This document defines the minimum required sections for a valid cooked artifact.
A well-done artifact MUST contain all of these sections:
## Dish
<1-2 sentence description of what we're building>
## Scope
### In Scope
- <what's included>
### Out of Scope
- <what's explicitly excluded>
### Non-goals
- <what we're intentionally not solving>Why required: Without clear scope, implementation will drift.
## Assumptions
- <assumption 1>
- <assumption 2>
## Non-goals
- <what this feature will NOT do>Why required: Unstated assumptions cause bugs. Non-goals prevent scope creep.
## Architecture Decision
### Approach
<selected approach>
### Alternatives Considered
1. **<Option A>**
- Pros: <list>
- Cons: <list>
- Decision: Rejected because <reason>
2. **<Option B>** (selected)
- Pros: <list>
- Cons: <list>
- Decision: Accepted because <reason>
### Trade-offs
- Sacrificing: <what we give up>
- Gaining: <what we get>Why required: Decisions without alternatives are not decisions.
## Patch Plan
### Files to Modify
1. `path/to/file.ts` - <what changes>
2. `path/to/another.ts` - <what changes>
### Commit Sequence
1. <first commit message>
2. <second commit message>
### High-risk Areas
- <area that needs extra attention>
### Dependencies
- <external dependency or prerequisite>Why required: Implementation without a plan is improvisation.
## QA Plan
### Test Cases
1. <happy path test>
2. <edge case test>
3. <boundary test>
### Edge Cases
- <edge case 1>
- <edge case 2>
### Acceptance Criteria
- [ ] Given <context>, when <action>, then <result>
- [ ] Given <context>, when <action>, then <result>
### Regression Checks
- <existing feature to verify>Why required: Untested code is broken code.
## Documentation Updates
### Files to Update
- `README.md` - <what to add/change>
- `docs/api.md` - <what to add/change>
### New Documentation Needed
- <topic>: <brief description>
### Migration Guide (if breaking change)
- <step 1>
- <step 2>Why required: Undocumented features don't exist to users.
## Pre-mortem (REQUIRED - 3 scenarios)
1. <what could go wrong> -> mitigation: <how we prevent/detect>
2. <what could go wrong> -> mitigation: <how we prevent/detect>
3. <what could go wrong> -> mitigation: <how we prevent/detect>
## Security Review
- Reviewed: yes
- Issues found: <list or "none">
- Risk level: low/medium/high
## Rollback Plan
1. <rollback step 1>
2. <rollback step 2>Why required: Optimism is not a risk management strategy.
A microwave artifact MUST contain:
## Problem Statement
<what's broken + how to reproduce>## Fix Plan
<minimal fix description>## Why Safe
<1 sentence explaining why this is low risk>## Tests
- <test 1: verifies the fix>- Problem statement is specific and measurable
- Scope has explicit "out of scope" items
- At least 2 alternatives were considered
- Implementation plan lists actual file paths
- At least 3 test cases defined
- At least 3 pre-mortem scenarios
- Rollback plan exists
- Problem is reproducible
- Fix is minimal (not a rewrite)
- "Why safe" is convincing
- At least 1 test verifies the fix
An artifact is INVALID if:
| Issue | Why invalid |
|---|---|
| Missing scope | Can't verify completion |
| No alternatives | Not a decision, just a guess |
| No tests | Can't verify correctness |
| No risks | Ignoring reality |
| Scope too large | Should be multiple cooks |
| "TBD" in required fields | Incomplete thinking |
- CHEF_MATRIX.md - Who produces what
- ANTI_PATTERNS.md - What NOT to do
.claude/templates/- Artifact templates