Skip to content
FunCoding

Search

Search docs, Skills and MCP

api-patterns

API design, versioning, testing, schema validation, and contract testing patterns for REST and GraphQL APIs.

测试531skills/api-patterns/SKILL.md

Install

Send this to Claude Code, Codex or Cursor. The agent checks the Skill for safety first and installs it only after you confirm.

读取 https://funcoding.ai/skills/vibeeval/vibecosystem/api-patterns/install.md ,按里面的步骤帮我安装这个 Skill。

SKILL.md

API Patterns

REST and GraphQL API design patterns for consistent, versioned, and well-tested interfaces.

API Versioning

URL-Based Versioning

// routes/v1/markets.ts
// routes/v2/markets.ts
// URL: /api/v1/markets, /api/v2/markets

// Express router setup
import { Router } from 'express'

const v1Router = Router()
const v2Router = Router()

app.use('/api/v1', v1Router)
app.use('/api/v2', v2Router)

// Deprecation header middleware
function deprecationWarning(version: string, sunsetDate: string) {
  return (_req: Request, res: Response, next: NextFunction) => {
    res.setHeader('Deprecation', 'true')
    res.setHeader('Sunset', sunsetDate)
    res.setHeader('Link', `</api/v${parseInt(version) + 1}>; rel="successor-version"`)
    next()
  }
}

v1Router.use(deprecationWarning('1', 'Sat, 01 Jan 2027 00:00:00 GMT'))

Header-Based Versioning

// Accept: application/vnd.api+json;version=2
function versionMiddleware(req: Request, res: Response, next: NextFunction) {
  const accept = req.headers['accept'] || ''
  const match = accept.match(/version=(\d+)/)
  req.apiVersion = match ? parseInt(match[1]) : 1
  next()
}

Schema Validation with Zod

Request + Response Validation

import { z } from 'zod'

// Request schema
const CreateMarketSchema = z.object({
  name: z.string().min(1).max(200),
  description: z.string().max(2000).optional(),
  category: z.enum(['sports', 'politics', 'crypto', 'tech']),
  closeAt: z.string().datetime(),
  initialLiquidity: z.number().positive().max(1_000_000)
})

// Response schema - strip internal fields
const MarketResponseSchema = z.object({
  id: z.string().uuid(),
  name: z.string(),
  category: z.string(),
  status: z.enum(['open', 'closed', 'resolved']),
  volume: z.number(),
  createdAt: z.string().datetime()
})

type CreateMarketDto = z.infer<typeof CreateMarketSchema>
type MarketResponse = z.infer<typeof MarketResponseSchema>

// Validation middleware
function validate<T>(schema: z.ZodSchema<T>) {
  return (req: Request, res: Response, next: NextFunction) => {
    const result = schema.safeParse(req.body)
    if (!result.success) {
      return res.status(400).json({
        success: false,
        error: 'Validation failed',
        code: 'VALIDATION_ERROR',
        details: result.error.flatten()
      })
    }
    req.validated = result.data
    next()
  }
}

// Usage
router.post('/markets', validate(CreateMarketSchema), async (req, res) => {
  const dto = req.validated as CreateMarketDto
  const market = await marketService.create(dto)
  const response = MarketResponseSchema.parse(market)
  res.status(201).json({ success: true, data: response })
})

Standardized Error Responses

// Always: { success, error, code, details? }
interface ApiError {
  success: false
  error: string
  code: string
  details?: unknown
  requestId?: string
}

interface ApiSuccess<T> {
  success: true
  data: T
  meta?: { total?: number; page?: number; limit?: number }
}

const ERROR_CODES = {
  VALIDATION_ERROR: 400,
  UNAUTHORIZED: 401,
  FORBIDDEN: 403,
  NOT_FOUND: 404,
  CONFLICT: 409,
  RATE_LIMITED: 429,
  INTERNAL: 500
} as const

function apiError(
  res: Response,
  code: keyof typeof ERROR_CODES,
  message: string,
  details?: unknown
): Response {
  return res.status(ERROR_CODES[code]).json({
    success: false,
    error: message,
    code,
    details,
    requestId: res.locals.requestId
  } satisfies ApiError)
}

Pagination Patterns

interface CursorPage<T> {
  items: T[]
  nextCursor: string | null
  prevCursor: string | null
  hasMore: boolean
}

async function paginateWithCursor<T extends { id: string; createdAt: Date }>(
  query: (cursor: string | null, limit: number) => Promise<T[]>,
  cursor: string | null,
  limit = 20
): Promise<CursorPage<T>> {
  // Fetch one extra to detect hasMore
  const items = await query(cursor, limit + 1)
  const hasMore = items.length > limit
  const page = hasMore ? items.slice(0, limit) : items

  return {
    items: page,
    nextCursor: hasMore ? Buffer.from(page[page.length - 1].id).toString('base64') : null,
    prevCursor: cursor,
    hasMore
  }
}

