The Agent Skills Directory

[SAFE]: The skill is focused on design exploration and documentation. It follows established software design principles and does not perform sensitive operations.
[INDIRECT_PROMPT_INJECTION]: The skill ingests user requirements to guide sub-agents in generating designs. This creates a standard data ingestion surface. As the output consists of design signatures and prose for human review, and the writing is limited to specific project documentation files (.gsd/DECISIONS.md and S##-CONTEXT.md), this is considered a functional requirement of the design process rather than a security risk.
[COMMAND_EXECUTION]: The skill utilizes a platform-provided capability to spawn parallel agents for task decomposition. This is a legitimate use of agentic features to achieve the skill's objective of generating multiple perspectives concurrently.

design-an-interface