The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or unauthorized data exfiltration behaviors were detected. The skill adheres to security best practices for handling physiological research data.
[COMMAND_EXECUTION]: The skill uses localized Python scripts for processing. These scripts utilize a shared utility module (_common.py) that enforces strict security controls, such as rejecting symlinks, limiting file sizes to 64MB, and ensuring all file operations occur within a declared root directory to prevent path traversal.
[EXTERNAL_DOWNLOADS]: The skill requires neurokit2==0.2.13 from PyPI. This is a pinned dependency from a well-known service, meeting the criteria for a safe external reference.
[INDIRECT_PROMPT_INJECTION]: The skill processes external CSV files containing physiological data. It mitigates indirect injection risks by performing strict numeric casting (converting all input cells to floats) and applying structured schema validation before the data is interpreted by the agent.

neurokit2