grey-haven-api-design
About
This skill provides Grey Haven's RESTful API design standards for FastAPI development. It covers creating endpoints with proper routes, Pydantic schemas, HTTP status codes, pagination, error handling, and OpenAPI documentation. Use it when building APIs, implementing server functions, or working with FastAPI, Pydantic, validation, and authentication.
Quick Install
Claude Code
Recommended/plugin add https://github.com/greyhaven-ai/claude-code-configgit clone https://github.com/greyhaven-ai/claude-code-config.git ~/.claude/skills/grey-haven-api-designCopy and paste this command in Claude Code to install this skill
Documentation
Grey Haven API Design Standards
RESTful API design for FastAPI backends and TanStack Start server functions.
Follow these standards when creating API endpoints, defining schemas, and handling errors in Grey Haven projects.
Supporting Documentation
- examples/ - Complete endpoint examples (all files <500 lines)
- fastapi-crud.md - CRUD endpoints with repository pattern
- pydantic-schemas.md - Request/response schemas
- tanstack-start.md - Server functions
- pagination.md - Pagination patterns
- testing.md - API testing
- reference/ - Configuration references (all files <500 lines)
- fastapi-setup.md - Main app configuration
- openapi.md - OpenAPI customization
- error-handlers.md - Exception handlers
- authentication.md - JWT configuration
- cors-rate-limiting.md - CORS and rate limiting
- templates/ - Copy-paste ready endpoint templates
- checklists/ - API design and security checklists
Quick Reference
RESTful Resource Design
URL Patterns:
- ✅
/api/v1/users(plural nouns, lowercase with hyphens) - ✅
/api/v1/organizations/{org_id}/teams(hierarchical) - ❌
/api/v1/getUsers(no verbs in URLs) - ❌
/api/v1/user_profiles(no underscores)
HTTP Verbs:
GET- Retrieve resourcesPOST- Create new resourcesPUT- Update entire resourcePATCH- Update partial resourceDELETE- Remove resource
HTTP Status Codes
Success:
200 OK- GET, PUT, PATCH requests201 Created- POST request (resource created)204 No Content- DELETE request
Client Errors:
400 Bad Request- Invalid request data401 Unauthorized- Missing/invalid authentication403 Forbidden- Insufficient permissions404 Not Found- Resource doesn't exist409 Conflict- Duplicate resource, concurrent update422 Unprocessable Entity- Validation errors
Server Errors:
500 Internal Server Error- Unhandled exception503 Service Unavailable- Database/service down
Multi-Tenant Isolation
Always enforce tenant isolation:
# Extract tenant_id from JWT
repository = UserRepository(db, tenant_id=current_user.tenant_id)
# All queries automatically filtered by tenant_id
users = await repository.list() # Only returns users in this tenant
FastAPI Route Pattern
from fastapi import APIRouter, Depends, HTTPException, status
router = APIRouter(prefix="/api/v1/users", tags=["users"])
@router.post("", response_model=UserRead, status_code=status.HTTP_201_CREATED)
async def create_user(
user_data: UserCreate,
db: Session = Depends(get_db),
current_user: User = Depends(get_current_user),
) -> UserRead:
"""Create a new user in the current tenant."""
repository = UserRepository(db, tenant_id=current_user.tenant_id)
user = await repository.create(user_data)
return user
See examples/fastapi-crud.md for complete CRUD endpoints.
Pydantic Schema Pattern
from pydantic import BaseModel, EmailStr, Field, ConfigDict
class UserCreate(BaseModel):
"""Schema for creating a new user."""
email: EmailStr
full_name: str = Field(..., min_length=1, max_length=255)
password: str = Field(..., min_length=8)
class UserRead(BaseModel):
"""Schema for reading user data (public fields only)."""
id: str
tenant_id: str
email: EmailStr
full_name: str
created_at: datetime
model_config = ConfigDict(from_attributes=True)
See examples/pydantic-schemas.md for validation patterns.
TanStack Start Server Functions
// app/routes/api/users.ts
import { createServerFn } from "@tanstack/start";
import { z } from "zod";
const createUserSchema = z.object({
email: z.string().email(),
fullName: z.string().min(1).max(255),
});
export const createUser = createServerFn({ method: "POST" })
.validator(createUserSchema)
.handler(async ({ data, context }) => {
const authUser = await getAuthUser(context);
// Create user with tenant isolation
});
See examples/tanstack-start.md for complete examples.
Error Response Format
{
"error": "User with ID abc123 not found",
"status_code": 404
}
Validation errors:
{
"error": "Validation error",
"detail": [
{
"field": "email",
"message": "value is not a valid email address",
"code": "value_error.email"
}
],
"status_code": 422
}
See reference/error-handlers.md for exception handlers.
