ligandmpnn

Warn

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • EXTERNAL_DOWNLOADS (MEDIUM): The skill directs users to clone code from https://github.com/dauparas/LigandMPNN.git. Because this user/repo is not on the Trusted External Sources list, the dependency is unverifiable and presents a potential supply-chain risk.
  • COMMAND_EXECUTION (LOW): The skill executes local Python and shell commands (modal run, python run.py). This is standard behavior for scientific software but is inherently a privileged operation.
  • PROMPT_INJECTION (LOW): The skill exposes an indirect prompt injection surface by processing PDB files. Ingestion points: --pdb-path parameter. Boundary markers: Absent. Capability inventory: Subprocess calls to Python scripts. Sanitization: Not documented or present in the instruction set.
  • DATA_EXFILTRATION (SAFE): No hardcoded credentials or sensitive data exfiltration patterns were detected.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 17, 2026, 06:09 PM