cuopt-server-api-python
Pass
Audited by Gen Agent Trust Hub on Oct 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill's deployment instructions suggest using mutable Docker tags (e.g., 'nvidia/cuopt:latest-cu13'). This introduces a potential supply chain risk where the underlying image could be updated without notice, although the images originate from a trusted organization.
- [DATA_EXFILTRATION]: The skill transmits optimization data to a network endpoint (defaulting to localhost:8000). The documentation explicitly states that the server implementation does not include built-in authentication or TLS, advising users to secure the service behind a gateway.
- [INDIRECT_PROMPT_INJECTION]: 1. Ingestion points: The skill ingests a 'reqId' from the server's response (SKILL.md and asset client scripts). 2. Boundary markers: The skill includes explicit instructions to treat the 'reqId' as untrusted input. 3. Capability inventory: The skill utilizes the 'requests' library for network operations (POST and GET). 4. Sanitization: All scripts and instructions mandate the use of a strict regular expression ('[A-Za-z0-9_-]{1,64}') to validate the ID before it is interpolated into subsequent request URLs.
Audit Metadata