Show HN: Claude Code Subagents – 100 domain-expert helpers for dev workflows
github.comI built a collection of 100+ specialized subagents for Claude Code that act like on-call experts (Python, React, Postgres, Docker, Stripe, Kafka, Prometheus, etc.). They can be auto-invoked by context or explicitly called (“use the postgres-expert to design indexes”), and each carries a focused system prompt + quality checklist. The runner picks a Claude model per task to balance speed/cost.
- Covers languages, frameworks, DBs/ORMs, DevOps, testing, ML, observability, auth, payments, messaging, and more. - Zero server: they’re just files in ~/.claude/agents/. - MIT-licensed and contributions welcome.
I’d love feedback on missing domains, rough edges in prompts, and real-world cases where this helped (or failed). Thanks! For foundational models, I feel like this is working against the model, they already are trained as experts, they need guidance, not prose about how to be an expert # Effective steering
stack: "FastAPI + SQLAlchemy + Redis"
scale: "10k RPS, sub-50ms P99"
deployment: "K8s, multi-region"
constraints: ["async-first", "12-factor", "observability"] # Not this
python-expert: "You are an expert in advanced Python..." # This
context: "Building FastAPI backend, PostgreSQL, Redis cache, Docker deployment"
constraints: "Sub-100ms response times, 10k concurrent users"
preferences: "Async-first, type hints, structured logging" I stopped telling ai how to do their jobs a long time ago, and started context management, I get crazy better results. The only time i need to bash training in is when it doesn't know an API, then I spawn a research agent to create an updated training prompt for an API, or command, then import it as needed. Keeps the primary context window cleaner for longer.