grill-with-docs – skill Lập trình cho agent AI
Phiên “thẩm vấn” dựa trên tài liệu — đối chiếu kế hoạch với ngôn ngữ hiện có của dự án (CONTEXT.md) và các quyết định đã ghi (docs/adr/), đồng thời cập nhật các tệp đó trực tiếp khi thuật ngữ và quyết định được làm rõ. Dùng khi người dùng muốn kiểm tra kế hoạch dựa trên thuật ngữ nghiệp vụ đã được ghi trong tài liệu, hoặc nhắc đến “grill with docs”
Cài skill grill-with-docs bằng một lệnh
Chọn công cụ bạn đang dùng, copy lệnh và dán vào terminal. Skill sẽ nằm trong thư mục skills của công cụ đó và sẵn sàng dùng ở phiên làm việc tiếp theo.
Claude Code
mkdir -p ~/.claude/skills && curl -fsSL https://skill.dn1nguoi.com/z/grill-with-docs-7631e.tgz | tar xz -C ~/.claude/skills
Gọi trong Claude Code: /grill-with-docs – rồi mô tả công việc bạn muốn làm.
Codex
mkdir -p ~/.codex/skills && curl -fsSL https://skill.dn1nguoi.com/z/grill-with-docs-7631e.tgz | tar xz -C ~/.codex/skills
Gọi trong Codex: $grill-with-docs – rồi mô tả công việc bạn muốn làm.
Hermes Agent
mkdir -p ~/.hermes/skills && curl -fsSL https://skill.dn1nguoi.com/z/grill-with-docs-7631e.tgz | tar xz -C ~/.hermes/skills
Gọi trong Hermes Agent: /grill-with-docs – rồi mô tả công việc bạn muốn làm.
DeepSeek
mkdir -p ~/.deepcode/skills && curl -fsSL https://skill.dn1nguoi.com/z/grill-with-docs-7631e.tgz | tar xz -C ~/.deepcode/skills
Gọi trong DeepSeek: /grill-with-docs – rồi mô tả công việc bạn muốn làm.
Nội dung SKILL.md
Nội dung SKILL.md gốc bằng tiếng Anh.
Grill with Docs
> Derived from [Matt Pocock's grill-with-docs](https://github.com/mattpocock/skills/tree/main/skills/engineering/grill-with-docs) (MIT, © 2026 Matt Pocock). Matt's interview discipline + docs-anchored grilling rules preserved verbatim under MIT. Additions in this repo: 3 stdlib validators (CONTEXT.md linter, ADR scanner, glossary↔code consistency check), 3 in-depth references each citing 7+ authoritative sources, cs-grill-with-docs agent, /cs-grill-with-docs command. See [Wrapper additions](#wrapper-additions) below.
<what-to-do>
Interview me relentlessly about every aspect of this plan until we reach a shared understanding. Walk down each branch of the design tree, resolving dependencies between decisions one-by-one. For each question, provide your recommended answer.
Ask the questions one at a time, waiting for feedback on each question before continuing.
If a question can be answered by exploring the codebase, explore the codebase instead.
</what-to-do>
<supporting-info>
Domain awareness
During codebase exploration, also look for existing documentation:
File structure
Most repos have a single context:
/
├── CONTEXT.md
├── docs/
│ └── adr/
│ ├── 0001-event-sourced-orders.md
│ └── 0002-postgres-for-write-model.md
└── src/If a CONTEXT-MAP.md exists at the root, the repo has multiple contexts. The map points to where each one lives:
/
├── CONTEXT-MAP.md
├── docs/
│ └── adr/ ← system-wide decisions
├── src/
│ ├── ordering/
│ │ ├── CONTEXT.md
│ │ └── docs/adr/ ← context-specific decisions
│ └── billing/
│ ├── CONTEXT.md
│ └── docs/adr/Create files lazily — only when you have something to write. If no CONTEXT.md exists, create one when the first term is resolved. If no docs/adr/ exists, create it when the first ADR is needed.
During the session
Challenge against the glossary
When the user uses a term that conflicts with the existing language in CONTEXT.md, call it out immediately. "Your glossary defines 'cancellation' as X, but you seem to mean Y — which is it?"
Sharpen fuzzy language
When the user uses vague or overloaded terms, propose a precise canonical term. "You're saying 'account' — do you mean the Customer or the User? Those are different things."
Discuss concrete scenarios
When domain relationships are being discussed, stress-test them with specific scenarios. Invent scenarios that probe edge cases and force the user to be precise about the boundaries between concepts.
Cross-reference with code
When the user states how something works, check whether the code agrees. If you find a contradiction, surface it: "Your code cancels entire Orders, but you just said partial cancellation is possible — which is right?"
