ncbi-blast-skill

Pass

Audited by Gen Agent Trust Hub on Sep 20, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • External Package Installation: The skill requires the requests library to manage API communications and provides instructions for its installation via pip. This involves downloading code from the standard public package registry (PyPI).
  • Network Communication: The script interacts with the National Center for Biotechnology Information (NCBI) BLAST API. This communication is essential for sequence search functionality and includes rate-limiting logic to follow provider guidelines.
  • File Output Capability: The skill can write raw search results to the local filesystem, either in a default temporary directory or a user-specified path. This enables the management of large datasets that exceed context limits.
  • Indirect Prompt Injection Surface: The tool processes data from an external source (NCBI search hits), which may include untrusted strings such as query titles or sequence descriptions.
  • Ingestion points: Data is fetched from the NCBI BLAST API and parsed into the agent's context within scripts/ncbi_blast.py.
  • Boundary markers: The skill does not employ specific delimiters for external content, though it favors compact summaries to limit exposure.
  • Capability inventory: The associated script has the ability to perform network requests and write files to the disk.
  • Sanitization: The script parses structured JSON results but does not perform additional sanitization on the descriptive text fields of the biological hits.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 20, 2026, 12:20 PM
Security Audit — agent-trust-hub — ncbi-blast-skill