melange-qcom

Pass

Audited by Gen Agent Trust Hub on Aug 11, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes the melange-qcom CLI for model management, repository creation, and benchmark reporting. It uses shell variable interpolation to manage identifiers like repository names and model keys across multiple steps.
  • [EXTERNAL_DOWNLOADS]: The skill facilitates the download of AI model artifacts and performance reports from the Melange platform. These operations are intended for the primary purpose of model deployment on target devices.
  • [SAFE]: The skill demonstrates safe credential handling by instructing the agent to use environment variables or specific key files for headless environments and to avoid printing or persisting active credentials during deployment code generation.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 11, 2026, 09:09 PM
Security Audit — agent-trust-hub — melange-qcom