Does this skill work with any programming language?

Yes, it is language-agnostic. It relies on the specific testing, building, or linting commands appropriate for your particular project and environment.

When exactly does this skill trigger?

The skill triggers whenever Claude is about to claim work is complete, state that tests have passed, claim a bug is fixed, or attempt to create a pull request.

Why does this skill ban phrases like 'seems to work'?

To prevent false confidence and 'hallucinated' success. If the code hasn't been verified with fresh output, any claim of success is treated as an assumption rather than a fact.

What is the main purpose of the Verification Before Completion skill?

It serves as an evidence-first gate that prevents Claude from claiming work is done until it has successfully run verification commands and shared the output.

Verification Before Completion

Name: Verification Before Completion
Author: luanrodrigues

byluanrodrigues

•

보안 및 테스팅

Enforces evidence-based task completion by requiring fresh command execution and output verification before any success claims are made.

The Verification Before Completion skill implements a strict 'Iron Law' for software development, preventing AI assistants from making premature or unsubstantiated success claims. It acts as a mandatory gate function that requires the AI to identify a specific verification command, execute it, and provide the raw output as evidence before stating a task is complete. By banning phrases like 'should work' or 'appears to,' this skill reduces silent failures, improves developer trust, and ensures that code changes are truly functional and regression-free before they are committed or reviewed.

주요 기능

01Multi-point verification checklists covering tests, builds, and requirements

022 GitHub stars

03Strict requirement to paste full command output as objective evidence

04Mandatory execution of verification commands before claiming task completion

05Automated detection and prevention of non-committal phrases like 'should be working'

06Standardized 'Command-First' reporting structure for consistent feedback

사용 사례

01Confirming bug fixes through a verified red-green test cycle

02Validating build success and linter cleanliness before committing to a repository

03Ensuring all unit and integration tests pass before opening a pull request

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add luanrodrigues/ia-frmwrk verification-before-completion

For use in Claude.ai and ChatGPT

주요 기능

01Multi-point verification checklists covering tests, builds, and requirements

022 GitHub stars

03Strict requirement to paste full command output as objective evidence

04Mandatory execution of verification commands before claiming task completion

05Automated detection and prevention of non-committal phrases like 'should be working'

06Standardized 'Command-First' reporting structure for consistent feedback

사용 사례

01Confirming bug fixes through a verified red-green test cycle

02Validating build success and linter cleanliness before committing to a repository

03Ensuring all unit and integration tests pass before opening a pull request

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add luanrodrigues/ia-frmwrk verification-before-completion

For use in Claude.ai and ChatGPT