The Agent Skills Directory

[SAFE]: The skill is composed entirely of natural language instructions and formatting templates. No executable code, shell commands, or dynamic execution patterns are present in the skill definition.
[SAFE]: No remote dependencies or external script downloads are requested. The Python and PyTorch code blocks provided in the reference files are static examples intended for formatting guidance rather than execution.
[SAFE]: There is no evidence of obfuscation, persistence mechanisms, or credential harvesting. The skill operates strictly as a text-to-text transformer.

lecture-alchemist