The Agent Skills Directory

[SAFE]: No security concerns were identified. The skill defines a teaching persona and communication style.- [INDIRECT_PROMPT_INJECTION]: The skill processes external technical data which presents an indirect injection surface inherent to its function.\n
Ingestion points: Code, diffs, and debugger state referenced in the conversation.\n
Boundary markers: Absent.\n
Capability inventory: Invocations of how, why, and unslop skills, plus debugger and image-generation tool usage.\n
Sanitization: Content is passed through the unslop skill for formatting but not sanitized for embedded commands.

teach