Arena

Open a todolist with one entry per phase (Frame, Fan out, Cross-judge, Pick, Graft, Verify) before launching anything.

Phase A: Frame

Every candidate gets the same prompt, so the prompt is the contract.

1. State the artifact each candidate is producing.
2. Derive the rubric, 3-6 concrete gradeable criteria for what success looks like on this task. It is the picker's tool in Phase D. Candidates only see the task.
3. Pick the runners through /swarm: by default its panel of two, one on the session's model and one on the counterpart. When a seat cannot run, replace it so at least two runners remain, and say so: a counterpart seat takes /swarm's same-vendor fallback on the session's model, and another seat runs on a remaining runner's model. Spawn more runners when the arena covers multiple design directions. Use the same model N times when the work is generation-bound rather than judgment-sensitive. Pick each subagent's model and effort for the job through /swarm.
4. Give each candidate its own output location (a git worktree where possible, otherwise /tmp/arena-<slug>/candidate-<n>/), per /principle-separate-before-serializing-shared-state.

Complete when the artifact, the rubric, the runners, and one output location per candidate are named.

Phase B: Fan out

Spawn all N subagents at once, in the background, through /swarm, each with the task, the path to the shared grounding, its own output path, and instructions to produce the artifact and a short rationale naming the alternatives it considered and what it rejected.

If a candidate produces no output, proceed with N-1 and note the dropout.

Complete when every candidate has written its artifact and rationale or is noted as a dropout.

Phase C: Cross-judge

After all Phase B candidates complete, spawn one readonly judge subagent on the counterpart through /swarm's cross-vendor critic route. The judge is never from the parent's vendor, so take only the counterpart routes, T3 and then the CLI, never the same-vendor critic. If the counterpart resolves to fallback or neither route runs, skip the judge, pick the base alone in Phase D, and say so. It sees the rubric and the candidates by path label, scores each criterion, and recommends a base with rationale. It runs in parallel with your reading in Phase D. Don't spawn it while candidates are still writing.

Complete when the judge has returned per-criterion scores and a recommended base, or the skip and its reason are recorded.

Phase D: Pick a base

Read every candidate end to end before picking. Score each against the rubric criterion by criterion, not on holistic feel, and compare with the cross-judge when Phase C ran one. Agreement on the base confirms the pick. Disagreement means one of you is biased or the rubric was ambiguous. Read both rationales before deciding.

Pick the base a future maintainer can extend most easily without breaking invariants. When two feel tied, prefer the cleaner boundary or smaller API, per /principle-laziness-protocol.

Complete when one base is chosen, with your scores compared with the judge's when Phase C ran one.

Phase E: Graft

Walk each losing candidate once more for what is worth porting into the base, usually one or two things per candidate, not most of it. Fold each graft in by hand, per /principle-redesign-from-first-principles, so the result stays coherent under one mental model. Don't paste mechanically.

If the candidates converge on one shape, that is strong agreement. Note it and ship the consensus shape. No graft is needed. If they wildly diverge, Phase A was under-specified. Reframe and re-run rather than averaging the divergence.

Complete when every losing candidate has been walked and each graft or rejection has a reason.

Phase F: Verify

Verify the synthesized artifact like any other output, per /principle-prove-it-works. If verification finds a problem the arena missed, either Phase A was wrong (re-frame and re-run) or a candidate caught it and you missed the graft (go back to Phase E). Don't paper over it.

Complete when the synthesized artifact passes verification.

Outputs

One synthesized artifact, with a short synthesis note alongside naming the base and why, the cross-judge's verdict or that Phase C skipped the judge, each graft with its source candidate, what was rejected and why, any convergence or dropouts, and the verification result.