tooluniverse-chemical-safety
Pass
Audited by Gen Agent Trust Hub on Mar 8, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is designed to interact with established scientific and regulatory databases, including the FDA, PubChem, and ChEMBL. All external data retrieval is performed through these reputable services.
- [SAFE]: No hardcoded credentials, sensitive file paths, or suspicious network communication patterns were identified in the skill metadata, instructions, or associated scripts.
- [SAFE]: The logic focuses on chemical informatics and does not include instructions to override model safety guidelines or execute unauthorized system-level commands.
- [SAFE]: The Python test script provided is a functional validation tool that utilizes a vendor-specific library and does not contain any indicators of malicious activity or persistence.
- [SAFE]: Evaluated ingestion points from external scientific sources. The indirect injection surface is limited to curated content from trusted entities like the FDA, with no dangerous generic capabilities exposed to the processed data.
Audit Metadata