The Agent Skills Directory

[SAFE]: The skill implements a robust workflow for configuring repository-level instructions without any prompt injection or safety bypass mechanisms.\n- [SAFE]: Shell commands used for repository grounding (such as ls, find, cat, and grep) are restricted to read-only operations with limited output via head, making them safe for their intended discovery purpose.\n- [SAFE]: No network operations, data exfiltration patterns, or external dependencies were detected. All provided templates are for local adaptation and do not involve remote code execution.

init-copilot-review