tooluniverse-chemical-safety

Pass

Audited by Gen Agent Trust Hub on Apr 1, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill defines a rigorous, phase-based workflow for chemical analysis using legitimate external data sources such as PubChem, FDA, DrugBank, and CTD. No malicious patterns were identified in the instructions or the test scripts.
  • [PROMPT_INJECTION]: No evidence of prompt injection, role-play bypasses, or safety guideline overrides was found. The instructions focus on structured data retrieval, evidence grading (T1-T4), and objective reporting.
  • [DATA_EXFILTRATION]: No sensitive file access, credential harvesting, or unauthorized network operations were detected. Tool calls are strictly scoped to chemical database APIs.
  • [EXTERNAL_DOWNLOADS]: The skill references a Python dependency (tooluniverse[ml]) which is consistent with the skill's name and its requirement for machine learning tools (ADMET-AI). No unverified remote scripts or piped bash executions are present.
  • [COMMAND_EXECUTION]: The skill does not perform arbitrary shell command execution. The test_skill.py file contains legitimate testing logic for the integrated chemical database tools and does not exhibit malicious behavior.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 1, 2026, 03:20 AM