How does it improve developer productivity?

By preventing the AI from skipping critical testing steps, it reduces regressions and the time spent debugging 'hallucinated' success claims that weren't actually verified.

What is the Verification skill for Claude Code?

It is a specialized protocol that forces the AI to run real commands and verify outputs before reporting a task as finished, ensuring claims are based on evidence rather than assumptions.

Can I use this with any programming language?

Yes, it is entirely language-agnostic. It works by executing whatever CLI tools your project uses, such as npm, pytest, go test, or custom build scripts.

What happens if a verification command fails?

The skill requires the AI to report the failure accurately with evidence and return to the investigation phase rather than claiming the task is complete.

Rigorous Task Verification

Name: Rigorous Task Verification
Author: Veraticus

byVeraticus

セキュリティとテスト

Enforces an empirical evidence-first workflow by requiring command execution and output validation before any task is claimed as complete.

概要

The Verification skill is a specialized protocol designed to eliminate 'false positives' in AI-driven development. It implements a strict 'Gate Function' that prevents Claude from claiming work is finished, fixed, or passing without first providing fresh evidence from actual command outputs. By enforcing a Low-Freedom rigidity level, it ensures that every assertion—from bug fixes to build successes—is backed by real-time verification logs, checklists of success criteria, and exit code confirmations, making it an essential tool for high-integrity software engineering workflows.

主な機能

Automated integration with test-runner agents
0 GitHub stars
Evidence-before-claims enforcement protocol
Red-green TDD cycle verification
Line-by-line success criteria checklist validation
Pre-commit and build-state validation

ユースケース

Verifying bug fixes by running regression tests and confirming original symptoms are gone
Ensuring production-ready code by validating all success criteria before marking an epic as complete
Preventing premature commits by enforcing full build and linter passes with evidence

概要

主な機能

Automated integration with test-runner agents
0 GitHub stars
Evidence-before-claims enforcement protocol
Red-green TDD cycle verification
Line-by-line success criteria checklist validation
Pre-commit and build-state validation

ユースケース

Verifying bug fixes by running regression tests and confirming original symptoms are gone
Ensuring production-ready code by validating all success criteria before marking an epic as complete
Preventing premature commits by enforcing full build and linter passes with evidence