What is the 'Iron Law' of this skill?

The Iron Law states that no completion claims can be made without fresh verification evidence. If a verification command hasn't been run in the current context, the task cannot be claimed as passing.

How does it improve developer productivity?

While it adds a mandatory check, it significantly reduces time wasted on 'false starts,' broken builds, and rework caused by shipping unverified code or missing requirements.

How does it handle tasks performed by other agents?

It treats agent reports with skepticism, requiring an independent check of the VCS diff and manual execution of verification commands to confirm the reported state is accurate.

Does it work with different programming languages?

Yes, it is language-agnostic. It identifies the appropriate commands for your specific environment, whether it is npm test, cargo build, pytest, or custom shell scripts.

Verification Before Completion

Name: Verification Before Completion
Author: tajohnson

bytajohnson

0•

安全与测试

Enforces rigorous verification protocols and evidence gathering before any task is marked as complete or submitted for review.

This skill establishes a strict 'evidence-before-assertion' framework for AI-assisted development, preventing premature claims of completion or success. It implements a mandatory gate function that requires Claude to identify, run, and read the output of verification commands—such as test suites, linters, and build scripts—before committing code or creating pull requests. By eliminating reliance on assumptions and 'should work' statements, it ensures high-quality output, reduces broken builds, and maintains trust between the AI and human developers through verifiable proof of work.

主要功能

01Systematic 'Gate Function' to identify, run, and verify command outputs and exit codes

02Strict identification and prevention of vague language or rationalized excuses

03Independent verification of delegated agent tasks via VCS diff analysis

04Mandatory execution of fresh verification commands before any completion claim

05Red-Green TDD verification support to ensure regression tests are valid

060 GitHub stars

使用场景

01Ensuring 100% test pass rates before proposing a Pull Request

02Confirming successful project builds and linter runs after major refactors

03Verifying bug fixes through a complete red-green reproduction cycle

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add tajohnson/superpowers verification-before-completion

For use in Claude.ai and ChatGPT

主要功能

01Systematic 'Gate Function' to identify, run, and verify command outputs and exit codes

02Strict identification and prevention of vague language or rationalized excuses

03Independent verification of delegated agent tasks via VCS diff analysis

04Mandatory execution of fresh verification commands before any completion claim

05Red-Green TDD verification support to ensure regression tests are valid

060 GitHub stars

使用场景

01Ensuring 100% test pass rates before proposing a Pull Request

02Confirming successful project builds and linter runs after major refactors

03Verifying bug fixes through a complete red-green reproduction cycle

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add tajohnson/superpowers verification-before-completion

For use in Claude.ai and ChatGPT