grill-with-docs
/grill-with-docs
Przepytuje o plan i zapisuje ustalone terminy do słownika projektu oraz uzasadnione decyzje do ADR.
- Autor
- cravingmaker ↗
- Repozytorium
- skills ↗
- Licencja
- MIT
- Ostatnia zmiana repozytorium
- 10.11.2026
Ocena i bezpieczeństwo
- Pliki w paczce
- 23
- Skrypty
- 0
- Komentarze HTML
- 0
- Adresy URL w SKILL.md
- 1
Brak sygnałów. Repozytorium zmienione 10.11.2026, bez odwołań do wycofanych modeli i API.
MIT, otwarta: można używać i zmieniać, z zachowaniem informacji o autorze.
Co robi i kiedy
Przepytuje o plan i zapisuje ustalone terminy do słownika projektu oraz uzasadnione decyzje do ADR.
Oryginalny opis (EN)
Interview the user about a plan or design while recording the resolved terms in the project glossary and justified decisions as ADRs at the configured paths. Use when the user invokes grill-with-docs or asks to be grilled with the glossary and ADRs updated as decisions land.
Jak działa
- 1. Ground the interview
- 2. Interview and record
- 3. Optional builder check
- 4. Close
- Out of scope
- Provenance
Podgląd SKILL.md
--- name: grill-with-docs description: Interview the user about a plan or design while recording the resolved terms in the project glossary and justified decisions as ADRs at the configured paths. Use when the user invokes grill-with-docs or asks to be grilled with the glossary and ADRs updated as decisions land. license: MIT metadata: requires: "grilling domain-modeling" source: "https://github.com/mattpocock/skills/blob/49dd158d1076134a641b33efb035946536778336/skills/engineering/grill-with-docs/SKILL.md" source-revision: 49dd158d1076134a641b33efb035946536778336 source-license: MIT, see LICENSE-mattpocock and LICENSE-superpowers source-files: "skills/engineering/grill-with-docs/SKILL.md" local-changes: "Replaced Skill-tool calls with reads of the owned skills; reads the domain setup and existing decisions first; added Superpowers context grounding and an optional budgeted builder check; dropped the brainstorming hard gate, design approval, visual companion, and second spec pipeline." --- # Grill with docs Run the `grilling` interview with the `domain-modeling` discipline active, so the glossary and ADRs grow as decisions land. Read both skills before starting; this skill only says how they combine. ## 1. Ground the interview 1. Read `docs/agents/domain.md` for the glossary and ADR locations. Missing glossary or ADR files are normal; the `domain-modeling` skill creates them when the first entry is written. 2. If `docs/agents/domain.md` itself is missing, say so and name the `setup-agent-skills` skill as the step that records the paths. Offer to continue as a plain `grilling` interview, with no doc writes until setup exists or the user names the paths. 3. Read the glossary entries and ADRs in the area under discussion, the files the user pointed at, and the code they touch. Keep reading until you can explain how the relevant parts work today and what the change would touch. 4. Never ask what the project or the conversation already answers, and never reopen a decision the user or an accepted ADR already made. Raise a conflict with an ADR as a conflict, citing it. ## 2. Interview and record Work the `grilling` rounds. As each round lands: - Challenge and sharpen terms against the glossary, as the `domain-modeling` skill describes. - Write each resolved term to the configured glossary when it settles, within the task's write authority; otherwise include the proposed entry in the close. - Offer an ADR only when a decision meets all three `domain-modeling` criteria. Write it when the user accepts and the task permits doc writes; otherwise return the proposed ADR. Use the task's existing authority. A request to keep the glossary and ADRs updated authorizes those writes at the configured paths; loading this skill adds no authority. Preserve read-only and draft-only limits. Commits, product code, and tracker writes need their own authority. ## 3. Optional builder check When the settled understanding will be handed on to be built (to the `to-spec` skill or to another builder), and a fresh reader would add value, run one builder check on the close. Follow [references/builder-check.md](references/builder-check.md). A subagent start counts against the run's shared budget; without a subagent or budget, read the close as that builder yourself. ## 4. Close Post the `grilling` close, then list every glossary and ADR change with its path. Name the `to-spec` skill as the next step when the result is a build. Done when the interview closes under the `grilling` skill's scope rule and every resolved term and accepted ADR is written at its configured path or returned as a proposal when doc writes are outside the task's authority. ## Out of scope - A gate that blocks all building until a design document is approved. - A second spec or plan document; the `to-spec` skill writes the spec. - A browser companion or local server. ## Provenance Adapted upstream files, their hashes, and the local changes: [references/SOURCES.md](references/SOURCES.md).
