tools-archfit

Pass

Audited by Gen Agent Trust Hub on Aug 22, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes the archfit binary using common local paths (./archfit, ./.bin/archfit) or the system PATH. These commands are used to perform architectural analysis, health checks, and scorecard generation.
  • [SAFE]: The skill explicitly defines a 'Read-only discipline', instructing the agent to write command outputs to temporary directories outside the target repository and forbidding the creation of baselines, labels, or config patches without explicit user approval.
  • [SAFE]: The instructions mandate that the agent treat deterministic tool output as evidence for calibration rather than a replacement for independent judgment, reducing the risk of being misled by potentially poisoned tool configurations.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 22, 2026, 10:31 AM
Security Audit — agent-trust-hub — tools-archfit