The Agent Skills Directory

[SAFE]: The skill provides instructions and code templates for creating brand personalities and micro-interactions. It does not contain any executable scripts that run on the host environment, nor does it request access to sensitive files, credentials, or network resources.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest brand guidelines and user feedback to generate content. While this represents a data ingestion surface, the risk is negligible as the skill lacks high-privilege capabilities such as shell access or outbound network requests. The instructions focus on creative generation within the agent's safe boundaries.
[COMMAND_EXECUTION]: The workflow section contains descriptive shell-style blocks, but these consist entirely of comments (prefixed with #) and serve as logical placeholders for the agent's thought process rather than executable commands.

whimsy-injector