lamindb
Pass
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides thorough documentation and code examples for using LaminDB in biological research. No malicious patterns, such as prompt injection, unauthorized data access, or obfuscation, were identified.
- [SAFE]: Explicit security guidance is provided, instructing the agent to never log or transmit secrets like API keys or database URLs and to verify the existence of named environment variables instead.
- [SAFE]: The skill promotes data integrity through the use of curators and schemas, ensuring that external data from APIs or files is validated before registration.
- [SAFE]: External dependencies are restricted to well-known biological and machine learning libraries (e.g., pandas, anndata, wandb) with specific version pinning recommended for reproducibility and security.
Audit Metadata