What does this skill prevent?

It prevents 'false success' claims where an AI might assume code works without actually running the necessary tests, builds, or linters.

Can this skill be used for build checks?

Yes, it treats build exit codes as mandatory evidence, ensuring that code doesn't just look right but actually compiles and builds correctly.

Why is 'Red-Green' verification important?

It ensures that a regression test is actually valid by proving it fails when the fix is removed and passes when the fix is applied.

How does the 'Gate Function' work?

Before making any claim of success, the skill identifies the proof command, executes it, analyzes the exit code and failures, and only then reports status based on the evidence.

Does it work with delegated agents?

Yes, it requires independent verification of any changes made by other agents or sub-processes before accepting their success reports.

Verification Before Completion

Name: Verification Before Completion
Author: guaderrama

byguaderrama

0•

安全与测试

Enforces a mandatory evidence-first workflow by requiring successful command execution before any work is marked as complete.

Verification Before Completion is a high-integrity skill designed to eliminate false claims of success by enforcing a strict gate function. It requires Claude to execute fresh verification commands—such as tests, linters, or build scripts—and analyze the full output before asserting that a task is finished. By prioritizing evidence over assumptions, this skill prevents broken code from being committed, maintains project trust, and streamlines the development cycle through a rigorous validation process that ensures every claim is backed by reproducible logs.

主要功能

01Strict prohibition of speculative language like 'should' or 'probably'

020 GitHub stars

03Red-green cycle validation for reliable regression testing

04Mandatory execution of verification commands before PR creation

05Line-by-line requirement checklist verification

06Independent verification of changes made by delegated agents

使用场景

01Confirming build success and linter cleanliness before opening a pull request

02Ensuring unit tests pass successfully before committing code to a repository

03Verifying that a bug fix actually resolves the original symptom via reproduction

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add guaderrama/cabo-health-ai verification-before-completion

For use in Claude.ai and ChatGPT

Download Skill