// GET /api/markets?cursor=<base64>&limit=20
router.get('/markets', async (req, res) => {
  const cursor = req.query.cursor as string | null
  const limit = Math.min(parseInt(req.query.limit as string) || 20, 100)
  const decoded = cursor ? Buffer.from(cursor, 'base64').toString() : null

  const page = await paginateWithCursor(
    (c, l) => db.market.findMany({
      take: l,
      skip: c ? 1 : 0,
      cursor: c ? { id: c } : undefined,
      orderBy: { createdAt: 'desc' }
    }),
    decoded,
    limit
  )

  res.json({ success: true, ...page })
})

Offset Pagination (Simple use cases)

interface OffsetPage<T> {
  items: T[]
  total: number
  page: number
  limit: number
  totalPages: number
}

// GET /api/markets?page=2&limit=20

Rate Limiting

Token Bucket with Redis

import Redis from 'ioredis'

const redis = new Redis(process.env.REDIS_URL!)

async function tokenBucket(
  key: string,
  capacity: number,
  refillRate: number   // tokens per second
): Promise<{ allowed: boolean; remaining: number; resetIn: number }> {
  const now = Date.now()
  const bucketKey = `ratelimit:${key}`

  const script = `
    local key = KEYS[1]
    local capacity = tonumber(ARGV[1])
    local refill_rate = tonumber(ARGV[2])
    local now = tonumber(ARGV[3])

    local bucket = redis.call('HMGET', key, 'tokens', 'last_refill')
    local tokens = tonumber(bucket[1]) or capacity
    local last_refill = tonumber(bucket[2]) or now

    local elapsed = (now - last_refill) / 1000
    tokens = math.min(capacity, tokens + elapsed * refill_rate)

    if tokens >= 1 then
      tokens = tokens - 1
      redis.call('HMSET', key, 'tokens', tokens, 'last_refill', now)
      redis.call('EXPIRE', key, math.ceil(capacity / refill_rate) + 1)
      return {1, math.floor(tokens)}
    else
      return {0, 0}
    end
  `

  const [allowed, remaining] = await redis.eval(
    script, 1, bucketKey, capacity, refillRate, now
  ) as [number, number]

  return {
    allowed: allowed === 1,
    remaining,
    resetIn: allowed ? 0 : Math.ceil(1 / refillRate)
  }
}

function rateLimitMiddleware(capacity: number, refillRate: number) {
  return async (req: Request, res: Response, next: NextFunction) => {
    const key = req.user?.id || req.ip || 'anonymous'
    const result = await tokenBucket(key, capacity, refillRate)

    res.setHeader('X-RateLimit-Limit', capacity)
    res.setHeader('X-RateLimit-Remaining', result.remaining)

    if (!result.allowed) {
      res.setHeader('Retry-After', result.resetIn)
      return apiError(res, 'RATE_LIMITED', 'Too many requests')
    }
    next()
  }
}

API Endpoint Testing

import { describe, it, expect, beforeAll, afterAll } from 'vitest'
import supertest from 'supertest'
import { app } from '../app'
import { db } from '../db'

const request = supertest(app)

describe('POST /api/v1/markets', () => {
  let authToken: string

  beforeAll(async () => {
    authToken = await getTestToken()
  })

  it('creates a market with valid payload', async () => {
    const payload = {
      name: 'Will BTC reach 100k?',
      category: 'crypto',
      closeAt: new Date(Date.now() + 86400000).toISOString(),
      initialLiquidity: 1000
    }

    const res = await request
      .post('/api/v1/markets')
      .set('Authorization', `Bearer ${authToken}`)
      .send(payload)
      .expect(201)

    expect(res.body.success).toBe(true)
    expect(res.body.data).toMatchObject({
      id: expect.any(String),
      name: payload.name,
      status: 'open'
    })
  })

  it('rejects invalid payload with 400', async () => {
    const res = await request
      .post('/api/v1/markets')
      .set('Authorization', `Bearer ${authToken}`)
      .send({ name: '' })   // invalid
      .expect(400)

    expect(res.body.success).toBe(false)
    expect(res.body.code).toBe('VALIDATION_ERROR')
  })

  it('returns 401 without auth token', async () => {
    const res = await request.post('/api/v1/markets').send({}).expect(401)
    expect(res.body.code).toBe('UNAUTHORIZED')
  })
})

