bio-interdomain-hgt

Pass

Audited by Gen Agent Trust Hub on Sep 1, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements strong security and scientific integrity practices. It enforces a strict database availability gate using environmental variables ($BIO_DB_ROOT) and checksum-verifies all input databases (arbiter, labels, and comparison collections) before processing. This prevents the use of untrusted or modified datasets.
  • [SAFE]: The Python driver script (run_hgt_evidence.py) uses jsonschema for strict validation of execution manifests and evidence gates. It employs hashlib to ensure data integrity of biological databases, which is a best practice for scientific reproducibility and security.
  • [SAFE]: Command execution is limited to well-known bioinformatics tools (DIAMOND, MMseqs2, MAFFT, IQ-TREE) as described in the documentation. The instructions emphasize using default sensitivity levels to prevent resource exhaustion (DoS) on large datasets.
  • [SAFE]: The skill includes specific technical null hypotheses (e.g., assembly contamination, sampling bias) that must be explicitly evaluated, reducing the risk of deceptive results or metadata poisoning by scientific misinterpretation.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 1, 2026, 03:17 PM
Security Audit — agent-trust-hub — bio-interdomain-hgt