skills/akillness/jeo-skills/headroom/Gen Agent Trust Hub

headroom

Pass

Audited by Gen Agent Trust Hub on Sep 1, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes local CLI tools headroom and graphify via Python's subprocess.run to perform environment diagnostics and context scope checks.
  • Evidence: In scripts/jeo-code-policy-hook.py, the run_read_only function executes commands with stdin=subprocess.DEVNULL, a fixed timeout, and limited output capture to prevent side effects.
  • [PRIVILEGE_ESCALATION]: The installer script scripts/setup-claude-code-policy-hook.sh explicitly manages file permissions to ensure the agent's configuration remains private.
  • Evidence: The script enforces 0600 (owner-read/write only) permissions when creating or updating ~/.claude/settings.json and refuses to follow symlinks to prevent configuration hijacking.
  • [SAFE]: The skill references official Headroom documentation and repositories hosted on Vercel and GitHub, which are well-known services.
  • Evidence: URLs like headroom-docs.vercel.app and github.com/headroomlabs-ai/headroom are used for documentation and installation instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 1, 2026, 11:57 PM
Security Audit — agent-trust-hub — headroom