Zarządzanie i produkt · Planowanie
grilling
/grilling
Przepytuje plan, decyzję lub pomysł, aby sprawdzić, czy myślenie użytkownika wytrzymuje kolejne pytania.
- Autor
- yuan88964 ↗
- Repozytorium
- RelayStudio ↗
- Licencja
- MIT
- Ostatnia zmiana repozytorium
- 12.11.2026
Ocena i bezpieczeństwo
100/ 100Klasa: dobry
Format25 / 25
Frontmatter z nazwą i opisem, poprawna nazwa, opis do 1024 znaków, nazwa zgodna z folderem.
Odkrywalność25 / 25
Opis mówi, kiedy użyć skilla, ma sensowną długość i jest w 3. osobie.
Treść35 / 35
Co najmniej 150 słów i do 500 linii, nagłówki, listy kroków, przykłady albo kod.
Higiena15 / 15
Bez zaślepek typu TODO, bez ścieżek Windows, z licencją.
Ryzykobrak sygnałów w skanie
- Pliki w paczce
- 3
- Skrypty
- 0
- Komentarze HTML
- 0
- Adresy URL w SKILL.md
- 0
Przestarzałość
Brak sygnałów. Repozytorium zmienione 12.11.2026, bez odwołań do wycofanych modeli i API.
Licencja
MIT, otwarta: można używać i zmieniać, z zachowaniem informacji o autorze.
Co robi i kiedy
Przepytuje plan, decyzję lub pomysł, aby sprawdzić, czy myślenie użytkownika wytrzymuje kolejne pytania.
Wyzwalacze z opisu
grill
Oryginalny opis (EN)
Grill the user relentlessly about a plan, decision, or idea. Use when the user wants to stress-test their thinking, or uses any 'grill' trigger phrases.
Jak działa
- 0. Three-State Entry Contract (Standard / Normalize / Bailout)
- 1. Core Operator: Design-Tree Rounds & Plain-Markdown Batching
- 2. Caller-Keyed Branches (Select by Calling Container)
- 3. Subroutine Output & Return Contract (Recency Peak)
Sekcje z nagłówków SKILL.md, w oryginale · 60 linii, 640 słów
Podgląd SKILL.md
--- name: grilling description: Grill the user relentlessly about a plan, decision, or idea. Use when the user wants to stress-test their thinking, or uses any 'grill' trigger phrases. metadata: upstream: mattpocock/skills upstream-path: skills/productivity/grilling/SKILL.md upstream-commit: 4ae2317ddb22b046e26455e4f9c9b3aaaf62212d local-adaptation: "2026-11-07" --- # Relentless Design Interview (`grilling`) Shared design-tree interview subroutine that walks every open decision branch until the **frontier** is empty and shared understanding is locked. Upper container workflows (`Global_Agent_Rules`, `develop`, `to-spec`, `wayfinder`, `triage`, `improve-codebase-architecture`, `math-teaching`, `writing-fragments`, `loop-me`, `grill-with-docs`) invoke this skill to batch all currently askable questions in a single plain-Markdown turn, emit the **Settled Decision Summary** (`§3`), and immediately resume their own workflow. --- ## 0. Three-State Entry Contract (`Standard / Normalize / Bailout`) - **`Standard` (Facts verified, open decision frontier non-empty)**: Compute the current **frontier** (every decision whose prerequisites are already settled), batch all frontier questions into a single plain-Markdown message (`❓ **Qn**` + `➡️ <recommended answer>`) in the conversation body, and wait for the user's reply. - **`Normalize` (Frontier depends on checkable codebase, runtime, or document facts)**: Dispatch read-only exploration subagents (or run direct single-point checks) to verify all environment facts first—never ask the user a question that the filesystem, codebase, or docs can answer. Ask any independent frontier questions immediately while downstream fact-dependent branches wait for the subagent report. - **`Bailout` (Frontier already empty or running inside an unattended post-go-ahead execution wave)**: - *Frontier already empty*: Skip directly to `§3` and emit the **Settled Decision Summary**. - *Unattended execution wave (`develop` post-go-ahead)*: Abort interactive questioning at line 1 and emit `Error: user-visible decision required during unattended execution` + `Route: ask-user <decision>` (`Status: quarantined`). --- ## 1. Core Operator: Design-Tree Rounds & Plain-Markdown Batching Map the problem space as a **design tree** where each high-level decision unlocks the downstream decisions that hang off it: 1. **Batch the entire current frontier into one plain-Markdown message**: Output all independent frontier questions together in the conversation body in place of