grill
Skill79 starsChanged 28 days agoMITView on GitHub
Use when the user says 先討論 / wants to align on a fuzzy idea before building, or before any non-trivial implementation whose requirements are still ambiguous — interviews the user until shared understanding is confirmed, one question at a time, each with a suggested answer; facts answerable from the filesystem/code are looked up, never asked. Optionally maintains the repo's CONTEXT.md glossary when vocabulary settles during discussion. Also invoked by other skills (spec / prd-create) as their interviewing discipline. NOT for already-frozen specs (just implement) or for debugging (use diagnose).
87 other projects use a version of this.
Adopted by
Most recently taken up by
- Yesterday-AI/ytstacktheir own version, first seen Nov 8, 2026
- therealtinhtute/mono-harnesstheir own version, first seen Nov 8, 2026
- gaoyu06/harnesstheir own version, first seen Nov 8, 2026
- afawfawffaf/my-skillstheir own version, first seen Nov 8, 2026
- BSSE23047/nauraiz-devtheir own version, first seen Nov 8, 2026
- Mahesh-Bishnoi/skillstheir own version, first seen Nov 8, 2026
Used with
Projects with it also use write-skill, root-cause and 10 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
NoaiRoxSkillwrite-skill
Use when creating a new skill, editing an existing skill, or verifying a skill works before deployment.
Also in 5 projects that use this
Ingramradical235Skillroot-cause
Structured root cause analysis using WHERE (MECE decomposition) then WHY (causal structure diagrams) pipeline. Includes coin-flip detection, logical-leap detection, and vicious cycle intervention. Use when buffer enters YELLOW/RED, when metrics stall, or when Actions execute but don't produce results.
Also in 4 projects that use this
Neko-Catpital-LabsSkillprove-it
Shared evidence rule for investigating, diagnosing, debugging, proving, or explaining what happened or why for any problem or situation: do not assert something is fixed, working, passing, merged, or running unless you personally observed it this turn and can show what you saw; do not assert it was caused by X until a repro and controlled isolation prove that cause.
Also in 4 projects that use this
sahinSkillexecute
Also in 5 projects that use this
assisjpSkillfile-pr
Prepare and open or update an authorized pull request with a concise problem-first description and verifiable evidence. Use when the user asks to create, file, open, or update a PR. Do not use merely to review a diff or when publication is not authorized.
Also in 3 projects that use this
Arman-KudaibergenovSkillwrite-plan
Этот скилл MUST быть вызван когда есть design.md и нужно разбить доработку на атомарные задачи (2-5 мин) с точными путями файлов и критериями проверки в tasks.md. SHOULD также вызывать как промежуточный шаг между brainstorm и subagent-dev. Do NOT использовать для обсуждения идеи — используй brainstorm; для выполнения задач — используй subagent-dev.
Also in 3 projects that use this
Other versions87 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 87
- kirodotdev/KiroCrew★ 4.8ksame name
- crbnos/carbon★ 2.8ksame name
- JuliusBrussee/cavekit★ 1.1ksame name
- SethGammon/Citadel★ 927same name
- nagi-studio/skills★ 470same name
- alamops/skills★ 53same name
- Kanevry/session-orchestrator★ 53same name
- oakreader/oakreader★ 46same name
- manusco/resonance★ 38same name
- shaokeyibb/anti-asu-skills★ 28same name
77 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository25 other files it gives its agents.
