agentleFS
58254727-blip58254727-blip/codex-task-control-tower/skills/verification-gate/SKILL.md

verification-gate

Skill1 starsChanged 3 months agoMITView on GitHub

Use when software work is about to be reported complete and the requested behavior, regressions, or release safety require evidence-based acceptance.

9 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

5 rules. 0 are common, 4 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 7 edited versions we read keep these.

  • 4

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Focused tests: run the narrow checks tied to the change.

    line 16
  2. Distinctive:

    Relevant regression: verify nearby behavior that shares the changed path.

    line 17
  3. Distinctive:

    Release: when publishing, validate the exact package or commit that will be released.

    line 19
  4. Distinctive:

    Use one gate result:

    line 27
  5. Very specific:

    Do not claim completed without inspectable evidence for the original behavior and requested outcome.

    line 33

“Common” is measured: kept by at least 6 in 10 of the 7 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

Verification Gate

Decide whether the requested outcome is proven. The gate checks behavior, not effort, elapsed time, or a successful-looking interface.

Required Checks

  1. Original behavior: reproduce the original failure, or establish the accepted baseline when the work is a feature.
  2. Requested outcome: inspect the real changed behavior or artifact.
  3. Focused tests: run the narrow checks tied to the change.
  4. Relevant regression: verify nearby behavior that shares the changed path.
  5. Privacy: scan outputs that may contain private or sensitive material.
  6. Release: when publishing, validate the exact package or commit that will be released.

Mark an unavailable check as unverified and state the evidence needed to resolve it. A check is not optional merely because it cannot run locally.

Result

Use one gate result:

  • passed: every applicable check has inspectable evidence.
  • failed: a check proves the outcome or regression requirement is not met.
  • unverified: required evidence is missing, inaccessible, or stale.

Do not claim completed without inspectable evidence for the original behavior and requested outcome. Report the result, evidence, skipped or unavailable checks, remaining risk, and next safe action.

The gate does not authorize production access, deployment, publishing, or any other operation outside the user's existing approval.

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.

Reviews can't be read right now.

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.