What is a Truth Score in this skill?

A Truth Score is a real-time reliability metric on a scale of 0.0 to 1.0 that evaluates the correctness and quality of code, agents, and specific tasks.

Is there a way to view metrics over time?

Absolutely. The skill includes an interactive dashboard and report generation tools that track quality trends, confidence intervals, and improvement metrics over 24h, 7d, or 30d periods.

Can I integrate these metrics into GitHub Actions or GitLab?

Yes, the skill provides structured JSON and HTML export options specifically designed for easy integration into standard CI/CD pipelines to enforce quality gates.

What criteria does the verification system evaluate?

The system performs comprehensive checks on code correctness (syntax/types), security vulnerabilities, performance impact, documentation completeness, and adherence to SOLID principles.

How does the automatic rollback feature function?

If a verification check falls below the defined threshold (defaulting at 0.95), the system utilizes Git integration to instantly revert the codebase to its last known good state.

Verification & Quality Assurance

Name: Verification & Quality Assurance
Author: ricable

byricable

•

Security & Testing

Ensures code reliability and agent output accuracy through automated truth scoring, quality verification, and instant rollback systems.

This skill provides an advanced framework for maintaining high-quality codebases within the Claude Flow ecosystem by implementing a 0.95 accuracy threshold for all agent outputs. It combines real-time truth scoring with automated security checks, best practice validation, and a Git-integrated rollback mechanism that reverts changes failing verification. Ideal for developers building complex AI-driven applications, it offers comprehensive monitoring through interactive dashboards and exportable CI/CD metrics to ensure production-grade reliability across every automated task.

Key Features

01Instant Git-based automatic rollback for changes failing quality thresholds

02Automated verification checks for security, performance, and best practices

03Seamless CI/CD integration with JSON, CSV, and HTML reporting capabilities

04Interactive web-based dashboards for live monitoring and trend analysis

05Real-time Truth Scoring metrics (0.0-1.0 scale) for code and agent reliability

064 GitHub stars

Use Cases

01Monitoring AI agent performance and task accuracy across large-scale projects

02Implementing automated quality gates in CI/CD pipelines to prevent regressions

03Ensuring production-ready code through continuous security and logic verification

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add ricable/ultimate-ai-agent verification-quality

For use in Claude.ai and ChatGPT

Download Skill