What is the primary purpose of this Claude Code skill?

The skill enforces a strict rule that the AI agent must run verification commands and prove work is correct before claiming a task is finished.

Why is this skill important for AI-assisted coding?

It prevents 'hallucinated success' where an agent might assume its logic is correct without actually executing the code to verify its behavior.

Does it support linter and build verification?

Absolutely. It requires explicit evidence of zero linter errors and successful build exit codes before any success claims can be made.

How does it handle statements like 'it should work'?

It treats phrases like 'should work' or 'seems correct' as red flags, stopping the agent until it provides empirical evidence via command output.

Can this skill be used for TDD (Test Driven Development)?

Yes, it specifically supports red-green cycles, requiring the agent to verify that tests fail when the fix is removed and pass when restored.

Verification Before Completion

Name: Verification Before Completion
Author: claudiodearaujo

byclaudiodearaujo

安全与测试

Enforces a strict evidence-based workflow by requiring successful command execution before any task completion claims are made.

关于

This skill establishes a rigorous 'Iron Law' for Claude Code, preventing the agent from claiming success or satisfaction until it has verified the results with fresh command output. It eliminates guesswork and false positives by mandating a specific gate function: identifying verification commands, executing them, and analyzing the full output—including exit codes—before committing code or creating pull requests. By replacing assumptions with empirical proof, it ensures high-quality contributions and prevents the common pitfall of shipping broken code that 'should have' worked.

主要功能

Supports TDD red-green cycles, linter checks, and build success validation
Mandates fresh verification evidence before any completion claims are allowed
Cross-references VCS diffs to ensure reported success matches actual changes
0 GitHub stars
Implements a five-step Gate Function to identify, run, and analyze command outputs
Detects and blocks 'red flag' language like 'should work' or 'probably fixed'

使用场景

Ensuring complex requirement checklists are fully met before task handoff
Preventing broken code from being pushed by enforcing local test runs
Validating bug fixes through reproduction and regression testing

What are Skills?·How to Install

Install with 🐟 Skill.Fish

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

For use in Claude.ai and ChatGPT

Download Skill

GitHub

关于