optimize-for-gpu
Pass
Audited by Gen Agent Trust Hub on Apr 3, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill facilitates the installation of various GPU acceleration libraries, such as CuPy, Numba, and the NVIDIA RAPIDS suite. These packages are retrieved from official registries including the NVIDIA package index.
- [REMOTE_CODE_EXECUTION]: The skill leverages JIT (Just-In-Time) compilation technologies provided by Numba CUDA and NVIDIA Warp. These tools compile Python source code into executable CUDA binary code to run on the GPU. This is the primary and intended mechanism for achieving high-performance numerical computing.
- [COMMAND_EXECUTION]: Instructions are provided for managing project dependencies using the
uvcommand-line tool. - [PROMPT_INJECTION]: The skill features a surface for indirect prompt injection as it is designed to process external data sources.
- Ingestion points: The skill reads data using methods like
cudf.read_parquet,kvikio.RemoteFile.open_s3_url, andcucim.CuImageacross multiple reference files. - Boundary markers: Not explicitly defined in instructions for handling untrusted data.
- Capability inventory: Significant capabilities including file system writes (via KvikIO), custom code execution (via Numba and Warp), and network access (via RemoteFile).
- Sanitization: None specified for external data inputs.
- Note: This vulnerability surface is inherent to the skill's purpose of data processing and optimization and is documented neutrally as it pertains to standard library usage.
Audit Metadata