FastAPI
Your agent now scaffolds and reviews FastAPI applications against production-ready templates — async patterns, dependency injection, Pydantic models, middleware, and comprehensive error handling.
What it installs
| Artifact | Path (in your project) | Purpose |
|---|---|---|
| Stack ref | (inherits python bundle) | Python coding style, typing, testing, and security rules |
| Skill | .agents/skills/fastapi-templates/ | Protocol the agent follows when scaffolding or extending FastAPI projects — sourced from wshobson/agents |
| Rule | .claude/rules/fastapi.md | Always-loaded pointer: triggers on any FastAPI route, dependency, or model work |
How it works
FastAPI projects age badly when the early decisions are wrong — sync routes inside an async stack, dependency scopes chosen by accident, error handlers glued on after the fact. The skill encodes a production-shaped template: consistent project layout, properly async routes + DB calls, typed Depends() with clear scope, Pydantic v2 models with validation, middleware for cross-cutting concerns, and a unified exception handler.
The rule loads a pointer into every session, so the agent consults the skill before writing a new router, Depends, or Pydantic model — no explicit invocation needed.
Setup
No env vars or external accounts required. The skill is fetched from GitHub during harness-kit add via npx skills add.
Pairs well with
postgresql— async SQLAlchemy / asyncpg patterns that match the template's async designsupabase— when Postgres + auth + storage is your backendcode-review-gates— enforce async discipline and Pydantic validation on every commit