- adrSkill
- alignSkill
- character-loraSkill
- conference-reportSkill
- ctf-kitSkill
- diagnoseSkill
- gpt-image-genSkill
- ios-resignSkill
- job-scoutSkill
- llm-benchmarkSkill
- llm-wiki-lintSkill
- md2pptSkill
- memory-lintSkill
- prd-breakdownSkill
- prd-createSkill
- prep-repoSkill
- repo-scanSkill
- review-gateSkill
- rewrite-toneSkill
- rewrite-twSkill
- searxngSkill
- skill-cronSkill
- specSkill
- workflow-routerSkill
- wrap-upSkill
The file
Files
grill — 問到對齊為止,一次一題
You are an interviewer whose only goal is shared understanding. 停止條件不是「問滿 N 題」,是「雙方對要做什麼的理解一致」。你的建議答案讓 user 可以一句「照你說的」就前進——拷問不等於把負擔丟回去。
CRITICAL — 停止句:共同理解未經 user 確認前,禁止動手實作、禁止寫任何交付物檔案。 唯讀查證(grep / 讀檔 / 跑唯讀指令)不受限,而且是義務(見 Step 2)。唯一寫入例外:Step 3 詞彙落檔——user 對該詞點頭的當下即為寫入授權。
Step 1: 宣告討論模式
一句話讓 user 知道現在是討論、不是動工:「先對齊,我一次問一題,你隨時可以喊開工。」
Step 2: 問答迴圈
每一輪:挑當下槓桿最大的一個未決點,然後照這張表分流:
| 未決點類型 | 判準 | 動作 |
| 可自查 fact | filesystem / code / git / 一條唯讀指令查得到(「這欄位存在嗎」「現在回傳什麼格式」「測試過不過」) | 自己查,不准問 user |
| User 持有的 fact | repo 查不到、只有 user 知道的事實(痛點場景 / 現場狀況 / 外部系統行為 / 未入庫的約束) | 問 user,可附「我猜是 X」當建議答案 |
| Decision | 取捨、偏好、業務判斷、優先順序(「要不要相容舊格式」「先做哪個」) | 問 user |
問 decision 的鐵律:
- 一次只問一題,等回覆再問下一題——一次丟五題,user 只會挑好答的答。
- 每題附你的建議答案 + 一句理由,讓 user 可以只回「1」或「照你說的」。
- 沿決策樹走:先問上游(會改變後續問題的),再問下游。
Step 3: 詞彙維護(選配,觸發才做)
討論中發現同一個概念出現多個叫法、或 user 用了模糊詞 → 當場處理:
- 提議一個 canonical 詞:「所以我們統一叫『孵化』?」
- User 點頭 → 當場 inline 更新 repo 根目錄
CONTEXT.md,不攢批、不留到收尾:
- **孵化**:從掉落物培育出新寵物並登錄圖鑑的整條流程。_避免_:抽卡、生成、hatch規則:lazy 建檔(第一個詞敲定才建 CONTEXT.md);只收這個專案獨有的概念(通用技術詞如 timeout、retry 不收);嚴禁實作細節——它是詞彙表,不是 spec。
Step 4: 收斂
判斷理解已對齊時,輸出共同理解摘要:
## 對齊摘要
- 要做:<一句話>
- 不做:<明確排除項>
- 關鍵決策:<D1: ...> <D2: ...>
- 未確認假設:<有就列,沒有寫「無」>User 確認摘要 → 可把已定內容整理成需求/派工草稿。準備開始非小幅實作或交給另一個 agent 執行時,先呼叫 review-gate:以草稿和既有來源為對象,採派工/需求判準,由盲審與有脈絡複核產出處置。main 修正必要問題與驗證;審查 READY 後才交棒。有執行授權就接著做,不增加一次例行確認。
要無人值守派工時,有裝 owl plugin 就交給 owl:brief 寫工單;沒裝就照 交接紀律 的最小底線,只補缺少的執行資訊。已對齊的工作不重啟訪談。缺 review-gate 或雙 agent 能力時回報能力缺口,不能把自審當已完成獨立審查。
純討論、詞彙落檔不啟動這一關。若下游 spec 已對同一版本跑過 review-gate,沿用其結果,不重複加一輪;測試、驗收與 UI QA 仍由原工作流程負責。
不阻塞條款
僅限背景 / 無人值守場景(user 本來就不在)→ 以各題的建議答案為假設繼續,但每個假設在摘要的「未確認假設」欄明文列出,不假裝那些是拍板過的。互動 session 裡 user 只是還沒回 → 等,不自行解除閘門。
Anti-patterns
- ❌ 整份問卷 — 一次丟 5+ 題編號清單(下游 skill 若自有批次問法慣例,該處從其慣例)
- ❌ 問 grep 一下就有答案的事 — fact 上桌問 user 是把查證成本外包給人
- ❌ 開放題不附建議 — 「你覺得呢?」是把思考丟回去,不是拷問
- ❌ 討論中偷跑 — 「我先順手把架子搭起來」= 違反停止句
- ❌ 詞彙記在腦中 — 敲定的詞不落 CONTEXT.md,下個 session 從零再猜一次
- ❌ 拿題數當停止條件 — 問滿三題就開工;停止條件只有「理解一致」
Important rules
- 停止條件 = 共同理解確認,不是題數
- 一次一題、附建議答案
- Fact 自己查、decision 才問人 — 本 skill 最鋒利的一條
- 詞彙敲定當場落檔,lazy 建立、只收專案獨有詞、禁實作細節
- 確認前禁動手;user 不在場則假設明文化、繼續走
- 被下游 skill 呼叫時(spec / prd-create),本 skill 管 fact/decision 切分與附建議答案;批次 vs 逐題節奏從呼叫方慣例(見 Anti-patterns 第 1 條),章節結構歸呼叫方
Acknowledgments
訪談紀律(一次一題、fact/decision 切分、shared understanding 停止條件)參考 mattpocock/skills(MIT)的 grilling,詞彙表機制參考其 domain-modeling / CONTEXT-FORMAT,中文重寫合併為單一 skill。
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

