doca-gpunetio-ib-write-lat
Pass
Audited by Gen Agent Trust Hub on Jul 18, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill instructions do not contain any patterns attempting to override agent safety guidelines, bypass constraints, or extract system prompts.
- [DATA_EXPOSURE]: No hardcoded credentials, API keys, or sensitive file access patterns were found. The skill operates within standard DOCA SDK environment paths (e.g.,
/opt/mellanox/doca) and advises users to treat GPU handles as temporary credentials, following security best practices. - [EXTERNAL_DOWNLOADS]: All external references point to trusted official NVIDIA documentation and GitHub repositories. No remote scripts or binaries are downloaded or executed from untrusted third-party sources.
- [COMMAND_EXECUTION]: The skill includes instructions for building and running benchmark binaries using standard development tools like
meson,ninja, andnvcc. These commands are strictly tied to the primary purpose of the skill (benchmarking performance) and do not include arbitrary or dangerous shell execution. - [OBFUSCATION]: No obfuscated commands, homoglyphs, or hidden instructions were detected. The included
.sigfile is a standard Sigstore digital signature bundle used for provenance verification. - [DYNAMIC_EXECUTION]: The build process (source compilation) is standard for C/CUDA-based developer tools and does not involve the dynamic execution of untrusted or remote code.
- [INDIRECT_PROMPT_INJECTION]: While the skill involves processing output from a benchmark binary, this data is structured performance metrics (latency numbers) and does not represent a significant surface for indirect injection attacks.
Audit Metadata