skills/alibaba/mnn/metal-optimize/Gen Agent Trust Hub

metal-optimize

Pass

Audited by Gen Agent Trust Hub on Nov 10, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill documents optimization strategies, development standards, and benchmarking procedures for the MNN Metal backend. All provided instructions and scripts are consistent with legitimate software engineering practices for performance-critical libraries.
  • [COMMAND_EXECUTION]: Local shell commands are suggested for standard build processes using cmake and make, as well as for executing performance benchmarks like llm_bench. These operations are performed within the developer's local workspace.
  • [DATA_EXFILTRATION]: No network activity, unauthorized data access, or exfiltration patterns were detected. Data handling is limited to local configuration files and temporary profiling logs.
  • [CREDENTIALS_UNSAFE]: No hardcoded credentials, API keys, or private secrets were found. The environment variables listed in the registry are dedicated to hardware-specific performance tuning and profiling controls.
  • [OBFUSCATION]: The skill contains no obfuscated code or hidden instructions. Metal shader code is presented as transparent C++ string literals for embedding into the library's source code, which is standard for the MNN architecture.
  • [DYNAMIC_EXECUTION]: The skill describes the legitimate runtime compilation of shaders by the Metal graphics driver, which is the intended functional design for GPU-accelerated computing on the target platform.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 10, 2026, 07:42 PM