How does this skill handle 'false' successes or assumptions?

It flags phrases like 'should work' or 'seems correct' as red flags and prevents moving forward until the actual verification command is run and the exit code is checked.

Does this work with delegated agent tasks?

Absolutely. It mandates that Claude must independently verify any changes made by other agents through diffs and fresh testing rather than blindly trusting their success reports.

What is the primary goal of the Verification Before Completion skill?

The primary goal is to ensure that Claude never claims a task is finished or successful without providing fresh, empirical evidence from command-line outputs.

Can this skill be used for bug fixing?

Yes, it specifically requires a red-green cycle where the original symptom is proven to pass only after it was proven to fail, ensuring the fix is valid.

Verification Before Completion

Name: Verification Before Completion
Author: zoosewu

byzoosewu

0•

Security & Testing

Enforces rigorous, evidence-based verification protocols before allowing any claims of task completion or code success.

This skill implements a strict 'Iron Law' for Claude, mandating that no status claims—such as tests passing, builds succeeding, or bugs being fixed—can be made without fresh, verifiable evidence. It acts as a gate function that requires the AI to identify verification commands, execute them, and analyze the full output before asserting success. By preventing rationalizations like 'it should work' or 'it seems fixed,' this skill ensures high-integrity development workflows, reduces regressions, and builds developer trust through mandatory red-green testing cycles and exit-code validation.

Key Features

01Evidence-based verification for tests, builds, linters, and requirement checklists

020 GitHub stars

03Enforcement of Red-Green TDD cycles for regression testing and bug fixes

04Identification and prevention of common success-bias 'Red Flags'

05Independent verification of changes made by delegated agents through VCS diffs

06Mandatory gate function requiring fresh command output before any success claims

Use Cases

01Validating full build and linting compliance across the entire project scope

02Verifying that a bug is actually resolved by reproducing the failure before confirming the fix

03Ensuring 100% test pass rates before committing code or opening a pull request

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add zoosewu/mackup verification-before-completion

For use in Claude.ai and ChatGPT

Download Skill