molecular-docking
Pass
Audited by Gen Agent Trust Hub on Jul 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides structured guidance for scientific workflows involving AutoDock Vina, OpenBabel, and Meeko. It contains best practices for ligand and receptor preparation, grid box sizing, and validation protocols.
- [COMMAND_EXECUTION]: The skill provides templates for command-line tool usage (e.g.,
obabel,mk_prepare_ligand.py). These are standard operations for the intended bioinformatics use case and do not involve arbitrary command execution or shell injection patterns. - [EXTERNAL_DOWNLOADS]: The skill references specific versions of scientific packages in its metadata. These are well-known software packages in the drug discovery domain and are used appropriately within the workflow.
- [PROMPT_INJECTION]: The skill has a potential surface for indirect prompt injection as it processes external scientific data (SMILES strings, PDB files). However, the instructions are focused on providing specific pipeline parameters, and there are no signs of malicious instruction embedding or safety bypass attempts.
Audit Metadata