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

turbovec

Pass

Audited by Gen Agent Trust Hub on Oct 30, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns or security vulnerabilities were detected. The skill is designed as a library wrapper with explicit security controls.
  • [SAFE]: The provided scripts/preflight.py script implements robust security measures for directory traversal. It uses os.O_NOFOLLOW and dir_fd to prevent symbolic link attacks and includes an explicit blocklist for sensitive directories and files (e.g., .ssh, .aws, private keys, and environment files).
  • [SAFE]: The skill instructions emphasize a 'human-in-the-loop' approach, explicitly requiring the agent to obtain user approval before performing side-effect operations such as package installation (pip install), file system writes, or processing private data.
  • [SAFE]: Integrity verification is enforced through the use of a manifest (source.json) containing SHA256 hashes for critical upstream files, ensuring the agent operates on a known-good version of the library code.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 30, 2026, 09:02 AM