skills/dartsim/dart/dart-ci/Gen Agent Trust Hub

dart-ci

Pass

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: LOWCOMMAND_EXECUTION
Full Analysis
  • [Command Execution] (LOW): The skill utilizes the gh (GitHub CLI) and pixi tools to interact with remote repositories and local build environments. These commands (gh run rerun, pixi run lint) are standard for CI/CD maintenance and do not involve arbitrary or obfuscated code execution.
  • [Indirect Prompt Injection] (LOW): The skill identifies failure modes by reading external data sources such as GitHub Actions logs (gh run view --log-failed).
  • Ingestion points: GitHub job logs and PR check outputs via gh CLI.
  • Boundary markers: None; the skill provides raw command templates.
  • Capability inventory: Ability to rerun jobs (gh run rerun) and execute local linting/testing (pixi run).
  • Sanitization: None; it relies on the user or agent to interpret log data safely. The risk is low as it primarily informs diagnostic decisions rather than automated critical writes.
Audit Metadata
Risk Level
LOW
Analyzed
Feb 16, 2026, 02:45 AM