boltz

Pass

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: LOWEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • Unverifiable Dependencies & Remote Code Execution (LOW): The skill utilizes external packages including boltz, modal, and biopython. These are sourced from non-whitelisted repositories, requiring verification of package integrity.
  • Indirect Prompt Injection (INFO): The skill processes research data from FASTA files.
  • Ingestion points: complex.fasta (referenced in SKILL.md)
  • Boundary markers: Absent; sequences are processed directly
  • Capability inventory: boltz predict and modal run (SKILL.md)
  • Sanitization: No sanitization logic for input sequences is documented
  • Command Execution (LOW): The skill provides instructions for running prediction tasks via the command line, including the execution of the modal_boltz.py script. While documented and intended, these operations involve the execution of shell commands and external scripts.
Audit Metadata
Risk Level
LOW
Analyzed
Feb 16, 2026, 12:45 AM