Before following this workflow, read the Codex runtime contract. It defines plugin paths, model configuration, delegation, and persistence.
Arena
Fan out parallel attempts, pick a base, graft the best ideas, and verify the result. Resolve role overrides through CODEX.md before each spawn.
Track Frame, Fan out, Cross-judge, Pick, Graft, and Verify before launching.
Phase A: Frame
Every candidate gets the same prompt, so the prompt is the contract.
- State the artifact each candidate is producing.
- Derive the rubric. Derive 3-6 concrete gradeable criteria for success. The rubric is the picker's tool in Phase D. Candidates only see the task.
- Pick the runners. Use
arena runnersfrom~/.codex/pstack-models.mdwhen present. Otherwise default to one each ongpt-6-astra @ xhighandgpt-6-luna @ xhigh. Spawn more when the arena covers multiple design directions. Same model N times when the work is generation-bound rather than judgment-sensitive. Pick reasoning effort for the job under the Subagents section of poteto-mode, using only the active execution tool’s supported values. - Assign output paths. Each candidate writes to its own location (a git worktree where possible, otherwise
/tmp/arena-<slug>/candidate-<n>/), per the separate-before-serializing-shared-state principle skill.
Phase B: Fan out
Spawn all N candidates asynchronously, in rolling waves bounded by the host's available capacity. Give each the task, the absolute path to the shared grounding, its own output path or worktree, and instructions to produce both the artifact and a short rationale.
If a candidate produces no output, proceed with N-1 and note the dropout.
Phase C: Cross-judge
After all Phase B candidates complete, choose one model from the arena cross-judge pool in ~/.codex/pstack-models.md when present. Otherwise use gpt-6-astra @ xhigh and gpt-6-luna @ xhigh. Prefer a different configured model family from the parent's when the host exposes one; report when the panel lacks family diversity. Spawn one read-only judge subagent on that model. It sees the rubric and the candidates by path label, scores each criterion, and recommends a base with rationale. It runs in parallel with the parent's reading in Phase D, not with the candidates themselves. Don't spawn the judge while candidates are still writing.
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. 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 the Laziness Protocol.
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 the redesign-from-first-principles principle skill, 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.
Phase F: Verify
Verify the synthesized artifact like any other output, per the prove-it-works principle skill. 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.
Outputs
One synthesized artifact, with a short synthesis note alongside naming the base and why, the cross-judge's verdict, each graft with its source candidate, what was rejected and why, any convergence or dropouts, and the verification result.
