vendor-contracts
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze and review external, untrusted contract documents such as MSAs and DPAs, which provides a surface for indirect prompt injection. Malicious instructions hidden within these documents could potentially influence the agent's negotiation advice or behavior.
- Ingestion points: User-supplied contract text and document files (MSAs, DPAs, SLAs) processed by the agent.
- Boundary markers: The skill does not implement delimiters or explicit instructions to treat external data as untrusted or to ignore embedded commands.
- Capability inventory: Includes the execution of a local Python script (scripts/check-output.py) to validate the agent's output.
- Sanitization: No sanitization or filtering of the external contract data is performed by the validation script or instructions.
Audit Metadata