alterlab-boltz

Pass

Audited by Gen Agent Trust Hub on Jul 6, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill installs the boltz Python package and downloads model weights from its official repository. This is standard behavior for deploying and running the Boltz-2 model.
  • [DATA_EXFILTRATION]: The skill provides an option to fetch Multiple Sequence Alignments (MSAs) from a hosted server. The documentation correctly notes that this involves sending sequences externally and advises caution for sensitive data.
  • [COMMAND_EXECUTION]: The skill uses Python and shell commands to perform scientific predictions. These operations are directly related to the skill's stated purpose and operate within the expected tool environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 6, 2026, 11:26 PM
Security Audit — agent-trust-hub — alterlab-boltz