reading-business-cards
Pass
Audited by Gen Agent Trust Hub on Jul 30, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill utilizes standard, well-known Python libraries for image processing, including OpenCV, NumPy, Pillow, and Scikit-Image.- [SAFE]: Network communication is restricted to the official Anthropic API endpoint (api.anthropic.com) for the sole purpose of card transcription.- [SAFE]: Credential management for API keys uses standard and secure patterns, prioritizing environment variables and designated configuration files like .env over hardcoding.- [SAFE]: The extraction prompt (haiku_extract.md) uses structured delimiters and explicit constraints to mitigate the risk of indirect prompt injection from text found on the business cards.- [SAFE]: All file system operations are localized to user-provided work directories or standard data paths (/mnt/user-data/outputs), with no attempts at privilege escalation or persistence.
Audit Metadata