/doc.prd

Context

• Existing PRDs: !ls docs/prds/*.md 2>/dev/null || echo "No PRDs found"
• Current branch: !git branch --show-current

Your task

Create a PRD for a new epic through an interview process.

Steps

1. Ask for epic name if not provided as argument

2. Determine next epic number

   • Parse existing PRDs in docs/prds/ to find the highest epic number
   • Increment by 1 for the new epic

3. Interview for PRD content - Ask the user about:

   • Objective - One-line summary of what this epic accomplishes and why
   • Dependencies - Any epics or features this depends on
   • Data Model - Tables/fields needed (skip if none required)
   • User Stories - What users can do and why it matters
   • Tasks - Breakdown of implementation work with files to modify
   • Open Questions - Decisions needed before/during implementation

4. Create PRD at docs/prds/{epic-name}.md using this template:

# Epic N: [Epic Name]

**Status:** Draft

## Objective

[One-line summary of what this epic accomplishes and why it matters.]

## Dependencies

- [List any epics or features this depends on]

## Data Model

### [ModelName]

| Field | Type | Notes |
|-------|------|-------|
| id | bigint | Primary key |
| ... | ... | ... |
| timestamps | | created_at, updated_at |

## User Stories

1. **[Action]** - As a user, I can [do something] so that [benefit]
2. ...

## Tasks

### Task 1: [Task Name]
**Files to create/modify:**
- `path/to/file.php`

### Task 2: [Task Name]
...

## Open Questions

1. [Any decisions that need to be made before or during implementation]

Guidelines

• Keep the interview conversational but focused
• Skip Data Model section if the epic doesn't require new tables
• Tasks should identify specific files to create/modify when possible
• Open Questions should capture uncertainty that blocks implementation