personal-tool-builder

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS] (SAFE): The skill provides example configurations for Node.js and Python projects. It lists standard, well-known libraries as dependencies.
  • Evidence: Mentions common packages like commander, chalk, ora, inquirer, conf, and click in educational snippets.
  • [COMMAND_EXECUTION] (SAFE): The skill discusses command-line tool development but does not execute any arbitrary or dangerous commands. All code provided is for illustrative purposes.
  • Evidence: Code snippets show standard CLI argument parsing and basic file system operations (fs.readFileSync, fs.writeFileSync).
  • [DATA_EXFILTRATION] (SAFE): The skill emphasizes 'Local-First' architecture, specifically advising that data stays on the user's machine.
  • Evidence: Snippets demonstrate storing data in local directories like ~/.mytool/ using JSON or SQLite, with no network-based exfiltration patterns detected.
  • [PROMPT_INJECTION] (SAFE): No instructions designed to override agent behavior or bypass safety filters were found. The skill maintains a consistent instructional tone.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 05:02 PM