- name
SKILL.md
技能檔案保留在發布者的儲存庫中。
在終端機中貼上複製的指令。它會將這個 SKILL.md 安裝到對應的代理中。
npx skills add backnotprop/pstack --skill arena -a cursor -y在終端機中貼上複製的指令。它會將這個 SKILL.md 安裝到對應的代理中。
npx skills add backnotprop/pstack --skill arena -a claude-code -y在終端機中貼上複製的指令。它會將這個 SKILL.md 安裝到對應的代理中。
npx skills add backnotprop/pstack --skill arena -a codex -y在終端機中貼上複製的指令。它會將這個 SKILL.md 安裝到對應的代理中。
npx skills add backnotprop/pstack --skill arena -a grok -y在 Grok Bot 聊天中傳送複製的訊息。Bot 會從 GitHub 取得這個 SKILL.md,並加入你所有 Bot 共用的技能中。
Install this skill: https://github.com/backnotprop/pstack/tree/HEAD/skills/arena來自發布者
Open a todolist with one entry per phase (Frame, Fan out, Cross-judge, Pick, Graft, Verify) before launching anything.
Every candidate gets the same prompt, so the prompt is the contract.
arena runners line in the pstack settings file (~/.cursor/rules/pstack-models.mdc in Cursor, ~/.claude/rules/pstack-models.md in Claude Code, ~/.agents/pstack-models.md in other harnesses). If the file or that line is missing, default to one each on claude-fable-5-2-high, claude-opus-5-5-high-fast, and gemini-4-carbon-high. An auto or inherit-parent entry in this line means the parent model, so omit model for it. If your subagent tool rejects a configured entry, run that seat on its family's default and say so. Families go by prefix, and the longest matching prefix wins: claude-fable-*, claude-*, and gemini-*. Drop a leading cursor- before you compare. Any other slug's family is its text before the first hyphen. A family's default is its entry in the default list above. If the family has no default, or your subagent tool rejects that default, use the closest valid slug of the same family from its error message or your harness's model list, or drop the seat and say so when there is none. If dropped seats leave fewer than two runners, add runners until there are two, or one when the line has one entry. Run each added runner on a remaining runner's model, or omit model so it runs on the parent model when none remain, and say so. 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 reasoning level and fast or standard speed for the job, per the Subagents section of the poteto-mode skill./tmp/arena-<slug>/candidate-<n>/), per the separate-before-serializing-shared-state principle skill.Spawn all N subagents in one message with run_in_background: true, 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.
After all Phase B candidates complete, choose one model from the arena cross-judge pool line in the pstack settings file (~/.cursor/rules/pstack-models.mdc in Cursor, ~/.claude/rules/pstack-models.md in Claude Code, ~/.agents/pstack-models.md in other harnesses). If the file or that line is missing, choose from claude-fable-5-2-high, claude-opus-5-5-high-fast, and gemini-4-carbon-high. Spawn one readonly judge subagent. The judge is never from the parent's vendor. A vendor is a family as Phase A step 3 defines families, except that claude-fable-* and claude-* are one vendor. Try only the entries from another vendor, in pool order. An auto or inherit-parent entry is the parent model, so it never judges. With the default pool, a Claude parent gets gemini-4-carbon-high. A parent from any other vendor gets claude-fable-5-2-high first, then claude-opus-5-5-high-fast. If your subagent tool rejects an entry, use a valid slug of the same model at the closest level or speed from the error message or your harness's model list, or move to the next entry when there is none. If no entry from another vendor runs, try the default pool's entries from another vendor the same way. If none of those runs, skip the judge, pick the base alone in Phase D, and say so. If you do not know the parent's model, try the pool's real slugs in pool order and say that the vendor check did not run. The judge 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.
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 the Laziness Protocol.
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.
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.
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.
Spawn N parallel candidates at the same task, pick a base, graft the strongest parts of the losers into it. Use for /arena, 'arena this', 'throw it in the arena', or when one attempt at a non-trivial
---name: arenadescription: Spawn N parallel candidates at the same task, pick a base, graft the strongest parts of the losers into it…---# skills/arena/SKILL.md總覽
Spawn N parallel candidates at the same task, pick a base, graft the strongest parts of the losers into it. Use for /arena, 'arena this', 'throw it in the arena', or when one attempt at a non-trivial artifact would lock in the wrong shape.
常見問題
從發布者下載 skills/arena/SKILL.md,或複製路徑後開啟儲存庫。PlugMCP 只收錄這項技能,不代管檔案。
發布者將它存放在 https://github.com/backnotprop/pstack。
Claude、Cursor、Codex 和 Grok Build 可以讀取你從發布者儲存庫加入的 SKILL.md。在聊天中傳送技能的 GitHub 資料夾連結,Grok Bot 就會安裝它。