The Agent Skills Directory

[SAFE]: The skill consists entirely of natural language instructions designed to guide the agent's behavior during session start. No executable scripts (.sh, .py, .js) or binaries are included.
[SAFE]: No external network connections, data exfiltration patterns, or hardcoded credentials were found in the file.
[SAFE]: The routing logic is designed as a suggestion layer with explicit instructions to avoid blocking the user or overriding user-specified commands, minimizing the risk of behavioral subversion.

intent-detection