# API Design

> Use when: design a clear, consistent, and evolvable API contract before implementing it.

- Skill: `kimtth/api-design` (Agent Skill)
- Install (CLI): `npx skillmds@latest add kimtth/api-design`
- Raw SKILL.md: https://api.skillmd.com/api/skills/kimtth/api-design/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Integrations & APIs
- Author: kimtth (https://skillmd.com/u/kimtth)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/kimtth/api-design

---


Goal: an API that is predictable, hard to misuse, and safe to evolve.

Use for:
- designing a new API surface or endpoint set
- reviewing naming, consistency, and error semantics
- planning backward-compatible changes

Workflow:
1. Model the resources and operations from the consumer's view.
2. Use consistent naming, casing, and pluralization.
3. Define clear request/response shapes and error formats.
4. Choose explicit status codes and meaningful error bodies.
5. Plan versioning and backward compatibility up front.
6. Document the contract and validate it against real usage.

Principles:
- make the common case simple and the right thing easy
- consistent conventions beat per-endpoint cleverness
- return structured, actionable errors
- design for additive, non-breaking evolution

Rules:
- never break a published contract silently; version it
- validate inputs and reject ambiguity explicitly
- keep responses predictable in shape and naming
- document before consumers depend on undocumented behavior