external-cannbot-ops-ascendc-direct-invoke-template

Pass

Audited by Gen Agent Trust Hub on Sep 7, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill serves as a project template for developers working with Huawei's CANN (Compute Architecture for Neural Networks) to implement custom NPU operators. All components, including the shell scripts (run.sh), C++ source files, and Python testing utilities, are consistent with this development use case.
  • [COMMAND_EXECUTION]: The skill uses standard build tools such as cmake and make to compile NPU kernels. Shell scripts perform environment setup by sourcing the local Ascend environment configuration (set_env.sh), which is required for CANN development.
  • [EXTERNAL_DOWNLOADS]: No suspicious external downloads were detected. The Python scripts rely on standard machine learning libraries like numpy, torch, and torch_npu for data processing.
  • [DATA_EXFILTRATION]: The skill reads and writes binary data files locally (e.g., input_x.bin, output.bin) for functional verification of the kernels. There is no evidence of unauthorized data access or network-based exfiltration.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 7, 2026, 03:53 PM
Security Audit — agent-trust-hub — external-cannbot-ops-ascendc-direct-invoke-template