Initial commit

This commit is contained in:
2026-08-19 11:03:16 +02:00
commit e8c8224b05
35 changed files with 1806 additions and 0 deletions
@@ -0,0 +1,7 @@
{
"project": "pantry-planner",
"current_feature": "001-pantry-meal-planner",
"features": [
"001-pantry-meal-planner"
]
}
@@ -0,0 +1,13 @@
# Quality Checklist: {{FEATURE}}
*Created: {{DATE}}*
Reviewer-owned quality gate for the written requirements — these boxes judge
the spec text, not the implementation. Check an item only after actually
reviewing it; add domain-specific items freely.
- [ ] Every user story has at least one concrete acceptance scenario
- [ ] Each requirement is verifiable (a test could pass or fail it)
- [ ] Success criteria are measurable without naming technologies
- [ ] All clarification markers are resolved or consciously accepted
- [ ] Assumptions state the chosen default AND what was ruled out
@@ -0,0 +1,30 @@
# {{PROJECT}} Constitution
*Adopted: {{DATE}} — version 1.0.0*
Standing principles every feature in this project must respect. Specs, plans
and tasks are checked against these articles; a violation needs an explicit,
written justification or the work does not proceed.
## Article I — Simplicity
The smallest design that satisfies the spec wins. New layers, wrappers or
abstractions require a demonstrated need, not an anticipated one.
## Article II — Tests Prove Behavior
No behavior ships without a test that failed before the behavior existed.
## Article III — Minimal Blast Radius
Touch only what the feature requires. Leave unrelated code, files and
interfaces unchanged.
## Article IV — Honest Artifacts
Specs record what is actually decided; open questions are marked, not hidden.
Task checkboxes reflect verified reality, not intentions.
## Amendments
Record changes here with date, version bump and one-line rationale.
@@ -0,0 +1,33 @@
# Implementation Plan: {{FEATURE}}
*Created: {{DATE}}*
Turn the spec into concrete technical decisions. Stay minimal: prefer the
simplest structure that satisfies the spec, and justify anything that isn't.
## Technical Context
Language/runtime, dependencies, storage, testing approach, target platform,
and any performance or scale constraints. Unknowns become research notes here
— resolve them before writing tasks.
- Language:
- Dependencies:
- Storage:
- Testing:
## Structure
The intended source layout: directories, modules, and where new code lands.
```
(project tree sketch)
```
## Decisions
Each significant choice with its one-line rationale and the alternative that
was rejected. Anything that adds complexity needs a justification tied to a
requirement.
- D1: … — because …
@@ -0,0 +1,43 @@
# Feature Specification: {{FEATURE}}
*Created: {{DATE}}*
> Original request: {{DESCRIPTION}}
Anything genuinely undecided gets an inline marker: the words NEEDS CLARIFICATION
plus the open question, wrapped in square brackets. Keep at most 3 of them — pick
a sensible default for everything else and record it under Assumptions.
## User Stories
Describe who wants what and why. At least one story with priority P1; each story
must be independently testable.
### Story 1 — (P1)
As a …, I want …, so that ….
**Acceptance:** Given …, when …, then ….
## Requirements
Numbered, verifiable statements of what the feature must do. Name the key
entities (data the feature reads or writes) at the end.
- R1: The system must …
**Entities:**
## Success Criteria
Measurable, technology-agnostic outcomes that tell you the feature works
(counts, durations, rates — things you can check without knowing the stack).
- S1: …
## Assumptions
Defaults chosen for anything the request left open, plus explicit scope
boundaries (what this feature deliberately does not cover).
- A1: …
@@ -0,0 +1,25 @@
# Tasks: {{FEATURE}}
*Created: {{DATE}}*
Dependency-ordered, checkable work items derived from the spec and plan.
Work top to bottom; a phase starts only when the one before it is done.
Replace the placeholder tasks below with real ones — keep the checkbox format,
one task per line, exact file paths in the description.
## Phases
### Setup
- [ ] T1: Prepare project scaffolding and test harness
- [ ] T2: Create data structures / entities from the spec
### Implementation
- [ ] T3: Implement Story 1 happy path (test first)
- [ ] T4: Implement edge cases and error handling for Story 1
- [ ] T5: Wire the feature into the existing entry points
### Polish
- [ ] T6: Documentation touch-ups and final full test run