check-work
Pass
Audited by Gen Agent Trust Hub on Nov 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill requires the agent to ingest and verify data from untrusted sources, including claims, evidence, and content from shared VM paths or thread messages provided by other workers.
- Ingestion points: The skill pulls content from cited shared-VM paths and thread messages.
- Boundary markers: The instructions do not specify any delimiters or safety prompts to separate the external data from the skill's instructions.
- Capability inventory: The skill instructions grant the ability to open and read data from a shared filesystem and messaging threads.
- Sanitization: No explicit content filtering, validation, or sanitization of the retrieved evidence is described.
Audit Metadata