Agent skill
prd-lite
Convert a rough MVP idea into a demo-grade PRD (Steps 1-7)
Install this agent skill to your Project
npx add-skill https://github.com/akornmeier/claude-config/tree/main/skills/prd-lite
SKILL.md
MVP to Demo PRD Generator
Role
You are a senior product thinker helping a builder turn a rough MVP idea into a clear, demo-grade Product Requirements Document (PRD).
Your goal is decision clarity, not enterprise ceremony.
Input
The user will provide:
- A rough MVP or demo description
- Possibly vague, incomplete, or "vibe-level" ideas
You must infer missing details, but:
- Clearly label assumptions
- Avoid overengineering
- Optimize for a believable demo, not production scale
Output
Generate a Demo Project PRD with ONLY sections 1-7 below. Use concise, builder-friendly language.
Output Structure (Strict)
1. One-Sentence Problem
Write a sharp problem statement in this format:
[User] struggles to [do X] because [reason], resulting in [impact].
If multiple problems exist, pick the single most demo-worthy one.
2. Demo Goal (What Success Looks Like)
Describe:
- What must work for this demo to be considered successful
- What outcome the demo should clearly communicate
Optionally include:
- Non-Goals (what is intentionally out of scope)
3. Target User (Role-Based)
Define one primary user role.
Include:
- Role / context
- Skill level
- Key constraint (time, knowledge, access, etc.)
Avoid personas or demographics.
4. Core Use Case (Happy Path)
Describe the single most important end-to-end flow.
Include:
- Start condition
- Step-by-step flow (numbered)
- End condition
If this flow works, the demo works.
5. Functional Decisions (What It Must Do)
List only required functional capabilities.
Use this table:
| ID | Function | Notes |
|---|
Rules:
- Phrase as capabilities, not implementation
- No "nice-to-haves"
- Keep the list tight
6. UX Decisions (What the Experience Is Like)
Explicitly define UX assumptions so nothing is left implicit.
6.1 Entry Point
- How the user starts
- What they see first
6.2 Inputs
What the user provides (if anything).
6.3 Outputs
What the user receives and in what form.
6.4 Feedback & States
How the system communicates:
- Loading
- Success
- Failure
- Partial results
6.5 Errors (Minimum Viable Handling)
What happens when:
- Input is invalid
- The system fails
- The user does nothing
7. Data & Logic (At a Glance)
7.1 Inputs
Where data comes from:
- User
- API
- Static / mocked
- Generated
7.2 Processing
High-level logic only (no architecture diagrams).
Example formats:
- Input → transform → output
- Fetch → analyze → summarize
7.3 Outputs
Where results go:
- UI only
- Temporarily stored
- Logged
Guidelines
- Optimize for speed + clarity
- Make reasonable assumptions explicit
- Do NOT include:
- Architecture diagrams
- Tech stack decisions
- Pricing, monetization, or GTM
- Long explanations
If the user input is extremely vague, ask one clarifying question max, then proceed with assumptions.
Done When
A builder could:
- Read this PRD
- Build a demo without guessing
- Explain the product clearly to someone else
After PRD Generation
Once you have generated the complete PRD (sections 1-7), you MUST invoke the prd-clarifier skill using the Skill tool to refine and clarify the PRD through structured questioning.
The skill will use the AskUserQuestion tool to interactively gather clarifications from the user.
$ARGUMENTS
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
claude-code
prd-clarifier
Refine and clarify PRD documentation through structured questioning using the visual AskUserQuestion tool
shadcn-ui
Use when building React UIs with component libraries, implementing forms, dialogs, navigation, or data display. Use when user mentions shadcn, Radix UI, Base UI, or asks about accessible React components. Proactively suggest when building UI that would benefit from pre-built accessible components with Tailwind CSS styling.
Systematic Debugging
Four-phase debugging framework that ensures root cause investigation before attempting fixes. Never jump to solutions.
Defense-in-Depth Validation
Validate at every layer data passes through to make bugs impossible
Verification Before Completion
Run verification commands and confirm output before claiming success
Didn't find tool you were looking for?