any host modal or popup question tool (`ask_question` / `AskUserQuestion`) and in place of one-by-one trickling: ```markdown ❓ **Q1** - **<question title>**: <question body, context, and trade-offs across options> ➡️ **推荐方案**:<your recommended default answer, worded so "yes / 同意" accepts it directly> --- ❓ **Q2** - **<question title>**: <question body, context, and trade-offs across options> ➡️ **推荐方案**:<your recommended default answer, worded so "yes / 同意" accepts it directly> ``` 2. **Advance the frontier round by round**: Each user reply settles the current tier and unblocks downstream branches. Recompute the newly exposed frontier and ask the next batched round until zero unsettled branches remain. --- ## 2. Caller-Keyed Branches (`Select by Calling Container`) | Branch | Calling Containers | Completion & Go-Ahead Behavior | | :--- | :--- | :--- | | **Branch A: Standalone / Pre-Execution Decision Gate** | Direct user `/grill-me`, `Global_Agent_Rules` pre-execution gate, `develop` (`§1 Normalize`), `to-spec` | When the frontier reaches zero, emit the **Settled Decision Summary** (`§3`) and ask the user for explicit confirmation (`"以上决策是否对齐?确认后即可开工 / 生成规格书"`) before any file edits or spec synthesis. | | **Branch B: Embedded Workflow Subroutine** | `wayfinder`, `triage`, `improve-codebase-architecture`, `math-teaching`, `writing-fragments`, `loop-me`, `grill-with-docs` | Keep rounds tightly scoped to the caller's active phase (e.g., module seam trade-offs in `improve-codebase-architecture`, thesis/story beats in `writing-fragments`, lesson scope in `math-teaching`). Once the scoped frontier is empty, emit the **Settled Decision Summary** (`§3`) and immediately resume the calling skill's next step. | --- ## 3. Subroutine Output & Return Contract (`Recency Peak`) `grilling` completes as soon as the decision frontier is empty. Close the interview by emitting this 3-line **Settled Decision Summary**: ```text 1. Resolved Decisions: <numbered list of locked architectural, behavioral, or scope choices> 2. Assumed / Deferred Constraints: <verified environment facts + explicitly out-of-scope items> 3. Next Step / Go-Ahead Gate: <confirmation prompt in Branch A, or target step resumed in Branch B> ``` **Immediate Return Rule**: Once the **Settled Decision Summary** is emitted (and confirmed in Branch A), **this subroutine is complete**—immediately return control to the calling container workflow (`develop`, `to-spec`, `wayfinder`, `triage`, `improve-codebase-architecture`, `math-teaching`, `writing-fragments`, `loop-me`, or `grill-with-docs`) at the exact step that invoked `grilling` without re-reading the caller's `SKILL.md`.
Pliki w paczce (3)
- LICENSE 1 KB
- SKILL.md 5 KB
- agents/openai.yaml 1 KB
Wersje
| Repozytorium | Wynik | Licencja |
|---|---|---|
| Ta strona · yuan88964/RelayStudio | 100 | MIT |
| nikospavlopoulos/devops-mutant-powers | 100 | MIT |
| RIJUDEVx/agent-skills | 100 | MIT |
| SimplySylvia/Claude-Quorum | 100 | MIT |
| FrancoStino/opencode-skills-collection | 100 | MIT |
| cravingmaker/skills | 100 | MIT |
| helmholtzai-consultants-munich/helmholtz-agent-skills | 100 | Apache-2.0 |
| bendrucker/claude | 100 | MIT |
| timharris707/skills | 100 | MIT |
Migawka bazy z 13.11.2026 · wersja paczki ae8c4229937b (skrót SHA-256 całej paczki)
Pobrania w ostatnich 30 dniach: 0 · ranking
Zgłoś problem z tym skillem: szkodliwa treść, błędny opis albo prośba autora o usunięcie.
Podobne skille
- 100grilling-sleek
Dokładnie przepytuje użytkownika o plan, decyzję lub pomysł, gdy chce on sprawdzić swoje myślenie. - 90grill-me
Przepytuje plan, decyzję lub pomysł bez litości, by sprawdzić myślenie przed budową. - 87grill-me
Przepytuje wytrwale użytkownika o plan, decyzję lub pomysł, by sprawdzić jego tok myślenia przed działaniem. - 82workflow-grilling
Zadaje użytkownikowi wnikliwe pytania o plan lub pomysł, żeby sprawdzić jego tok myślenia, zanim podejmie decyzję. - 100ra-grilling
Dopytuje użytkownika o plan, decyzję lub architekturę, aż każda gałąź drzewa projektu zostanie rozstrzygnięta. - 90grill
Dociekliwie przepytuje użytkownika o plan lub decyzję, zanim zacznie się wdrażanie.