discernment-nudge
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted architectural plans and financial projections which could contain hidden instructions designed to manipulate the agent's output.
- Ingestion points: Documents such as system design specifications, ADRs, and financial plans.
- Boundary markers: Includes explicit instructions to treat reviewed content as data (ASI01) to mitigate instruction hijacking.
- Capability inventory: The skill has permissions for file modification and command execution, which could be leveraged if an injection succeeds.
- Sanitization: Guidance provided to derive questions strictly from task objectives rather than embedded instructions.
Audit Metadata