eval-audit
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill requires the agent to ingest and analyze untrusted data, specifically evaluation traces and judge prompts, which could contain instructions intended to manipulate the audit results.
- Ingestion points: Evaluation artifacts including traces, evaluator definitions, and labeled data (SKILL.md).
- Boundary markers: Absent; there are no specific instructions or delimiters provided to ensure the agent treats the ingested data as non-executable text.
- Capability inventory: The skill is primarily instructional and does not define network, file-write, or shell execution tools in the provided files.
- Sanitization: Absent; there is no guidance on sanitizing or escaping the contents of the traces before analysis.
Audit Metadata