agentsclimarketplace

Build verify

Skill sh-ai-x/dev-harness-kit/skills/build-verify

verification-before-completion. No "done" without quoted exit code + test count + build log (MUST-L3, hook stop-verify).From its SKILL.md

Install
npx -y skills add sh-ai-x/dev-harness-kit --skill build-verify

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

1.8 KB, as published. Nobody here has run it

← Skills index

build-verify — Evidence-Before-Done

Iron Law (no exceptions)

No "done" without running the actual command and quoting evidence.

Verification priority (strong → weak)

1. exit code (test runner / linter / build)
2. test count (passed/failed/skipped)
3. lint count
4. runtime log (last 30 lines)
5. file path + line number cited

Rules

  • "should work" / "probably fine" / "done" ❌
  • On "passing" claim, stop-verify hook auto-warns on stderr:

    [verify-gate] You said it works but cited no output/exit/test evidence. [verify-gate] Run the verify command and quote the output.

  • 2-layer (MUST-9): (a) hook auto-check + (b) skill advisory
  • Regression fixtures (fixtures/real-bugs/) recommended

Hook integration

stop-verify.sh is auto-active in Build / Review / Security / Ship stages. When user says "done" / "finished", hook receives stop event and runs verify command + outputs result to stderr.

Usage example

User: "TDD done"
Hook: [stop-verify] running pytest...
       passed 12, failed 0, error 0
       OK — quoted 12/12.
       Iron Law L3 verified.

OR (failure):

User: "done"
Hook: [verify-gate] You said it works but cited no output/exit/test evidence.
       Run the verify command and quote the output.

Hand-off

On verify pass → state_codec.transition_stage(root, "review") auto. On fail → loop back to the per-step harness runner (lib/execute.py).

Keep looking

Skills are one crate of 608,666. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.