How does the automatic rollback feature work?

If a verification check falls below your specified threshold (default 0.95), the system utilizes Git integration to instantly revert files to their last known good state.

Is the quality threshold customizable?

Yes, you can set global, environmental, or per-task thresholds, allowing you to require higher accuracy (e.g., 0.99) for production and more leniency for development.

Can I integrate this skill with GitHub Actions?

Yes, the skill provides structured JSON output and CLI commands specifically designed for easy integration with GitHub Actions, GitLab CI, and other automation platforms.

What specific quality checks does the skill perform?

It evaluates code correctness (syntax and logic), security (vulnerability and secret detection), performance (complexity and memory), and documentation accuracy.

What is a Truth Score in this context?

A Truth Score is a real-time reliability metric ranging from 0.0 to 1.0 that evaluates the quality, correctness, and statistical confidence of code changes or agent tasks.

Verification & Quality Assurance

Name: Verification & Quality Assurance
Author: Microck

byMicrock

•

セキュリティとテスト

Ensures codebase reliability through real-time truth scoring, automated quality verification, and an intelligent rollback system for AI-generated outputs.

This skill serves as a rigorous safety and validation layer for AI-driven development by quantifying the reliability of agent outputs on a 0.0 to 1.0 scale. It automates high-standard checks for security, performance, and architectural best practices, while providing a critical failsafe mechanism that automatically reverts changes failing to meet defined quality thresholds. Designed for production environments, it offers comprehensive monitoring via live dashboards and seamless integration with CI/CD pipelines to maintain high-confidence code standards.

主な機能

01Real-time Truth Scoring (0.0-1.0 scale) for agents and codebase tasks

0289 GitHub stars

03Intelligent automatic rollback system for changes failing quality thresholds

04Automated verification for security, performance, and coding best practices

05Comprehensive CI/CD integration with JSON, CSV, and HTML reporting

06Interactive monitoring dashboards with live reliability and trend analysis

ユースケース

01Monitoring and benchmarking the long-term reliability of different AI coding agents

02Implementing a safety net that reverts low-quality refactors in high-stakes environments

03Automating the validation of AI-generated pull requests before production deployment

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add microck/ordinary-claude-skills verification-quality-assurance

For use in Claude.ai and ChatGPT

Download Skill