grill-with-docs

name	grill-with-docs
description	Grilling session that challenges your plan against the existing domain model, sharpens terminology, and updates documentation (CONTEXT.md, ADRs) inline as decisions crystallise. Use when user wants to stress-test a plan against their project's language and documented decisions, or during spec/planning when domain clarity matters.

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.

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?"

Surface integration contracts

When the discussion touches an integration surface — an API being consumed, a schema being conformed to, an event format being produced — ask where the source of truth lives. "You're saying the task API accepts these fields — is that defined in a Zod schema? A types file? Where would I look to verify that?"

Capture the answer in the decisions artifact under a dedicated section or inline with the relevant decision. This gives the planner concrete source-of-truth references to ground its claims against.

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.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.

Manifestation Plan

Grill artifacts are ephemeral by default. A decision that needs to outlive this session must escape into a durable artifact before the grill closes. There are three durable destinations:

• ADR (docs/adr/NNNN-slug.md) — a cross-cutting constraint or design rule that constrains future work.
• Agent-file update (AGENTS.md, skill files) — a convention agents must follow.
• Code — the decision IS the implementation and flows through a plan to a worker.

Plans are ephemeral happy-path output — they consume a grill and produce code/ADRs/agent updates, then die.

Decision rule: "If someone deletes this grill artifact tomorrow, does this decision lose its reason for existing?"

• Yes + constrains future work → ADR
• Yes + IS the code → flows through plan to worker, no separate tracking needed
• Yes + how agents should work → agent-file update
• No → stays in the grill, dies with it

Code comments are not a tracked manifestation type. Workers decide form (comment, variable name, test name) as normal craft.

Self-containment rule: Manifestation artifacts are self-contained OR reference an ADR. They never reference a grill or a plan.

Exit gate: The grill cannot close until every resolved decision is either listed in the Manifestation Plan or explicitly marked working-memory-only with a one-line rationale.

The grill agent never edits a durable file itself; it proposes entries for the planner. Use the format in ADR-FORMAT.md.

Write a decisions artifact at the end

When the user signals they're done (or all branches are resolved), write a decisions artifact using write_artifact:

write_artifact(name: "grill/YYYY-MM-DD-<topic>.md", content: "...")

Artifact format

# Grill Session: {Topic}

**Date:** YYYY-MM-DD
**Status:** Resolved | Partially resolved

## Decisions

### {Decision 1 title}
**Question:** {The question that was explored}
**Decision:** {What was decided}
**Rationale:** {Why — in one or two sentences}

### {Decision 2 title}
...

## Manifestation Plan
- **{ADR | agent-file | code}**: {path or "(new ADR — slug TBD)"} — {one-line what + why}

## Domain Updates
- {List CONTEXT.md terms added/modified and any ADRs created, with file paths}

## Open Questions
- {Anything deferred or left unresolved}

## Key Constraints Surfaced
- {Non-obvious constraints discovered during grilling}

Keep it tight — one section per decision, no filler. The "Domain Updates" section references the CONTEXT.md changes and ADRs that were written inline during the session, so the next agent (spec or planner) knows what documentation was touched.

This document is the handoff to whoever acts on the decisions.