Pagination
Offset-based (simple):
@router.get("", response_model=PaginatedResponse[UserRead])
async def list_users(skip: int = 0, limit: int = 100):
users = await repository.list(skip=skip, limit=limit)
total = await repository.count()
return PaginatedResponse(items=users, total=total, skip=skip, limit=limit)
Cursor-based (recommended for large datasets):
@router.get("")
async def list_users(cursor: Optional[str] = None, limit: int = 100):
users = await repository.list_cursor(cursor=cursor, limit=limit)
next_cursor = users[-1].id if len(users) == limit else None
return {"items": users, "next_cursor": next_cursor}
See examples/pagination.md for complete implementations.
Core Principles
1. Repository Pattern
Always use tenant-aware repositories:
- Extract
tenant_idfrom JWT claims - Pass to repository constructor
- All queries automatically filtered
- Prevents cross-tenant data leaks
2. Pydantic Validation
Define schemas for all requests/responses:
{Model}Create- Fields for creation{Model}Read- Public fields for responses{Model}Update- Optional fields for updates- Never return password hashes or sensitive data
3. OpenAPI Documentation
FastAPI auto-generates docs:
- Add docstrings to all endpoints
- Use
summary,description,response_description - Document all parameters and responses
- Available at
/docs(Swagger UI) and/redoc(ReDoc)
See reference/openapi.md for customization.
4. Rate Limiting
Protect public endpoints:
from app.core.rate_limit import rate_limit
@router.get("", dependencies=[Depends(rate_limit)])
async def list_users():
"""List users (rate limited to 100 req/min)."""
pass
See templates/rate-limiter.py for Upstash Redis implementation.
5. CORS Configuration
Use Doppler for allowed origins:
# NEVER hardcode origins in production!
allowed_origins = os.getenv("CORS_ALLOWED_ORIGINS", "").split(",")
app.add_middleware(
CORSMiddleware,
allow_origins=allowed_origins,
allow_credentials=True,
)
See reference/cors-rate-limiting.md for complete setup.
When to Apply This Skill
Use this skill when:
- ✅ Creating new FastAPI endpoints or TanStack Start server functions
- ✅ Designing RESTful resource hierarchies
- ✅ Writing Pydantic schemas for validation
- ✅ Implementing pagination, filtering, or sorting
- ✅ Configuring error response formats
- ✅ Setting up OpenAPI documentation
- ✅ Implementing rate limiting or CORS
- ✅ Designing multi-tenant API isolation
- ✅ Testing API endpoints with pytest
- ✅ Reviewing API design in pull requests
- ✅ User mentions: "API", "endpoint", "REST", "FastAPI", "Pydantic", "server function", "OpenAPI", "pagination", "validation"
Template References
These API design patterns come from Grey Haven's actual templates:
- Backend:
cvi-backend-template(FastAPI + SQLModel + Repository Pattern) - Frontend:
cvi-template(TanStack Start server functions)
Critical Reminders
- Repository pattern - Always use tenant-aware repositories for multi-tenant isolation
- Pydantic schemas - Never return password hashes or sensitive fields in responses
- HTTP status codes - 201 for create, 204 for delete, 404 for not found, 422 for validation errors
- Pagination - Use cursor-based for large datasets (better performance than offset)
- Error format - Consistent error structure with
error,detail, andstatus_codefields - OpenAPI docs - Document all parameters, responses, and errors with docstrings
- Rate limiting - Protect public endpoints with Upstash Redis (100 req/min default)
- CORS - Use Doppler for allowed origins, never hardcode in production
- JWT authentication - Extract
tenant_idfrom JWT claims for multi-tenant isolation - Testing - Use FastAPI TestClient with
doppler run --config test -- pytest
Next Steps
- Need endpoint examples? See examples/ for FastAPI CRUD and TanStack Start
- Need configurations? See reference/ for OpenAPI, CORS, error handlers
- Need templates? See templates/ for copy-paste ready endpoint code
- Need checklists? Use checklists/ for systematic API design reviews
GitHub Repository
Related Skills
evaluating-llms-harness
TestingThis Claude Skill runs the lm-evaluation-harness to benchmark LLMs across 60+ standardized academic tasks like MMLU and GSM8K. It's designed for developers to compare model quality, track training progress, or report academic results. The tool supports various backends including HuggingFace and vLLM models.
content-collections
MetaThis skill provides a production-tested setup for Content Collections, a TypeScript-first tool that transforms Markdown/MDX files into type-safe data collections with Zod validation. Use it when building blogs, documentation sites, or content-heavy Vite + React applications to ensure type safety and automatic content validation. It covers everything from Vite plugin configuration and MDX compilation to deployment optimization and schema validation.
langchain
MetaLangChain is a framework for building LLM applications using agents, chains, and RAG pipelines. It supports multiple LLM providers, offers 500+ integrations, and includes features like tool calling and memory management. Use it for rapid prototyping and deploying production systems like chatbots, autonomous agents, and question-answering services.
Algorithmic Art Generation
MetaThis skill helps developers create algorithmic art using p5.js, focusing on generative art, computational aesthetics, and interactive visualizations. It automatically activates for topics like "generative art" or "p5.js visualization" and guides you through creating unique algorithms with features like seeded randomness, flow fields, and particle systems. Use it when you need to build reproducible, code-driven artistic patterns.
