chembl-database

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes a local Python script (scripts/chembl_api.py) using uv run. This follows best practices by centralizing API logic, enforcing rate limits, and providing a controlled interface for the agent to interact with external data.
  • [EXTERNAL_DOWNLOADS]: Network operations are directed to https://www.ebi.ac.uk/, the official and reputable domain for the European Bioinformatics Institute. The skill fetches scientific data (bioactivity, chemical structures) consistent with its stated purpose.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data from a remote API and writes it to local files, creating a data ingestion surface.
  • Ingestion points: Bioactivity data and molecular descriptions are fetched via API calls in scripts/chembl_api.py.
  • Boundary markers: The agent is instructed to process structured JSON output, which provides natural boundaries for the data.
  • Capability inventory: Capabilities are restricted to making network GET requests to a specific trusted domain and writing results to user-defined file paths.
  • Sanitization: The script employs standard URL encoding for query parameters and JSON serialization for output, ensuring the data is handled safely.
  • Conclusion: Given the reputable data source and the tool's specialized bioinformatic focus, the risk is assessed as safe.
  • [SAFE]: The skill includes a benign license compliance workflow that notifies the user of data terms and creates a local record of the notification, following standard administrative practices.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 03:48 PM