What types of checks does Verification Guard support?

It supports a wide range of checks including unit tests, linter outputs, build exit codes, regression tests, and manual requirement checklists.

Why is the 'Red-Green' cycle important in this skill?

The Red-Green cycle ensures that a test actually detects the bug it claims to fix; by verifying the test fails without the fix, it prevents false confidence.

When should I apply this skill in my workflow?

It should be applied before every commit, pull request, or transition to a new task to maintain high code integrity and trust.

How does this skill prevent premature completion claims?

It enforces a mandatory 'Gate Function' where Claude must identify a proof command, run it, and analyze the full output before making any success assertions.

Can this skill be used when delegating to other AI agents?

Yes, it requires independent verification of the agent's changes via VCS diffs and test runs rather than trusting the agent's own success report.

Rigorous Verification Guard

Name: Rigorous Verification Guard
Author: cloudfieldcz

bycloudfieldcz

•

安全与测试

Enforces a strict evidence-based workflow requiring successful command execution and output validation before any task is marked as complete.

Rigorous Verification Guard is a high-integrity skill designed to eliminate premature completion claims and 'hallucinated' success in the development process. It implements a non-negotiable 'Gate Function' that requires Claude to identify, execute, and analyze specific verification commands—such as test suites, linters, and build scripts—before committing code or creating pull requests. By prioritizing fresh, empirical evidence over assumptions or 'should-work' logic, this skill ensures that all changes are genuinely passing and that requirements are fully met, significantly reducing technical debt and rework.

主要功能

01Comprehensive checks for builds, linters, and unit tests

022 GitHub stars

03Red-green cycle validation for regression tests and bug fixes

04Independent verification of delegated agent outputs and VCS diffs

05Checklist-based requirement validation to ensure feature parity

06Mandatory gate function requiring fresh verification evidence for all claims

使用场景

01Ensuring large-scale refactors maintain build stability and linting standards

02Pre-PR validation to ensure zero-failure status before human review

03Confirming complex bug fixes through reproducible failure-to-success cycles

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add cloudfieldcz/cf-powers verification-before-completion

For use in Claude.ai and ChatGPT

Download Skill