Is this skill language-agnostic?

Yes, it functions by identifying the relevant terminal commands for your specific environment, regardless of the programming language used.

Does it support Test-Driven Development (TDD)?

Absolutely. It enforces a red-green cycle where the model must prove the test fails without the fix and passes with it.

How does this skill prevent false success claims?

It mandates that Claude must run a fresh verification command and analyze the full output before it is allowed to use success-oriented language like 'fixed' or 'complete'.

Can I use this for build and linting checks?

Yes, it specifically targets builds, linters, and compilers, ensuring that the 'exit 0' status is confirmed before moving to the next task.

Verification Before Completion

Name: Verification Before Completion
Author: claudiodearaujo

byclaudiodearaujo

0•

Security & Testing

Enforces a rigorous evidence-based protocol that requires fresh command output before any task is claimed as finished or fixed.

The Verification Before Completion skill implements a mandatory 'Gate Function' to prevent AI hallucinations and premature success claims. It operates on the iron law of 'Evidence Before Assertions,' requiring Claude to identify, execute, and analyze specific verification commands before committing code or reporting progress. By eliminating phrases like 'should work' and 'probably fixed,' this skill ensures that every status update is backed by real-time exit codes and failure counts, significantly reducing technical debt and broken builds in collaborative environments.

Key Features

01Line-by-line requirements checklist validation

020 GitHub stars

03Independent verification of delegated agent tasks and VCS diffs

04Prevention of success rationalization and partial check shortcuts

05Strict red-green regression testing cycles for verified bug fixes

06Mandatory gate function requiring fresh command evidence for all status claims

Use Cases

01Ensuring bug fixes are validated against the original failing symptoms

02Maintaining high trust in automated workflows by requiring exit code 0 evidence

03Preventing the creation of Pull Requests before confirming linter and build success

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add claudiodearaujo/izacenter verification-before-completion

For use in Claude.ai and ChatGPT

Download Skill