The Agent Skills Directory

[SAFE]: The skill reads local project files, specifically thumbnail-config.md, brand-kit.md, and about-me.md, to gather design preferences and brand identity. This is a standard and safe operation for project-scoped tools and does not target sensitive system directories.
[SAFE]: No remote code execution, external package installations, or unverifiable dependencies were identified in the skill instructions.
[SAFE]: The skill's primary output is a formatted text prompt intended for manual use in an external image generation environment, which does not grant the agent any elevated privileges or persistence on the host system.
[SAFE]: The instructional language used for workflow automation, such as the auto-start directive, is consistent with legitimate skill design and does not attempt to bypass safety filters or override core agent constraints.

youtube-thumbnail