gnomad-database

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill performs network requests to the gnomAD GraphQL API (gnomad.broadinstitute.org) to fetch genomic data. These operations are directed at a well-known scientific resource and include built-in rate-limiting to comply with the service's usage policies.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied genomic identifiers such as gene symbols, rsIDs, and chromosomal coordinates.
  • Ingestion points: Input parameters for the scripts in the scripts/ directory (e.g., --gene, --variant_id, --rsid).
  • Boundary markers: The implementation uses GraphQL variables for all queries, which effectively separates user-supplied data from the query logic, preventing injection into the API requests.
  • Capability inventory: The skill is capable of making network requests to gnomAD and writing the resulting data to local files specified by the agent.
  • Sanitization: Input is handled via the argparse module and passed as structured variables to the HTTP client, ensuring that data is not interpreted as executable code or query structure.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 07:45 AM
Security Audit — agent-trust-hub — gnomad-database