Breaking Change Detection Checklist

Before releasing a new API version, verify:

BREAKING changes (require version bump):
  [ ] Removed a field from response
  [ ] Changed field type (string → number)
  [ ] Renamed a field
  [ ] Changed HTTP method
  [ ] Removed an endpoint
  [ ] Changed required → optional (ok) vs optional → required (breaking)
  [ ] Changed error codes/format

NON-BREAKING changes (safe to ship):
  [ ] Added new optional fields to response
  [ ] Added new optional query parameters
  [ ] Added new endpoints
  [ ] Added new enum values (check client handling)
  [ ] Relaxed validation rules

OpenAPI Spec Generation

// Use zod-to-openapi or tsoa
import { extendZodWithOpenApi } from 'zod-to-openapi'
import { z } from 'zod'

extendZodWithOpenApi(z)

const MarketSchema = z.object({
  id: z.string().uuid().openapi({ example: 'abc-123' }),
  name: z.string().openapi({ example: 'Will BTC hit 100k?' }),
  status: z.enum(['open', 'closed', 'resolved'])
}).openapi('Market')

// Auto-generate spec at /api/docs.json

Plan-Based Authorization

Tier-Aware Middleware

enum PlanTier {
  FREE = 'free',
  PRO = 'pro',
  ENTERPRISE = 'enterprise'
}

interface PlanLimits {
  tier: PlanTier
  rateLimit: number          // requests per minute
  maxItems: number           // max resources
  features: Set<string>      // enabled features
}

const PLAN_LIMITS: Record<PlanTier, PlanLimits> = {
  [PlanTier.FREE]:       { tier: PlanTier.FREE, rateLimit: 60, maxItems: 100, features: new Set(['read']) },
  [PlanTier.PRO]:        { tier: PlanTier.PRO, rateLimit: 600, maxItems: 10_000, features: new Set(['read', 'write', 'export']) },
  [PlanTier.ENTERPRISE]: { tier: PlanTier.ENTERPRISE, rateLimit: 6000, maxItems: Infinity, features: new Set(['read', 'write', 'export', 'audit', 'sso']) },
}

function requireFeature(feature: string) {
  return (req: Request, res: Response, next: NextFunction) => {
    const plan = PLAN_LIMITS[req.user.planTier as PlanTier]
    if (!plan.features.has(feature)) {
      return apiError(res, 'FORBIDDEN', `Feature "${feature}" requires ${PlanTier.PRO} plan or higher`)
    }
    next()
  }
}

function requireQuota(countFn: (userId: string) => Promise<number>) {
  return async (req: Request, res: Response, next: NextFunction) => {
    const plan = PLAN_LIMITS[req.user.planTier as PlanTier]
    const current = await countFn(req.user.id)
    if (current >= plan.maxItems) {
      return apiError(res, 'FORBIDDEN', `Plan limit reached (${plan.maxItems} items). Upgrade to increase.`)
    }
    next()
  }
}

// Usage
router.post('/exports', requireFeature('export'), async (req, res) => { /* ... */ })
router.post('/projects', requireQuota(countUserProjects), async (req, res) => { /* ... */ })

Serverless Rate Limiting

Sliding Window without Redis

// In-memory sliding window — single-instance only
// WARNING: Resets on cold start (serverless). For production multi-instance,
// use Redis/Upstash. Add MAX_ENTRIES cap to prevent memory exhaustion from IP flooding.
const windows = new Map<string, number[]>()

function slidingWindowRateLimit(
  key: string,
  maxRequests: number,
  windowMs: number
): { allowed: boolean; remaining: number; retryAfter: number } {
  const now = Date.now()
  const windowStart = now - windowMs

  // Get or create timestamps array
  const timestamps = windows.get(key) ?? []
  const valid = timestamps.filter(t => t > windowStart)

  if (valid.length >= maxRequests) {
    const oldestInWindow = valid[0]
    const retryAfter = Math.ceil((oldestInWindow + windowMs - now) / 1000)
    return { allowed: false, remaining: 0, retryAfter }
  }

  valid.push(now)
  windows.set(key, valid)
  return { allowed: true, remaining: maxRequests - valid.length, retryAfter: 0 }
}

// Cleanup stale entries periodically
setInterval(() => {
  const cutoff = Date.now() - 60_000
  for (const [key, timestamps] of windows) {
    const valid = timestamps.filter(t => t > cutoff)
    if (valid.length === 0) windows.delete(key)
    else windows.set(key, valid)
  }
}, 60_000)

