifactDDL, migration, ALTER TABLE, schemaCheck generated migration/DDL files or run the configured schema command
Browser/visual/manualbrowser, DevTools, Network, visual, screenshot, UX, 浏览器, 视觉Mark manual (○)
External servicethird-party APIs, paid APIs, production-only servicesMark manual or mock-only (○)

Step 3: Repair Failed Automatic Checks

For each automatic check:

  1. Run the check.
  2. Mark ✓ if it passes.
  3. If it fails, inspect the relevant files and fix implementation.
  4. Re-run only the failing check.
  5. Mark ✗→✓ if the repair succeeds.
  6. After three failed repair attempts, return ✗ with the last error summary and request arbitration.

Do not rewrite the spec to make verification easier. Do not weaken test assertions unless the spec explicitly says test expectations are still drafts.

Step 4: Output

Return this exact shape for run-pending-specs to paste into the session:

## Verification Results
- ✓ npm run build passed
- ✗→✓ API returned 500 first; fixed missing null guard; now returns 200
- ○ Visual layout requires browser review

Legend:

  • ✓ automatic check passed.
  • ✗→✓ automatic check failed, implementation was fixed, then passed.
  • ✗ automatic check still fails and needs arbitration.
  • ○ cannot be automatically verified in this environment.

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.