arena

Files (1)

• SKILL.md

SKILL.md (first lines)

# Arena

Fan out N parallel attempts at the same task. Read every candidate end to end. Pick the strongest as the base. Graft the best ideas from the others into it. Verify the synthesized result.

## Start

Open a todolist with one entry per phase before launching anything. The arena runs autonomously and the list keeps phases from silently disappearing.

1. Frame
2. Fan out
3. Cross-judge
4. Pick
5. Graft
6. Verify

## Phase A: Frame

The N candidates will receive the same prompt, so the prompt is the contract. Get it right before spawning anything.

1. State the artifact each candidate is producing.
2. Derive the rubric. State what success looks like for *this* task, then turn it into 3-6 concrete gradeable criteria. Concrete: `Adds a --dry-run flag that skips writes`. Vague: `code is correct`. The rubric is the picker's tool in Phase D; candidates only see the task.
3. Pick the runners. Use `arena-runners` from `.amp/orbstack.json` when present. Otherwise use `low`, `medium`, `high`, and `ultra`. Amp's model router supplies the available models behind those modes. Spawn more only when the arena covers additional design directions.
4. Choose the executor. Design-only candidates can run `local`. Code-writing candidates run in separate Orbs.
5. Assign a unique branch to every writing Orb. Each must commit, push, and report its branch and commit SHA. N candidates writing to the same branch is shared mutable state and fails the **separate-before-serializing-shared-state** principle test.

## Phase B: Fan out

Issue one `orbstack_run_agent` call per candidate in the same tool batch. Each receives the same task and grounding, its selected Dial `mode`, its executor, its unique branch when writing, and instructions to produce both the artifact and a short rationale.

The rationale is mandatory. Without it, the parent cannot tell whether a candidate's structure is principled or accidental, which makes Phase E grafting unreliable. Each rationale names the alternatives the candidate considered and what it rejected.

If a candidate fails to produce output, proceed with N-1 and note the dropout in the synthesis record.

## Phase C: Cross-judge

After all Phase B candidates complete, choose a judge Dial mode from `arena-cross-judge` in `.amp/orbstack.json` when present. Otherwise use `high`. Run one local, read-only judge with `orbstack_run_agent`. It sees the rubric, candidate reports, and fetched candidate commits, scores each criterion, and recommends a base with rationale. Run it in parallel with the parent's reading in Phase D, never while candidates are still writing.

## Phase D: Pick a base

Read every candidate end to end before picking. Skimming N candidates surfaces only the candidate whose surface looks most familiar.

Score each candidate against the rubric criterion by criterion, not on holistic feel. Compare against 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 on which candidate a future maintainer can extend most easily without breaking invariants. Prefer the cleaner boundary or smaller surface area when two feel tied, per the Laziness Protocol.

Record the pick and the reason in a short synthesis note alongside the base artifact, including the cross-judge's verdict.

## Phase E: Graft

Walk each losing candidate once more and identify what is worth porting into the base. The signal is 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. Don't paste mechanically. The result has to remain coherent under one mental model.

Record what was grafted, from which candidate, and what was rejected and why. The rejection notes are the highest-signal part of the record. Future readers learn from what you considered and dropped, not just what you kept.

When N candidates converge on the same shape, that is a strong agreement signal. Note the convergence in the record and ship the consensus shape. No graft is needed. When N candidates wildly diverge, Phase A was under-specified. Reframe and re-run rather than averaging the divergence.

## Phase F: Verify

The synthesized artifact has to hold up under the same scrutiny as any other output, per the **prove-it-works** principle skill. The arena does not earn you a pass.

If verification surfaces a problem the arena did not catch, either Phase A was wrong (re-frame and re-run) or one candidate caught it and you missed the graft (go back to Phase E). Don't paper over.

## Outputs

One synthesized artifact. One short synthesis note alongside, naming the base, the grafts (with source candidate), the rejections, the dropouts if any, and the verification result.

From MRL-00/orbstack, MIT.

More from MRL-00/orbstack