Verification & Quality Assurance

This Skill automates verification by calculating truth scores (0.0-1.0) for code and agent output, then triggers automatic rollback when scores fall below your threshold—default 0.95. Integrate it into your CI/CD pipeline to prevent low-quality code from merging.

Truth scoring is a real-time reliability metric that assigns a 0.0-1.0 score to code, agents, and tasks based on verification checks. The Skill runs security, performance, and functional assessments to generate these scores, helping you identify and quantify output quality.

Yes. The Skill exports dashboard-ready metrics and integrates directly into CI/CD workflows via Claude Flow, Git, and Node.js 18+. Use it to gate deployments—automatically reject or roll back changes when truth scores miss your defined threshold.

Run CLI-based verification checks and specify your custom threshold (the Skill defaults to 0.95). When code or agent output scores fall below that limit, automatic rollback triggers immediately, reverting the changes without manual intervention.

You need Claude Flow (npx claude-flow@alpha), a Git repository, and Node.js 18 or later. The Skill requires these to run truth scoring, execute verification checks, and perform automatic rollback on your codebase.

Yes. The Skill delivers real-time truth scores and automated verification checks designed for dashboards and live monitoring. Use it to track agent reliability continuously and trigger rollback or alerts when output quality degrades below your threshold.