design-an-interface

To compare API design approaches, this skill generates multiple radically different interface designs using parallel sub-agents and then performs a prose-based trade-off analysis on simplicity, flexibility, and encapsulation. It forces variation to prevent settling on a suboptimal first idea.

Exploring module shapes is done by generating three or more distinct interface options with different constraints, presenting each with a signature and concrete usage example. This process reveals hidden complexities and helps you choose a shape aligned with the module's real needs.

Parallel brainstorming for trade-off analysis works by deploying parallel sub-agents to create radically divergent API shapes for a target module. The resulting designs are then compared based on implementation efficiency and generality without actually implementing the code.

Yes, you can generate interface signatures and concrete usage examples without implementing the module. The skill focuses on caller-focused output and depth-focused comparison, deliberately stopping before code implementation to evaluate design efficiency.

You need to design an API interface twice when you want to reduce the risk of choosing a suboptimal first idea. By producing radically different interface designs and comparing their trade-offs, you ensure the final shape is easier to learn and harder to misuse.

No dependencies or components are required to compare interface designs. You simply describe your target module in one paragraph, and the skill produces multiple interface options to evaluate simplicity, generality, and encapsulation.