Anthropic hackathon winner open-sources Claude Code setup with 68 subagents and 286 skills
Roger_M_Taylor · x · 2026-09-16
The winner of an Anthropic hackathon open-sourced his entire Claude Code engineering setup (ECC) under MIT license: 68 subagents, 286 skills, 94 commands. The design swaps a lone assistant for a whole engineering department:
- Planning: one sentence in, a signed-off plan out before any code
- Review: fresh-context diff reads with a dedicated reviewer per language
- Build repair: a fixer per toolchain, including PyTorch and CUDA
- Security: OWASP sweeps plus a scanner hunting injection holes in agent configs
- Architecture: catches design mistakes before they become migrations
A companion article on "Loop Engineering" covers the four loop types in Claude Code, the evaluator that decides whether loops work, and four bills nobody warns you about. A directly reusable multi-agent engineering blueprint.
More from coding & agent
- Engineer uses LLMs to turn knowledge into playable simulations, builds a chip fab game — thehiphopswami · 2026-09-16
- Open-Source Skill Strips AI Slop: 41 Patterns Caught Across 4 Review Gates — Roger_M_Taylor · 2026-09-16
- Muse Code Wins Over Skeptical Dev as Alexandr Wang Touts the Model — alexandr_wang · 2026-09-16
- Figma Links Aren't Enough: What Context to Give AI Coding Agents Before They Touch UI — T07NAD0 · 2026-09-16
- prxhub MCP connector lets AI agents search, cite and publish .prx research bundles — modelcontextprotocol · 2026-09-16
- Building reliable agent harnesses: lessons from 4,000 runs of an open-source deep research agent — Public_Umpire_1099 · 2026-09-16