# Agent Trust Auditor

> - Design and run verification for AI-agent output so you never take an agent's word for anything. Use when the user asks "can I trust my agent's output", "my agent made something up", "the agent fabricated a citation", "how do I verify agent work", "review agent PRs", "audit what the agent did", "agent QA", "build a judge layer", "when can my agent run autonomously", or wants checks, evals, approval gates, or acceptance criteria for delegated AI work. Covers the delegability test, workflow constitutions, cost-tiered auditing, actor/judge separation with allow-block-revise-escalate outcomes, the four-class action taxonomy, auditions before autonomy, and completion-vs-acceptance monitoring. For platform-level agent governance or compliance infrastructure, use the sibling platform-security-playbook skill instead; for application-layer RAG/LLM quality metrics, tracing, and offline eval suites, use llm-evals-engineer.

## Facts
- Page: https://tashan.sh/capability/skill-cloud-byte-consulting-agent-trust-auditor
- tashan id: skill:Cloud-Byte-Consulting/agent-trust-auditor
- Source: https://github.com/Cloud-Byte-Consulting/plugins
- Type: skill
- Category: security
- tashan score: not scored (catalogued only — too little public evidence)
- Adoption: 9.0
- Upkeep: 95.0
- Freshness: 90.0
- Evidence coverage: 84% of the inputs this score can use
- Health: active
- Instruction depth: not yet graded
- License: Apache-2.0
- Official: no

## Install

```sh
cp -r agent-trust-auditor ~/.claude/skills/
```

## Security audit
Not scanned. We audit npm-published capabilities; this one has no npm package we can resolve, or has not reached the queue. This is not a clean bill of health.

---
Measured 2026-08-22 by tashan (https://tashan.sh) from public evidence. Scorer s5.