Update CONTEXT.md inline
When a term is resolved, update CONTEXT.md right there. Don't batch these up — capture them as they happen. Use the format in [CONTEXT-FORMAT.md](./CONTEXT-FORMAT.md).
CONTEXT.md should be totally devoid of implementation details. Do not treat CONTEXT.md as a spec, a scratch pad, or a repository for implementation decisions. It is a glossary and nothing else.
Offer ADRs sparingly
Only offer to create an ADR when all three are true:
1. Hard to reverse — the cost of changing your mind later is meaningful
2. Surprising without context — a future reader will wonder "why did they do it this way?"
3. The result of a real trade-off — there were genuine alternatives and you picked one for specific reasons
If any of the three is missing, skip the ADR. Use the format in [ADR-FORMAT.md](./ADR-FORMAT.md).
</supporting-info>
Wrapper Additions
The additions below are not part of Matt's upstream skill. They operationalize the upstream's rules into deterministic, stdlib-only validators that pair naturally with the interview loop.
Workflow (with wrapper tools)
1. Pre-flight (before the first question):
- Run
scripts/context_md_linter.py CONTEXT.mdif aCONTEXT.mdexists — confirms the glossary is well-formed before grilling against it. - Run
scripts/adr_scanner.py docs/adr/ifdocs/adr/exists — surfaces numbering gaps, malformed ADRs, status-frontmatter inconsistencies. - Run
scripts/glossary_code_consistency.py --context CONTEXT.md --code src/— flags defined-but-unused terms (dead glossary) and code-only common nouns that may need definitions. Use these flags as opening grill questions.
2. During the session (Matt's rules apply):
- One question per turn, walking depth-first.
- When a term is sharpened: edit
CONTEXT.mdimmediately; re-runcontext_md_linter.pyif the edit is structural. - When an ADR is warranted: write it under
docs/adr/; re-runadr_scanner.pyto confirm numbering.
3. Closing:
- Final
glossary_code_consistency.pyrun to confirm no new orphan terms were introduced. - Summarize: terms added/refined, ADRs written, scenarios discussed, open items.
Tools (stdlib-only)
| Tool | One-line role |
|---|---|
| scripts/context_md_linter.py | Validate CONTEXT.md against the CONTEXT-FORMAT.md structure. PASS/WARN/FAIL per rule. |
| scripts/adr_scanner.py | Walk docs/adr/, check NNNN-slug.md pattern, numbering integrity, body completeness. |
| scripts/glossary_code_consistency.py | Cross-reference bold terms in CONTEXT.md against codebase usage. Flag dead glossary + code-only common nouns. |
References (citations behind each rule)
- [
references/ubiquitous_language.md](references/ubiquitous_language.md) — why a glossary belongs in source control (Evans, Vernon, Khononov, Wlaschin, Brandolini, Avram & Marinescu, Fowler) - [
references/adr_practice.md](references/adr_practice.md) — when an ADR earns its keep (Nygard, Tyree & Akerman, Zimmermann Y-statements, MADR, ThoughtWorks Radar, adr-tools, Backstage) - [
references/context_md_as_artifact.md](references/context_md_as_artifact.md) — CONTEXT.md as living artifact (Khononov on language drift, Kernighan on naming, BoundedContext bliki, Confluent on data contracts, Brandolini on EventStorming glossary)
Companion
- Agent:
cs-grill-with-docs(see../../agents/cs-grill-with-docs.md) - Command:
/cs-grill-with-docs(see../../commands/cs-grill-with-docs.md)
---
Version: 1.0.0
Derived: Matt Pocock's grill-with-docs (MIT) + this repo's wrapper
Skill liên quan
- openhands – Giao việc lập trình cho OpenHands CLI (độc lập với mô hình, LiteLLM).
- career-ops – Trung tâm chỉ huy tìm việc đa CLI: đánh giá offer, quét cổng tuyển dụng, tùy chỉnh CV…
- python-error-handling – Các mẫu xử lý lỗi Python gồm kiểm tra đầu vào, phân cấp exception và xử lý lỗi một phần…
- openhue – Điều khiển đèn Philips Hue, cảnh sáng và phòng qua OpenHue CLI.
- case-review – Đánh giá cổng chất lượng cho gói hồ sơ reverse-engineering hoặc đánh giá: mức độ sẵn…
- cc-skill-backend-patterns – Mẫu kiến trúc backend, thiết kế API, tối ưu cơ sở dữ liệu và thực hành tốt phía server…
- cc-skill-coding-standards – Tiêu chuẩn mã nguồn, thực hành tốt nhất và mẫu chung cho phát triển TypeScript…
- cc-skill-continuous-learning – Biến một phiên gỡ lỗi đã hoàn thành hoặc lần người dùng sửa lại nhiều lần thành một quy…