| name | grill-with-docs-skill |
| description | A relentless interview to sharpen a plan or design, which also creates docs (CONTEXT.md glossary and ADRs) as we go. User-invoked orchestrator that pairs grilling-skill with domain-modeling-skill. |
| license | Apache-2.0 |
| compatibility | opencode |
| metadata | {"audience":"developers","workflow":"planning, alignment, documentation","trigger":"explicit-only","version":"1"} |
What I do
I orchestrate a focused session that combines two engines:
grilling-skill — relentlessly interviews the user about the plan, one question at a time with a recommended answer, until the decision tree is resolved
domain-modeling-skill — captures the crystallised terminology into CONTEXT.md and records hard-to-reverse decisions as ADRs, inline as they emerge
The result is alignment AND a durable artifact: the next session (or the next agent) inherits a sharpened glossary and recorded decisions, so the work doesn't have to be re-litigated.
When to use me
Use this skill when:
- You want to stress-test a plan or design before building
- You want the alignment session to ALSO produce domain docs (glossary + ADRs)
- You're starting a non-trivial feature and want to lock down terminology first
Trigger phrases:
- "grill with docs"
- "grill me and write it up"
- "stress-test this plan and capture it"
- "interview me and document"
If you want the interview WITHOUT doc capture, use grill-me-skill instead.
Core Workflow
Step 1: Load both engines
This skill delegates all behavior to two model-invoked engines. Do not reimplement them:
grilling-skill — owns the interview loop
domain-modeling-skill — owns doc capture
Step 2: Run the grilling session
Invoke grilling-skill to conduct the interview. Apply its rules throughout:
- One question at a time, each with a recommended answer
- Explore the codebase instead of asking when possible
- Walk the decision tree branch-by-branch
Step 3: Capture docs inline during the session
As grilling-skill resolves each branch, hand off to domain-modeling-skill to:
- Challenge and sharpen terminology against the existing
CONTEXT.md
- Update
CONTEXT.md inline the moment a term is resolved
- Offer an ADR only when a decision is hard-to-reverse + surprising + a real trade-off
The two engines run interleaved, not sequentially. Doc capture happens during the interview, not after.
Step 4: Conclude
When grilling-skill signals convergence (every branch resolved, or the user wants to build), summarise:
- Terms added/changed in
CONTEXT.md
- ADRs created (if any)
- Remaining open questions (if any)
Why This Combination Matters
With a ubiquitous language, conversations among developers and expressions of the code are all derived from the same domain model.
— Eric Evans, Domain-Driven Design
Agents dropped into a project usually decode jargon as they go, using 20 words where 1 will do. A shared CONTEXT.md fixes this concisely:
- Before: "There's a problem when a lesson inside a section of a course is made 'real' (i.e. given a spot in the file system)"
- After: "There's a problem with the materialization cascade"
This concision pays off session after session: variables/functions/files are named consistently, the codebase is easier to navigate, and the agent spends fewer tokens on thinking because it has a more concise language.
Rules
- Do not reimplement the engines. Delegate to
grilling-skill and domain-modeling-skill. This skill is an orchestrator, nothing more.
- Capture inline, not in a batch. Docs are written as terms/decisions crystallise, not collected and dumped at the end.
- Respect the user's appetite. If they want a shorter session, stop grilling earlier. Don't grill past their signal.
- Don't force ADRs. The three-criteria gate (hard-to-reverse + surprising + real trade-off) is enforced by
domain-modeling-skill. Most sessions create zero ADRs, and that's correct.
Integration with Other Skills
| Skill | Integration |
|---|
grilling-skill | The interview engine I orchestrate |
domain-modeling-skill | The doc-capture engine I orchestrate |
ticket-plan-workflow-skill | A grilled, documented plan feeds cleanly into ticket/branch/PLAN creation |
plan-execution-skill | The resolved plan + glossary give execution a precise vocabulary |
continuous-learning-skill | ADRs and glossary entries become durable learnings |
Example Usage
"Grill with docs — we're adding a notifications module"
The skill will:
- Load
grilling-skill and domain-modeling-skill.
- Begin the interview: "First — push, pull, or hybrid notifications? I recommend hybrid. Agree?"
- As terms resolve (e.g. "notification" vs "alert" vs "digest"), update
CONTEXT.md inline.
- If a hard-to-reverse decision emerges (e.g. "we'll use a message bus, not polling"), offer an ADR.
- Conclude with a summary of docs written and any open questions.
References
grilling-skill - Interview engine
domain-modeling-skill - Doc-capture engine