Contribution surface for agents and missions

[flyoung588]

Hi Jamie — Ray here, founder working in an adjacent devtools/AI-agents space. This is a genuine peer question, not a pitch.

I was reading through AGENT-11 and noticed the repo is very user-path optimized: install via secure-install.sh, restart Claude Code, then choose a workflow. The interesting part is that the real project surface seems to be the 11 specialist agents plus the 20 missions and the persistent context/handoff conventions, not just the shell installer.

One outside observation: for a project like this, the highest-value contributors probably are not “general OSS contributors.” They would need to understand Claude Code behavior, prompt/context design, and solo-founder product workflows well enough to improve an agent or mission without making the system noisier.

My question: for AGENT-11, are you intentionally keeping that agent/mission contribution surface curated for now, or have you already had to find people manually who can contribute at that level?

A short reply is plenty.

[TheWayWithin]

Hey Ray, good question and good read on the project.

Yes, intentionally curated for now. The quality bar for an agent or mission is prompt/context design fluency, not code volume, and that's a small pool of people. A mediocre agent addition makes the whole system noisier, which is worse than not having it.

That said, the door isn't closed. If someone understands Claude Code behaviour and solo-founder workflows well enough to improve a specialist or add a mission that earns its place, the path would be: open an issue first describing what the agent/mission does and why it belongs, then we'd iterate on the design before any PR. No formal contributor guide exists yet, but that's the shape it would take if demand warrants it.

Appreciate the thoughtful framing.