llm-evals-audit
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides instructional guidance for auditing evaluation artifacts and does not exhibit any malicious behavior.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to read and analyze potentially untrusted project files such as judge prompts and annotation data, which represents an indirect prompt injection attack surface. This is considered safe as the ingestion is necessary for the skill's primary function and the agent's capabilities are limited to generating an audit report.
- Ingestion points: Step 1 (artifact search and retrieval in SKILL.md)
- Boundary markers: None identified in the provided instructions
- Capability inventory: File reading and report generation
- Sanitization: No sanitization or validation of the content of ingested artifacts is mentioned
Audit Metadata