skills/glebis/claude-skills/anon/Gen Agent Trust Hub

anon

Pass

Audited by Gen Agent Trust Hub on Jun 12, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill enforces strict local-only data handling by applying 0600 permissions (read/write by owner only) to the mapping files that contain original PII.
  • [SAFE]: It proactively manages supply chain and exposure risks by automatically appending sensitive file patterns to the local .gitignore to prevent accidental commits of PII artifacts.
  • [SAFE]: Includes a defensive heuristic to detect and warn users when sensitive outputs are written to cloud-synced directories like Dropbox, iCloud, or Google Drive.
  • [SAFE]: The de-identification process utilizes a layered stack (Regex and Natasha) designed to operate entirely on the local machine without external data exfiltration.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 12, 2026, 10:20 PM
Security Audit — agent-trust-hub — anon