the winner of an Anthropic hackathon open sourced his full Claude Code setup - and this is absolute f*cking gold 68 subagents, 286 skills, 94 commands, MIT license ECC turns one lone Claude Code assistant into an entire engineering department a blueprint lands before any build, a failing test... lands before any fix, and every change gets read again by a context that never watched it get written • who does what > planning - give it one sentence, get back a plan you approve before a line of code exists > review - a fresh-context pass over your diff, with a separate reviewer per language > build repair - one dedicated fixer per toolchain, PyTorch and CUDA included > security - an OWASP sweep plus a scanner hunting injection holes in your agent config > architecture - catches design mistakes while they're still cheap to undo > domain work - database queries, ML pipelines, e2e tests, docs the security duo is the part almost nobody sets up an external OWASP audit runs four figures and a week of waiting this one finishes on your branch before lunch • what the skills cover > testing - tdd-workflow takes you red to green, eval-harness sits on top of it > language packs - Python, Go, Rust, C++, Django, Laravel, Spring Boot, Next.js > context - search-first reads the docs before writing, iterative-retrieval keeps your repo from flooding the window > shipping - Docker, CI/CD, health checks, rollbacks, migrations > beyond code - writing in your voice, market research, pitch decks fork it, cut it down, have your own version running by tomorrow start with one plan and one rules pack switching on all 286 skills at once is the fastest way to make everything worse whoever wires this in over a weekend spends the next quarter reviewing work instead of typing itshow more

68 subagents and 286 skills is impressive and terrifying at the same time 💀

So where's the repo then?

Wow is he running all this on a subscription or api?

@polydao sounds like a powerhouse! for a setup like that, keeping documentation updated is key. learned that the hard way.

Orchestration matters more than agent count. The fresh-context review and dedicated repair paths are the pieces that make parallel agents dependable, not just busy. I’d steal that pattern first.

That is a whole engineering department

68 subagents and 286 skills is a department on paper. The useful version is probably 8 of them. Which two would you turn on first: planning + review, or security + build repair?

how do you see teams adapting to this?

286 skills is the easy part, the hard part is the agent picking the right 3 before it burns the context window just reading the menu

This last line is the whole post. I keep the tools small and grant capabilities on purpose. A pile of skills is just a bigger default.

you've articulated the thing I've been failing to articulate for weeks now

this is the shift from model choice to runtime design. 68 subagents matter less than the receipts between them: who decided, what changed, and which context survived the handoff.

this has to be the sloppiest thread I've seen in months 🤣

68 个子代理，感觉我得先学会当主管。

When deploying this open-source setup, consider starting with just one plan and one rules pack. Activating all 286 skills at the same time can make the overall performance worse.

68 subagents is a fleet, so give every task one accountable coordinator. Require a handoff record—owner, inputs, artifact, next action—or parallelism just multiplies orphaned work.

the external OWASP audit takes so long and costs so much, this setup sounds like a major time-saver

这工程化程度简直是给个人开发者开了外挂

this is pretty wild. an entire engineering department in an agent. insane.

oh, it's the best tutorial, than kyou

Claude code is terrible:

Gold mine.