The Agent Skills Directory

[SAFE]: The skill operates as a purely advisory gate, providing textual suggestions to the user without writing code or performing network operations.
[COMMAND_EXECUTION]: The skill is configured with a restricted toolset (Grep, Glob, Read) and does not attempt to execute arbitrary commands or scripts.
[DATA_EXFILTRATION]: No network operations or external data transfers were detected. The analysis remains local to the repository and the user chat.
[INDIRECT_PROMPT_INJECTION]: While the skill ingests repository content to provide suggestions, it lacks the dangerous capabilities (such as file-writing or remote code execution) required to be exploited. It functions solely as an informational layer.

subtract-first