hatch3r: AI coding agents for any repo. One command. 12+ editors.

Hey Product Hunt! I'm Vic, one of the creators of hatch3r. The backstory: We use multiple AI coding tools daily — Cursor, Claude Code, Copilot — and the config fragmentation was killing us. Every project needed the same setup dance: code standards, review workflows, board management, MCP servers. All in different formats for each tool. So we built a reusable framework internally. It worked so well that we cleared 400+ GitHub issues in 10 days with it. At that point we figured: why keep this to ourselves? What makes hatch3r different: The architecture is an adapter pattern — one canonical source in .agents/ generates native configs for 12+ editors. You're never locked into a single tool. Switch from Cursor to Windsurf tomorrow? Your agent setup comes with you. The daily workflow is just two commands: - board-fill — turns your todo.md into GitHub issues with dependency graphs and priority ordering - board-pickup — auto-selects the next issue, delegates to sub-agents, runs a review loop (reviewer → fixer → re-review, up to 3 rounds), and creates a PR Yes, there are way more commands, but those two handle 80% of what you'll do day-to-day. What we want from you: hatch3r is community-driven. We want to hear: - What agents or skills are missing for your workflow? - Which editors should we support next? - What's confusing or broken in the install flow? - How can we simplify the solution? MIT licensed. No paid tier, no catch. We just think the collective knowledge of how to configure AI coding agents should be shared openly. Try it: npm i hatch3r & npx hatch3r init