nvshmem-enable-tma
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external CUDA source code provided by the user, creating a surface for potential indirect prompt injection.
- Ingestion points: User-provided application source files or source paths specified in the 'Inputs' section of SKILL.md.
- Boundary markers: The instructions require identifying a specific kernel and establishing a 'change boundary' to limit the scope of modifications.
- Capability inventory: The skill directs the agent to perform code assessment, patching, and review, which are standard agent capabilities.
- Sanitization: No explicit sanitization or validation of the input source code is defined in the instructions.
Audit Metadata