Skills
skills/flora131/atomic/skill-creator/Gen Agent Trust Hub

skill-creator

Pass

Audited by Gen Agent Trust Hub on Apr 7, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill invokes the claude CLI via subprocesses in scripts/run_eval.py and scripts/improve_description.py to simulate user queries and test skill triggering logic.
  • [COMMAND_EXECUTION]: System commands like lsof are used for local process management in eval-viewer/generate_review.py to handle port conflicts for the visualization server.
  • [EXTERNAL_DOWNLOADS]: The HTML-based evaluation viewer (viewer.html) fetches the SheetJS library from a well-known CDN (cdn.sheetjs.com) to provide spreadsheet visualization capabilities.
  • [DATA_EXFILTRATION]: The skill hosts a local web server on 127.0.0.1 via eval-viewer/generate_review.py to present evaluation results to the user; this is a controlled, local-only communication channel.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 7, 2026, 10:59 PM