The Agent Skills Directory

[SAFE]: The skill performs local-only text processing and adheres to privacy best practices. It is designed as a terminal step in a redaction workflow, ensuring sensitive information remains on the local machine.
[COMMAND_EXECUTION]: The skill provides a Python script to process local files. Analysis of the script shows it uses standard argument parsing and file I/O with no evidence of shell injection or unsafe subprocess execution.
[DATA_EXFILTRATION]: No network operations, external API calls, or remote transmissions were detected. The script is explicitly designed to only output summary counts to stdout, preventing restored PII from being captured in the agent's session history.

rehydrate