capture-learning

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • [Prompt Injection] (SAFE): No attempt to override agent instructions or bypass safety filters was found. The content is purely structural for documentation purposes.
  • [Data Exposure & Exfiltration] (SAFE): While the templates include placeholders for sensitive variables like 'API_KEY' and 'DATABASE_URL', these are standard documentation examples and do not contain or leak actual credentials. No network operations or file system access commands are present.
  • [Remote Code Execution] (SAFE): The bash snippets provided (e.g., 'pnpm install') are inert examples within documentation templates and are not executed by the skill.
  • [Indirect Prompt Injection] (SAFE): The skill serves as a static reference for templates and does not ingest external untrusted data or possess capabilities to act on such data.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:35 PM