Why is this better than standard AI reporting?

Standard AI may assume a fix worked based on the code change; this skill enforces empirical proof, significantly reducing the risk of shipping broken code or missing requirements.

What is the 'Iron Law' of this skill?

The Iron Law is the absolute requirement that no completion claims can be made without providing fresh, verifiable evidence from command outputs.

Can I use this for Test-Driven Development (TDD)?

Yes, it specifically supports Red-Green cycles by requiring the agent to prove a test fails before proving it passes, ensuring the test is actually valid.

How does this skill handle 'satisfaction' claims?

It prevents the AI from expressing satisfaction or saying 'Done!' until it has executed the necessary verification commands and confirmed the actual output matches the expected result.

Verification Before Completion

Name: Verification Before Completion
Author: claudiodearaujo

byclaudiodearaujo

•

安全与测试

Enforces strict evidence-based verification protocols before marking tasks as complete to ensure code reliability and integrity.

This skill establishes a rigorous 'Iron Law' for AI agents, preventing premature claims of success by requiring fresh, empirical evidence through command execution. It acts as a gate function that mandates running tests, builds, or linters and verifying their actual output before any work is committed or reported as finished. By eliminating assumptions and 'should-be' logic, it builds trust between the user and the AI, ensuring that every claim of a fix or feature completion is backed by concrete logs and exit codes, effectively preventing regressions and broken builds.

主要功能

01TDD Red-Green Support: Validates regression tests through verified failure and success cycles.

02Red Flag Detection: Identifies and stops vague language like 'should' or 'probably'.

03Agent Supervision: Independently verifies delegated tasks through VCS diffs and fresh logs.

04Requirement Checklists: Cross-references implementation against project plans line-by-line.

051 GitHub stars

06Mandatory Gate Function: Requires command execution before any success assertions.

使用场景

01Preventing broken builds by mandating fresh compilation checks before PR creation.

02Maintaining high trust in automated workflows by banning premature success reports.

03Ensuring bug fixes are actually effective by requiring red-green test verification.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add claudiodearaujo/sistema-de-narra-o-de-livro verification-before-completion

For use in Claude.ai and ChatGPT

Download Skill

主要功能

01TDD Red-Green Support: Validates regression tests through verified failure and success cycles.

02Red Flag Detection: Identifies and stops vague language like 'should' or 'probably'.

03Agent Supervision: Independently verifies delegated tasks through VCS diffs and fresh logs.

04Requirement Checklists: Cross-references implementation against project plans line-by-line.

051 GitHub stars

06Mandatory Gate Function: Requires command execution before any success assertions.

使用场景

01Preventing broken builds by mandating fresh compilation checks before PR creation.

02Maintaining high trust in automated workflows by banning premature success reports.

03Ensuring bug fixes are actually effective by requiring red-green test verification.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add claudiodearaujo/sistema-de-narra-o-de-livro verification-before-completion

For use in Claude.ai and ChatGPT

Download Skill