agentleFS
pabloleonalcaidepabloleonalcaide/development-skills/skills/api-design/SKILL.md

api-design

Skill1 starsChanged 36 days agoMITView on GitHub

Design stable public interfaces — HTTP endpoints, event payloads, module boundaries, SDK-style ports — and make retried operations safe with idempotency keys. Use when adding or changing an endpoint, a request/response or event contract, an error format, a public module interface, or any operation that can be retried (payments, webhooks, jobs, queue consumers). Invoked by /develop when the blueprint declares API contract impact.

191 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use db-design, everything-claude-code and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • yuexueyuSkill

    db-design

    数据库工程师技能 - Schema设计、索引优化、并发安全、数据库审计。当你涉及数据库表/字段/索引/SQL/迁移/MCP数据库操作、建表改表、写查询时必须使用此技能。即使用户只是说"加个字段"或"查下数据",也应触发。

    Also in 4 projects that use this

  • InfopibeSkill

    everything-claude-code

    Also in 9 projects that use this

  • mryeastcodeSkill

    everything-claude-code-conventions

    Development conventions and patterns for everything-claude-code. JavaScript project with conventional commits.

    Also in 8 projects that use this

  • tpavanipradeepSkill

    nextjs-turbopack

    Next.js 16+ and Turbopack — incremental bundling, FS caching, dev speed, and when to use Turbopack vs webpack.

    Also in 15 projects that use this

  • Indium-AI-LabsSkill

    data-pipeline-design

    Design batch or incremental ETL and ELT pipelines and warehouse models: grain, partitioning, idempotent loads, late-arriving data, data quality tests, PII controls, and cost monitoring, before building them. Use when adding or changing a dbt model, warehouse table, scheduled transformation, or ingestion job. Not for transactional schema design (use the database-design skill) or event-sourced stores (use the event-sourcing-architecture skill).

    Also in 3 projects that use this

  • thejordanleopoldSkill

    third-party-integration

    Use when integrating with external APIs, payment processors (Stripe), webhooks, or third-party services. Use when implementing API clients with authentication and rate limiting, handling webhooks with signature verification and idempotency, writing contract tests for external services, or isolating third-party dependencies behind adapters. Triggers: \"Stripe\", \"payment integration\", \"webhook\", \"third-party API\", \"external service\", \"API client\", \"integrate with\", \"signature verification\", \"idempotency key\".

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

2 rules. 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 165 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Every observable behavior of an interface becomes a contract once someone depends on it (Hyrum's Law).

    line 8
  2. Distinctive:

    Contract first: Write the request, response, error and event shapes (as types or a schema) before the implementation, and get them into the …

    line 19

“Common” is measured: kept by at least 6 in 10 of the 165 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

API design

Every observable behavior of an interface becomes a contract once someone depends on it (Hyrum's Law). Design the contract first, change it by adding, and make every operation that can be retried safe to retry.

When NOT to use

  • Internal functions with a single caller in the same module — that is ordinary code design.
  • Choosing where a seam goes for testing — that is tdd's interface-design guidance.

Process

  1. Contract first. Write the request, response, error and event shapes (as types or a schema) before the implementation, and get them into the blueprint. Name the consumers.
  2. One error format across the API: a stable machine-readable code, a human message, and optional details. Consumers branch on code, never on message.
  3. Change by adding. New optional fields, new endpoints, new event types. Renaming, retyping, removing a field or tightening validation is a breaking change → version it or run an expand/contract migration (deprecation).
  4. Validate at the edge only. The boundary turns raw input into typed values; inner layers trust their types and don't re-validate.
  5. Predictable naming. Resources as plural nouns, actions as HTTP verbs, the same name for the same concept everywhere (match the project glossary).
  6. Make illegal states unrepresentable. Discriminated unions for variants ({status: 'paid', paidAt} vs {status: 'pending'}), branded types for ids that must not be mixed.
  7. Retried operations get an idempotency key. Any operation that a client, a queue or a retry policy may deliver twice — open IDEMPOTENCY.md when you reach it.

Common Rationalizations

RationalizationReality
"Nobody uses that field"You can't see every consumer. Under Hyrum's Law someone does; prove zero consumers before removing.
"It's a small rename, clients will adapt"A rename is a delete plus an add. Every client breaks at once.
"Retries are rare, idempotency can wait"Retries happen exactly when things go wrong — timeouts, deploys, network blips — and that's when a double charge hurts.
"A timestamp in the key makes it unique"Unique per attempt, so every retry looks new. Derive the key from the intent.

Red Flags

  • A response field removed, renamed or retyped in the diff without a version or migration plan.
  • Error handling that matches on message strings.
  • Validation of the same input repeated in several layers.
  • An idempotency key built from Date.now(), a random UUID per attempt, or the request count.
  • "Check if exists, then create" without a unique constraint.

Verification

  • Contract types/schema written and listed in the blueprint, with consumers named
  • No breaking change, or a version/expand-contract step is planned
  • Errors use the shared format with a stable code
  • Every retryable operation has an intent-derived idempotency key and a test that sends it twice

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.