Verification & Quality Assurance

Verifies code quality with truth scoring, automated checks, and rollback for agent outputs.

1|Updated Jun 3, 2026
One-click install
npx skills add https://github.com/KentwareDemo/RuView --skill verification-quality-assurance-kentwaredemo
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: Verification & Quality Assurance
Source: https://github.com/KentwareDemo/RuView/tree/main/.claude/skills/verification-quality
Command: npx skills add https://github.com/KentwareDemo/RuView --skill verification-quality-assurance-kentwaredemo

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Agent-generated code and task outputs can silently introduce defects, security issues, or regressions. This Skill enforces measurable quality gates by scoring outputs on a 0.0-1.0 truth scale, running verification checks, and automatically rolling back changes that fall below a configurable threshold (default 0.95). ## Core Features & Use Cases - Truth Scoring & Metrics: View real-time reliability scores per file, agent, or task with trends, statistics, and export to JSON, CSV, or HTML reports. - Verification Checks: Validate code correctness, security, performance, documentation, and best practices on files, directories, or task outputs, with optional auto-fix. - Automatic Rollback: Revert failed changes via Git-based, selective, or dry-run rollback modes to restore the last known good state. - Use Case: In a CI/CD pipeline, run a verification check with JSON output, fail the build if the overall score drops below 0.95, and archive the verification report as a build artifact. ## Quick Start Ask the agent to run a verification check on the current project and show the truth score dashboard with any files below the 0.95 threshold.

Frequently Asked Questions about Verification & Quality Assurance

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run a code verification check with Claude Flow?

Run npx claude-flow@alpha verify check to verify the current directory, or pass --file or --directory to target specific paths. Add --threshold to set a custom pass score and --json for structured output suitable for CI pipelines.

How do I view truth scores for my codebase and agents?

Use npx claude-flow@alpha truth to display scores in table format. Filter by time period with --period 7d, by agent with --agent, or find low-scoring items with --threshold. A --watch flag enables live updates.

Can verification integrate with GitHub Actions or GitLab CI?

Yes, verification outputs JSON that CI pipelines can parse. Run the check with --json, read the overallScore field with jq, and fail the job when the score falls below your threshold, uploading the report as a build artifact.

How does automatic rollback work when verification fails?

Rollback uses Git to restore the last known good state via verify rollback --last-good. Selective mode reverts only failed files, --dry-run previews changes, and backups are created automatically before reverting.

What does the verification system check in my code?

It evaluates five areas: code correctness, best practices, security, performance, and documentation. Each area produces its own score, and the weighted overall score is compared against your configured threshold.

Why is my truth score low and how do I diagnose it?

Low scores indicate failing verification criteria such as security issues or poor documentation. Run truth --verbose --threshold 0.0 for a full breakdown, or verify check --verbose to see per-criterion scores and specific issues.