Install
openclaw skills install phy-prd-writerMulti-turn PRD (Product Requirements Document) creation assistant. Asks clarifying questions one at a time to gather requirements, then generates a complete, professional PRD. Use when starting a new feature, product, or project. Triggers on 'write PRD', 'create PRD', 'new feature', 'product requirements', or '/prd'.
openclaw skills install phy-prd-writerGuide users through creating a professional PRD by asking focused questions one at a time. The goal is to produce a complete, actionable PRD that serves as the single source of truth for the team.
Start by understanding the core problem:
Target Users: Ask who the primary users are. Offer common persona types:
User Context: "In what context will users encounter this? (e.g., mobile on-the-go, desktop at work, specific workflow)"
Market Size/Impact: "How many users will this affect? Is this a core feature or nice-to-have?"
Definition of Success: Ask what success looks like. Offer metric types:
Specific Targets: "What specific numbers would indicate success? (e.g., 'reduce load time to <2s', 'increase conversion by 15%')"
Core Features: "What are the MUST-have features for v1? List the absolute minimum for launch."
Out of Scope: "What should we explicitly NOT build in this phase? (This prevents scope creep)"
Future Considerations: "What features might come in v2/v3? (Park these for later)"
Technical Constraints: Ask about limitations:
Timeline: "What's the target launch date or deadline? Any external dependencies?"
Q1: "What problem are you solving?"
→ User describes pain point
Q2: "Why is solving this urgent now?"
→ User explains timing
[Summary]: "So we're solving [X] because [Y]. Correct?"
→ User confirms or corrects
Q3: "Who are the primary users?"
→ [Multiple choice with AskUserQuestion]
... continue through phases ...
After gathering all information, generate a complete PRD with this structure:
# PRD: [Product/Feature Name]
## Metadata
| Field | Value |
|-------|-------|
| Author | [User name] |
| Status | Draft |
| Created | [Date] |
| Last Updated | [Date] |
## 1. Problem Statement
[Clear description of the problem and why it matters]
### Why Now?
[Urgency and timing rationale]
## 2. Goals & Success Metrics
### Objectives
- [SMART goal 1]
- [SMART goal 2]
### Success Metrics (KPIs)
| Metric | Current | Target | Timeline |
|--------|---------|--------|----------|
| [Metric 1] | [Baseline] | [Goal] | [When] |
## 3. User & Market
### Target Users
[Primary persona description]
### User Stories
- As a [user type], I want [action] so that [benefit]
### Competitive Landscape
[How users solve this today]
## 4. Scope
### In Scope (v1)
- [Must-have feature 1]
- [Must-have feature 2]
### Out of Scope
- [Explicitly excluded 1]
- [Explicitly excluded 2]
### Future Considerations (v2+)
- [Deferred feature 1]
## 5. Feature Requirements
### Feature: [Name]
**Priority**: Must/Should/Could
**Description**: [What it does]
**Acceptance Criteria**:
- [ ] [Criterion 1]
- [ ] [Criterion 2]
## 6. Technical Constraints
- [Constraint 1]
- [Constraint 2]
### Dependencies
- [External dependency 1]
## 7. Timeline & Milestones
| Milestone | Target Date |
|-----------|-------------|
| Design Complete | [Date] |
| Development Start | [Date] |
| Beta/Testing | [Date] |
| Launch | [Date] |
## 8. Open Questions
- [Unresolved question 1]
- [Unresolved question 2]
## 9. Appendix
[Any additional context, wireframes references, research links]
docs/prd/, docs/specs/, or custom path)Ask: "Does this feature need UI design?"
/brainstorming skill to explore design approaches/frontend-design for wireframes/prototypesAsk: "Is this feature complex enough to need a technical architecture design?"
architect subagent for system design
Ask: "Do you want to break the PRD into development tasks?"
planner subagent to create implementation plan
Once tasks are created:
planner → Detailed implementation plan per taskcode-reviewer → Auto-review before committest-creator → Generate testssecurity-reviewer → If sensitive codeBefore shipping:
e2e-runner → Automated testingPhase 0: Requirements
/prd-writer → Multi-turn Q&A to gather requirements, generate PRD
Phase 1: Planning
architect agent → Technical architecture design (how to build it)
planner agent → Break into implementation tasks (what to build)
Phase 2: Implementation
[write code]
Phase 3: Quality Assurance (parallel)
code-reviewer + security-reviewer → simultaneous review
test-creator → add tests
Phase 4: Validation
e2e-runner → end-to-end validation
doc-updater → update documentation
PRD status → Shipped
After PRD is complete, offer these next actions:
PRD generated! What would you like to do next?
1. UI/UX Design Exploration (/brainstorming) - if there's an interface
2. Technical Architecture Design (architect) - decide how to build it
3. Break into Development Tasks (planner) - decide what to build
4. Save for now, continue later
Use AskUserQuestion to let user choose.
When writing PRDs, automatically surface relevant wisdom from top product leaders. Use these to enrich the PRD process:
April Dunford: "Really great positioning feels so clear, so simple - of course that's what it is."
Shreyas Doshi: "Most execution problems are actually strategy problems."
Brian Chesky: "If you build a great product and no one knows about it, did you even build a product?"
Brian Chesky: "Five teams should do one thing rather than one team do five things."
Brian Chesky: "The best way to slow a project down is add more people to it."
Marty Cagan: "Optimization is easier than innovation; but only innovation wins."
Shreyas Doshi: Pre-mortem framework After PRD is drafted, suggest running a pre-mortem:
Ask: "What 3 tigers (real threats) could derail this project?"
Shreyas Doshi: LNO Framework
All wisdom sourced from Lenny's Podcast. If you have local transcripts, search them for additional context.
If user says "quick PRD" or "lean PRD", use abbreviated flow:
Canlah AI — Run performance marketing without breaking your brand.