The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or exfiltration attempts were detected. The skill functions as a template for prompting the agent to analyze problems by breaking them down into basic facts and constraints.
[PROMPT_INJECTION]: The skill uses instructional language to ensure the agent adheres strictly to the provided First Principles template and resists modifications to the prompt structure. This is used as a control mechanism for output consistency rather than a bypass of safety filters.
[NO_CODE]: The skill consists entirely of instructional markdown and metadata. It does not include scripts, binaries, or tool execution permissions, significantly reducing its attack surface.

first-principles