# Verify Code Quality With Truth Scores

AI-assisted code changes can be hard to trust before merge. This skill helps teams score outputs, run verification checks, and connect quality gates to CI, rollback, and monitoring workflows.

## Install

```bash
npx skillstore add dnyoussef/verification-quality-assurance
```

## Metadata

- Status: approved
- Slug: dnyoussef-verification-quality-assurance
- Version: 2.0.0
- Author version: 2.0.0
- Skillstore revision: r1
- Version status: valid
- Tree hash: 20563cfe79b198a48011c4744c7f5e405109026f0a8ee53c9b1369a47253ee54
- Author: DNYoussef
- GitHub username: DNYoussef
- License: MIT
- Repository: https://github.com/DNYoussef/ai-chrome-extension/tree/main/.claude/skills/verification-quality
- Ref: c579efd02184f1128d6741d16e6c8d951d7d4208
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: confirmation\_required
- Manual install advisory: allowed
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: external\_commands, network, filesystem, env\_access
- Quality score: 38
- Quality tier: warning
- Public page: https://skillstore.pages.dev/skills/dnyoussef-verification-quality-assurance
- Manifest: https://skillstore.pages.dev/api/skills/dnyoussef-verification-quality-assurance/manifest

## Capabilities

- Shows truth scores for code, agents, and tasks on a zero to one scale.
- Runs verification checks for files, directories, tasks, and batch workflows.
- Describes rollback commands for failed verification states and selective recovery.
- Exports verification reports in table, JSON, CSV, HTML, and markdown formats.
- Documents CI quality gates, pre-commit hooks, dashboards, and metric export workflows.

## Use Cases

- Review Pull Requests: Run verification before merge and summarize weak areas that need fixes.
- Gate CI Pipelines: Add score thresholds and report artifacts to automated build checks.
- Monitor Agent Quality: Track agent output trends and identify tasks with declining reliability.

## Prompt Templates

### Check Current Scores

```
Show current truth scores for this project and summarize any items below 0.85.
```

### Verify Changed Files

```
Run verification for the files changed in this branch. Explain failed checks and suggest fixes.
```

### Create CI Gate

```
Create a CI quality gate with a 0.95 threshold. Include report export, artifact handling, and failure messaging.
```

### Design Release Workflow

```
Design a rollback-safe verification workflow for a multi-agent feature release. Include thresholds, monitoring, and manual approval points.
```

## Limitations

- Requires Claude Flow and Node.js to run the documented commands.
- Rollback workflows depend on a clean and reliable Git history.
- Truth scores are quality indicators, not formal proof of correctness.
- External monitoring exports require careful endpoint and secret configuration.

## Best Practices

- Run verification before merge and again after major automated edits.
- Use stricter thresholds for production code than for experiments.
- Review rollback actions manually before allowing destructive changes.

## Anti Patterns

- Do not treat a high score as a replacement for human review.
- Do not run rollback commands when uncommitted work is not backed up.
- Do not send metrics to external endpoints without approval.

## Security Audit

- Audited at: 2026-07-05T13:33:57.228\+00:00
- Summary: No prompt injection attempt was found in SKILL.md. The file is mostly documentation, and many static hits are Markdown formatting false positives, but it still instructs users to run unpinned npx commands, install shell hooks, export metrics to external endpoints, use a secret in a URL query string, and run destructive rollback commands.

## Stats

- Views: 279
- Downloads: 8
- Favorites: 0
- Popularity score: 0