test
Pass
Audited by Gen Agent Trust Hub on Nov 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data from test reports which could theoretically contain malicious instructions.
- Ingestion points: The skill reads summary reports produced by the
testersubagent and external report files specified via the--provenargument. - Boundary markers: The instructions do not specify the use of delimiters or specific escaping when processing text from these reports, creating a surface for embedded instructions.
- Capability inventory: The skill can read files, write to iteration logs, and authorize the
spawn_agentsubagent for task execution. - Sanitization: The skill includes explicit instructions to 'NEVER copy sensitive fixture data into the goal file,' which mitigates data exposure risks during report generation.
Audit Metadata