skills/matt-riley/agent-skills/fallow/Gen Agent Trust Hub

fallow

Pass

Audited by Gen Agent Trust Hub on Sep 28, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADS
Full Analysis
  • [SAFE]: The skill implements proactive security guardrails, specifically advising the agent to treat remote extends URLs in configuration files as untrusted input to prevent potential exploitation via remote configuration.
  • [SAFE]: It enforces a safe workflow for destructive operations by requiring a preview (fallow fix --dry-run) before any automated cleanup is applied, ensuring that the agent's actions are validated against user intent.
  • [EXTERNAL_DOWNLOADS]: The skill utilizes the 'fallow' package via npx, which is a standard method for executing JavaScript tools from the npm registry.
  • [SAFE]: The instructions demonstrate technical robustness by providing specific guidance on clean output handling (e.g., using --format json --quiet and redirecting stderr to /dev/null) and correct exit code interpretation (handling code 1 as a successful report of findings rather than a tool failure).
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 28, 2026, 05:29 PM