Can I use this skill in long-running sessions?

Yes, it is recommended to trigger the loop every 15 minutes or after finishing a major component to maintain a 'mental checkpoint' and keep the session stable.

Does it support languages other than TypeScript?

Yes, the skill includes predefined commands for both JavaScript/TypeScript environments (npm, pnpm, tsc) and Python environments (pyright, ruff).

What happens if a validation phase fails?

The skill instructs Claude to STOP and fix the identified issues immediately before proceeding to the next phase, preventing cascading errors.

How does the security scan phase work?

It uses targeted grep commands to search for potential secret leaks, such as API keys (sk-) and leftover debugging tools like console.log in the source directory.

When should I run the Verification Loop?

It is best used after completing a feature, before creating a PR, or after significant refactoring to ensure no regressions or type errors were introduced.

Verification Loop

Name: Verification Loop
Author: sori883

bysori883

0•

セキュリティとテスト

Implements a multi-phase validation workflow to ensure code quality, type safety, and security before finalizing changes.

The Verification Loop skill provides a structured framework for Claude Code to validate development work through six rigorous phases: build verification, type checking, linting, test suite execution with coverage targets, security scanning for secrets, and a comprehensive git diff review. It is designed to act as a quality gate before pull request creation or after major refactoring, ensuring that all code meets production standards and is free of common errors like leaked API keys or unnecessary debugging logs. By enforcing a standardized output report, it gives developers and AI agents clear visibility into the health of the codebase.

主な機能

01Multi-language support for automated build and type checking (TypeScript and Python).

02Proactive security scanning for leaked secrets and accidental console logs.

03Standardized verification reports for clear PASS/FAIL project status visibility.

04Comprehensive git diff analysis to identify unintended changes and edge cases.

050 GitHub stars

06Integrated test suite execution with coverage reporting and 80% threshold targets.

ユースケース

01Ensuring high code quality and type safety during intensive refactoring sessions.

02Automated security auditing to prevent the accidental commit of API keys or debugging statements.

03Validating complex code changes and fixing regressions before submitting a Pull Request.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add sori883/ai-settings verification-loop

For use in Claude.ai and ChatGPT

主な機能

01Multi-language support for automated build and type checking (TypeScript and Python).

02Proactive security scanning for leaked secrets and accidental console logs.

03Standardized verification reports for clear PASS/FAIL project status visibility.

04Comprehensive git diff analysis to identify unintended changes and edge cases.

050 GitHub stars

06Integrated test suite execution with coverage reporting and 80% threshold targets.

ユースケース

01Ensuring high code quality and type safety during intensive refactoring sessions.

02Automated security auditing to prevent the accidental commit of API keys or debugging statements.

03Validating complex code changes and fixing regressions before submitting a Pull Request.