skills/flagos-ai/skills/flagrelease/Gen Agent Trust Hub

flagrelease

Pass

Audited by Gen Agent Trust Hub on Mar 25, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses Docker CLI commands (docker ps, docker inspect) to identify running containers and verify their operational state before proceeding with the pipeline.
  • [EXTERNAL_DOWNLOADS]: Downloads Large Language Models, such as Qwen3-0.6B, for environment smoke testing and verification phases.
  • [COMMAND_EXECUTION]: Executes a series of orchestration steps that involve copying Python scripts (e.g., collect_env_info.py, run_offline_inference.py) into a target container and executing them to perform benchmarks and diagnostics.
  • [EXTERNAL_DOWNLOADS]: Utilizes network mirrors (configured via detect_network.py) to download and install specialized software packages including vLLM and the Flag-series libraries.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 25, 2026, 07:07 AM