doca-public-knowledge-map
Pass
Audited by Gen Agent Trust Hub on Jul 28, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied questions to provide documentation routing. This represents a standard information retrieval surface.
- Ingestion points: User queries regarding DOCA libraries, tools, and services are processed as input (SKILL.md).
- Boundary markers: The skill contains explicit instructions to refuse out-of-scope requests (e.g., hardware/firmware mutation) and route them to specific alternative skills.
- Capability inventory: The skill itself contains no scripts or tool execution capabilities; it is purely informational.
- Sanitization: Not applicable as the skill does not perform execution or interpolation into executable contexts.
- [COMMAND_EXECUTION]: The routing documentation (references/map.md) includes benign examples of shell commands (such as
ls,pkg-config, andcat) to help users verify their environment or find versions. These are provided as static educational examples for the user rather than instructions for the agent to auto-execute. - [EXTERNAL_DOWNLOADS]: The skill provides numerous URLs to documentation and resource pages. All identified URLs point to official NVIDIA domains (docs.nvidia.com, developer.nvidia.com, catalog.ngc.nvidia.com) or the official NVIDIA GitHub organization, which are trusted sources for this vendor-authored skill.
Audit Metadata