Does this work with any testing framework?

Yes, the skill is framework-agnostic. It relies on the AI identifying and running whatever specific commands your project uses for validation, such as npm test, pytest, or custom scripts.

What is the main purpose of the Verify skill?

It prevents Claude from making claims like 'tests pass' or 'bug fixed' without first running actual commands and reviewing the output to prove the claim is true.

How does this help with Pull Requests?

It ensures that any PR description or commit message is based on successful build and test outputs, significantly reducing the chance of merging broken code.

Why use this instead of just asking Claude to run tests?

This skill adds a systematic reasoning gate that makes verification a mandatory part of the workflow, reducing AI hallucinations and overconfidence regarding task status.

Strict Verification

Name: Strict Verification
Author: dodi-hq

bydodi-hq

0•

セキュリティとテスト

Enforces a rigorous evidence-based workflow by requiring verified command outputs before any task completion claim, commit, or pull request.

The Strict Verification skill establishes a mandatory quality gate for Claude Code, preventing premature success claims by requiring tangible proof of work. It forces the AI to follow a systematic five-step process: identifying the correct verification command, executing it, reading the full output and exit codes, confirming the results, and only then claiming success with evidence. This eliminates 'hallucinated' successes and ensures that every code change—from bug fixes to major features—is backed by passing tests, successful builds, or verified requirements.

主な機能

01Mandatory execution of verification commands and exit code checks

02Proactive detection of overconfident or unverified AI claims

03Strict 'Evidence-First' gate for all code claims

04Line-by-line requirements checklist validation

05Real-time bug fix verification through symptom reproduction

060 GitHub stars

ユースケース

01Verifying that a production build succeeds following dependency or configuration updates

02Confirming a bug resolution by re-running a reproduction script and verifying the fix

03Ensuring all unit and integration tests pass before performing a git commit

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add dodi-hq/dodi-skills verify

For use in Claude.ai and ChatGPT