The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use ripgrep (rg) to discover available style modules within the skill's references directory. This is a local discovery mechanism used for internal navigation of provided assets.
[PROMPT_INJECTION]: The skill is designed to process untrusted context such as user briefs, screenshots, and repository metadata, which creates an attack surface for indirect prompt injection.
Ingestion points: SKILL.md (briefs, repository metadata, screenshots, existing icons, brand assets).
Boundary markers: Absent.
Capability inventory: SKILL.md (local command execution, image generation, file-writing for exports).
Sanitization: Absent.
[SAFE]: The skill references official documentation and guidelines from trusted organizations including Apple, Google, Microsoft, and Mozilla to ensure design compliance across different operating systems.

app-icon-design