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

The core principle is 'The Iron Law': no completion claims are allowed without providing fresh verification evidence, such as terminal output from a test or build command.

What counts as a 'Red Flag' in this workflow?

Red flags include using words like 'should', 'probably', or 'seems to', or expressing satisfaction like 'Done!' or 'Great!' before running a verification command.

How does this skill improve AI code reliability?

It forces the AI to move through a mandatory 'Gate Function' (Identify, Run, Read, Verify) before making assertions, preventing it from assuming code works without actual execution.

Can this skill be used for regression testing?

Yes, it specifically includes patterns for TDD Red-Green verification, requiring the AI to prove a test fails before the fix and passes after the fix.

Verification Before Completion

Name: Verification Before Completion
Author: ironbob

byironbob

0•

Seguridad y Pruebas

Enforces a strict evidence-based workflow requiring fresh verification evidence before any task is claimed as complete.

This skill establishes a 'Gate Function' for AI agents, mandating that no status claims, PR creations, or commits occur without executing and verifying fresh command-line evidence. It prevents rationalization and 'hallucinated success' by requiring the AI to run tests, linters, or build commands and explicitly confirm the output before reporting progress. By prioritizing empirical evidence over confidence, it eliminates the risk of shipping broken code or missing requirements, ensuring high-integrity development cycles and building trust between the AI and the human developer.

Características Principales

010 GitHub stars

02Anti-rationalization checks to bypass common AI shortcuts and assumptions

03Strict red-green regression test verification patterns

04Systematic Gate Function for identifying and running proof commands

05Prevention of hallucinated success through fresh terminal output analysis

06Mandatory execution of verification commands before any status claim

Casos de Uso

01Ensuring all unit tests pass before a pull request is automatically generated

02Verifying build success and linter cleanliness after major code refactoring

03Validating bug fixes through actual reproduction and resolution testing

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add ironbob/superpower_light verification-before-completion

For use in Claude.ai and ChatGPT

Características Principales

010 GitHub stars

02Anti-rationalization checks to bypass common AI shortcuts and assumptions

03Strict red-green regression test verification patterns

04Systematic Gate Function for identifying and running proof commands

05Prevention of hallucinated success through fresh terminal output analysis

06Mandatory execution of verification commands before any status claim

Casos de Uso

01Ensuring all unit tests pass before a pull request is automatically generated

02Verifying build success and linter cleanliness after major code refactoring

03Validating bug fixes through actual reproduction and resolution testing

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add ironbob/superpower_light verification-before-completion

For use in Claude.ai and ChatGPT