pediatric-drug-dosing
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements all calculations locally using transparent Python logic. It does not perform any network operations (e.g., curl, wget, or requests) or external data exfiltration.
- [SAFE]: The script uses a hardcoded database (DRUG_DATABASE) for medication parameters. It does not download external scripts, dynamic libraries, or unverifiable dependencies.
- [SAFE]: User-provided inputs such as weight, age, and medication names are processed through standard mathematical formulas (Mosteller) and dictionary lookups. The skill does not use dangerous dynamic execution functions like eval(), exec(), or subprocess calls.
- [SAFE]: No persistence mechanisms, privilege escalation attempts, or sensitive file access patterns (e.g., searching for .ssh or .env files) were found in the instructions or scripts.
- [SAFE]: The metadata and instructions are consistent with the technical implementation, focusing on pediatric safety and clinical decision support without deceptive patterns or hidden commands.
Audit Metadata