The Agent Skills Directory

[SAFE]: The skill is purely instructional and provides a set of behavioral guidelines for AI agents to follow. These guidelines focus on efficiency, such as performing targeted file reads rather than tree-wide searches and preferring narrow patches over full-file rewrites.
[NO_CODE]: No executable scripts (Python, Node.js, or Shell) are present in the skill package. The files consist entirely of Markdown instructions and YAML configuration for agent interfaces.
[DATA_EXFILTRATION]: There are no commands or instructions that attempt to access sensitive files or exfiltrate data. Mention of tools like doc-to-lora-evaluator are provided as high-level workflow suggestions rather than executable code paths.

token-efficient-execution