Pliki w paczce (23)
- LICENSE-mattpocock 1 KB
- LICENSE-superpowers 1 KB
- SKILL.md 4 KB
- agents/openai.yaml 1 KB
- evals/examples.md 2 KB
- evals/results/2026-11-08-codex-gpt-6.1-sol-activation-1-core5fix.md 5 KB
- evals/results/2026-11-08-codex-gpt-6.1-sol-activation-2-core5fix.md 5 KB
- evals/results/2026-11-08-codex-gpt-6.1-sol-activation-3-core5fix.md 5 KB
- evals/results/2026-11-08-codex-gpt-6.1-sol-near-miss-1-core5fix.md 5 KB
- evals/results/2026-11-08-codex-gpt-6.1-sol-near-miss-2-core5fix.md 5 KB
- evals/results/2026-11-08-codex-gpt-6.1-sol-near-miss-3-core5fix.md 5 KB
- evals/results/2026-11-09-claude-claude-opus-5-5-activation-1-claude.md 4 KB
- evals/results/2026-11-09-claude-claude-opus-5-5-activation-2-claude.md 4 KB
- evals/results/2026-11-09-claude-claude-opus-5-5-activation-3-claude.md 4 KB
- evals/results/2026-11-09-claude-claude-opus-5-5-near-miss-1-claude.md 4 KB
- evals/results/2026-11-09-claude-claude-opus-5-5-near-miss-2-claude.md 4 KB
- evals/results/2026-11-09-claude-claude-opus-5-5-near-miss-3-claude.md 4 KB
- evals/results/2026-11-09-claude-claude-opus-5-5-outcome-1-claude.md 5 KB
- evals/results/2026-11-09-claude-claude-opus-5-5-outcome-2-claude.md 5 KB
- evals/results/2026-11-09-claude-claude-opus-5-5-outcome-3-claude.md 4 KB
- evals/results/2026-11-09-claude-claude-opus-5-5-outcome-4-claude.md 5 KB
- references/SOURCES.md 3 KB
- references/builder-check.md 2 KB
Wersje
| Repozytorium | Wynik | Licencja |
|---|---|---|
| Ta strona · cravingmaker/skills | 90 | MIT |
| kriegcloud/beep-effect | 100 | Apache-2.0 |
| drosam/agents | 100 | MIT |
| drvoss/everything-copilot-cli | 100 | MIT |
| dlebed/rotisserie | 100 | MIT |
| alirezarezvani/claude-skills | 100 | MIT |
| khaledsaeed18/dotclaude | 100 | MIT |
| ayoubben18/ab-method | 100 | MIT |
| lttr/claude-marketplace | 100 | MIT |
Migawka bazy z 13.11.2026 · wersja paczki 115d727950cc (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
Przepytuje plan lub decyzję i buduje glosariusz projektu oraz ADR w miarę ustalania kolejnych decyzji. - 90grill-and-records
Przeprowadza rozmowę o planie i zapisuje ustalone terminy w słowniku oraz decyzje w ADR. - 100grill-me
Testuje plan lub decyzję pytaniami o wybór opcji, a większe decyzje zapisuje jako dokumenty, na przykład ADR i terminy słownika. - 90grill-me
Wypytuje użytkownika pytanie po pytaniu, aż plan będzie rozstrzygnięty, i zapisuje trwałe decyzje w plikach projektu lub ADR. - 100decisions
Zapisuje decyzje architektoniczne jako krótkie ADR w docs/adr i prowadzi słownik pojęć projektu w GLOSSARY.md. - 100grill-docs
Przepytuje użytkownika o plan i zapisuje przebieg: transkrypt pytań i odpowiedzi w katalogu grills oraz każdą decyzję w dokumencie decyzji repozytorium.