Why is the Verification Before Completion skill necessary?

It prevents 'hallucinated success' where an AI might claim a task is finished based on code changes alone without actually running the tests or build commands to prove it works.

What is the 'Iron Law' mentioned in this skill?

The Iron Law states that no completion claims can be made without fresh, explicit verification evidence from a command output run in the current context.

How does this skill handle regression tests?

It enforces a Red-Green cycle: you must prove the test fails without the fix, then prove it passes with the fix, ensuring the test is actually valid and effective.

Does this skill work with linters and builds?

Yes, it treats linter clean-up and build success as distinct verification gates that require full command output (exit code 0) before any success assertions are allowed.

Verification Before Completion

Name: Verification Before Completion
Author: ZhanlinCui

byZhanlinCui

•

Security & Testing

Enforces rigorous command-based verification and evidence collection before any task is marked as complete or successful.

About

This skill implements a non-negotiable 'Iron Law' for Claude Code, preventing premature claims of success by requiring fresh, empirical evidence from verification commands. It acts as a critical gate function that mandates a systematic workflow: identifying the proof command, executing it, reading the full output, and verifying the result against the claim. By eliminating reliance on assumptions or 'should work' scenarios, it ensures high-integrity software delivery, maintains developer trust, and prevents regression errors caused by partial verification or AI overconfidence.

Key Features

Mandatory gate function requiring fresh command output before any status claim
17 GitHub stars
Rationalization prevention to eliminate words like 'should', 'probably', or 'seems to'
Comprehensive verification checklists for tests, linters, builds, and requirements
Standardized Red-Green-Refactor verification patterns for regression testing
Independent verification protocols for delegated agent tasks and VCS diffs

Use Cases

Validating bug fixes with evidence-based regression testing before opening a Pull Request
Ensuring full linter and build compliance before committing code to a repository
Verifying complex requirement checklists after significant architectural refactors

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add zhanlincui/ultimate-agent-skills-collection verification-before-completion

For use in Claude.ai and ChatGPT

Download Skill

GitHub