The Agent Skills Directory

Data Exposure & Exfiltration (SAFE): No hardcoded credentials or sensitive file path access patterns were found. The scripts check for the existence of user-specified protein and ligand files as part of their intended validation logic.
Dynamic Execution (SAFE): The setup_check.py script uses the __import__ function to verify that required libraries are installed. The package names are hardcoded in the script and do not come from external or untrusted sources.
Remote Code Execution (SAFE): The analyzed files do not contain logic for downloading and executing remote scripts. Documentation mentions that model weights may be downloaded upon first run, which is standard for deep learning tools.
Indirect Prompt Injection (SAFE): Although the script prepare_batch_csv.py ingests CSV data, the data is only used for structural validation and file existence checks. The script lacks capabilities for network communication or command execution that could be exploited via malicious data content.

diffdock