The Agent Skills Directory

[SAFE]: The skill contains no scripts, binaries, or automated command execution logic. It is a procedural guide that uses natural language instructions to define a planning framework.
[INDIRECT_PROMPT_INJECTION]: While the skill is designed to process untrusted user input describing goals and tasks, it incorporates robust mitigations including structured boundary markers (e.g., Destination vs. Not the Destination) and a mandatory confirmation gate before implementation, which reduces the risk of accidental instruction following from processed data.

wayfinder