API Key Authentication

import { randomBytes, createHash, timingSafeEqual } from 'crypto'

// SECURITY: For in-memory secret comparison, always use timingSafeEqual
// DB lookups by hash are safe (constant-time at DB level)

// Generate: give raw key to user, store hash
function generateApiKey(prefix: string): { raw: string; hash: string } {
  const raw = `${prefix}_${randomBytes(24).toString('base64url')}`
  const hash = createHash('sha256').update(raw).digest('hex')
  return { raw, hash }
}

// Verify: hash incoming key, compare with stored hash
async function verifyApiKey(rawKey: string): Promise<ApiKeyRecord | null> {
  const hash = createHash('sha256').update(rawKey).digest('hex')
  return db.apiKey.findFirst({
    where: { hash, revokedAt: null, expiresAt: { gt: new Date() } }
  })
}

// Middleware
async function apiKeyAuth(req: Request, res: Response, next: NextFunction) {
  const key = req.headers['x-api-key'] as string
  if (!key) return apiError(res, 'UNAUTHORIZED', 'API key required')

  const record = await verifyApiKey(key)
  if (!record) return apiError(res, 'UNAUTHORIZED', 'Invalid or expired API key')

  // Scope check
  if (!record.scopes.includes(req.method.toLowerCase())) {
    return apiError(res, 'FORBIDDEN', 'API key lacks required scope')
  }

  req.user = { id: record.ownerId, keyId: record.id }
  next()
}

Usage Metering and Quota Management

interface UsageRecord {
  tenantId: string
  metric: string      // 'api_calls' | 'storage_bytes' | 'ai_tokens'
  value: number
  period: string      // '2026-03' (monthly bucket)
}

async function trackUsage(tenantId: string, metric: string, increment: number): Promise<void> {
  const period = new Date().toISOString().slice(0, 7) // YYYY-MM
  await db.usage.upsert({
    where: { tenantId_metric_period: { tenantId, metric, period } },
    update: { value: { increment } },
    create: { tenantId, metric, period, value: increment }
  })
}

async function checkQuota(tenantId: string, metric: string, limit: number): Promise<boolean> {
  const period = new Date().toISOString().slice(0, 7)
  const usage = await db.usage.findUnique({
    where: { tenantId_metric_period: { tenantId, metric, period } }
  })
  return (usage?.value ?? 0) < limit
}

// Middleware
function meteringMiddleware(metric: string, increment = 1) {
  return async (req: Request, res: Response, next: NextFunction) => {
    const plan = PLAN_LIMITS[req.user.planTier as PlanTier]
    const withinQuota = await checkQuota(req.user.tenantId, metric, plan.rateLimit * 60 * 24)
    if (!withinQuota) {
      return apiError(res, 'RATE_LIMITED', `Monthly ${metric} quota exceeded. Upgrade plan.`)
    }
    await trackUsage(req.user.tenantId, metric, increment)
    next()
  }
}

Remember: Consistent versioning and validation contracts make APIs maintainable across client teams and breaking-change deployments.

Similar Skills

skill-creator
anthropics/skills180k

skill-creator

Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.

Testing

ponytail-audit
DietrichGebert/ponytail158k

ponytail-audit

Quality audit of a whole repo: bugs, security holes, what breaks under real load, risky code without tests, slow paths, and what to delete, merge or split. Ranked, each finding explained in plain English. One-shot report, changes nothing. Use for "audit this codebase", "review the whole repo", "find bloat", "what can I delete", /ponytail-audit.

Testing

ponytail-audit
DietrichGebert/ponytail158k

ponytail-audit

Quality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first.

Testing

ponytail-review
DietrichGebert/ponytail158k

ponytail-review

Quality review of a diff: bugs, security, real load, missing tests, speed, and what to delete. Each finding says what goes wrong and how to fix it.

Testing

ci-cd-and-automation
addyosmani/agent-skills103k

ci-cd-and-automation

Automates CI/CD pipeline setup. Use when setting up or modifying build and deployment pipelines. Use when you need to automate quality gates, configure test runners in CI, or establish deployment strategies.

Testing

idea-refine
addyosmani/agent-skills103k

idea-refine

Refines raw ideas into sharp, actionable concepts through structured divergent and convergent thinking. Use when an idea is still vague, when you need to stress-test assumptions before committing to a plan, or when you want to expand options before converging on one. Triggers on "ideate", "refine this idea", or "stress-test my plan".

Testing