[aaylward]

UI for muchq/MoonBase#1681 and muchq/MoonBase#1682 (muchq/MoonBase#1668 phase 4).

• A lone seat at an ordinary go table picks a bot strength and adds GNU Go: Beginner 1, Casual 4, Club 7 or Strongest 10. This mirrors chess's Stockfish picker. Not offered at a round robin pairing, to watchers, or with both seats taken.
• useGoTable.addBot(level) sends addBot {level}, held as seating until the next view or a refusal.
• A bot's seat (gnugo@4) reads "GNU Go 4" in the waiting list, on its clock, in the turn and acceptance lines, and in the result. GoPlayer.bot and addBot are in the wire types.
• After that the table plays as a two-seat one: pick a board and clock, then Start. The bot marks the dead and accepts in the scoring.

Deploy order: merge after muchq/MoonBase#1682 deploys with GO_ENGINE_URL set. Until then the hub refuses addBot with "no go engine".

Tests (written first, watched red): naming, the hook's seating hold, the picker's gating and disabled states, and bot names in the waiting list, clocks, turn line, acceptance line and result. The UI's new conditions are mutation-checked. 1676/1676; tsc and eslint clean.

The review panel ran (React, UX/a11y, wire, altitude). Its fixes are the second commit.

🤖 Generated with Claude Code

https://claude.ai/code/session_01Xae1QBa2qHX32DPVGSq19F

[cloudflare-workers-and-pages]

Deploying with    Cloudflare Workers

The latest updates on your project. Learn more about integrating Git with Workers.

Status	Name	Latest Commit	Preview URL	Updated (UTC)
✅ Deployment successful! View logs	muchq	79ff785	Commit Preview URL Branch Preview URL	Oct 20 2026, 01:05 PM

[cursor]

Review

Altitude looks right for MoonBase#1681/#1682 phase 4: lone-seat addBot{level}, GNU Go naming via gnugo@N, seating hold until view/refusal, gated off pairings/watchers/full tables, then ordinary two-seat Start. Mirrors chess’s Stockfish picker without inventing a shared bot abstraction. CI green.

Low — clock seat names aren’t pinned

PR text and the “wherever a seat is named” test claim clocks; the playing assertion only checks the status line (GNU Go 4 to move.). Clock does use seatName via the color nameOf, but nothing reads go-seat-white (chess pins Stockfish 1500 · black). A regression that left the seat label as gnugo@4 would still pass.

[aaylward]

Clock seat label: agreed, it's unpinned. It will ride this PR's next code push, if one comes, as an assertion on the bot's clock reading "GNU Go 4". It doesn't start a push on its own; the label goes through the same seatName the status-line test pins.

---

Generated by Claude Code