🎉 Community registrations are open — join as contributor, maintainer, early adopter, or member

[UditAkhourii]

The ADHD community is open. One short form, four roles.

If you build with coding agents, you have a place here. Sign up as a contributor, maintainer, early adopter, or just a member who wants to follow along.

👉 Join here → https://tally.so/r/WO1Nzj

What you get

• A direct line into the project. Frame contributions, eval problems, integrations, adopter onboarding — all coordinated here.
• Early access to v0.2 work: cross-model judging, anchor-stripping pre-pass, the hyperfocus companion skill, frame-fitness learning across runs.
• Co-author credit on the published paper if you contribute frames, eval problems, or surface a documented win/loss.

Where ADHD is so far

• Open-sourced and published as a preprint with the full eval methodology, 5/6 wins vs single-shot baseline, MIT.
• Featured by The New Stack.
• Adopted by repowire — its maintainer ported ADHD into the default orchestrator template (PR #313, merged with full attribution).
• Independently reviewed by @mxriverlynn with 11 sources and 8 validation rounds; every finding tracked openly as issues #16, #17, #18.
• 0.1.4 on npm as adhd-agent, installable on Claude Code, Cursor, Antigravity, Codex, Cline, Gemini CLI, Windsurf, and ~50 more via npx skills add UditAkhourii/adhd.

What we are working on next

The current open issues list is the live backlog. Highest-leverage items right now:

• #4 Anchor-stripping pre-pass (the most important architectural fix).
• #6 Cross-model critic (defuses the same-model-judging critique).
• #10 Frame-fitness learning across runs.
• #11 Hyperfocus companion skill.

If any of those is the thing you would build, sign up and say so in the form. Pairing on a real issue beats abstract collaboration every time.

---

No spam. Low-volume updates. You can unsubscribe by simply not replying to anything.

Building this in public, including the parts where it does not work.