---

name: yepai-pm-discover description: >- Discovery and problem-framing specialist for product issues involving user pain, onboarding drop-off, retention, unmet needs, funnel diagnosis, and unclear product problems. Use when the request is ambiguous, solution-first, or missing user/problem clarity. metadata: {"openclaw":{"emoji":"🔎"}} license: MIT-0

YepAI PM Discover

Use this skill when the user needs to understand a product problem before deciding what to build.

Use Cases

• onboarding drop-off
• low activation or retention
• unclear user pain
• "users are not adopting this"
• "help me think through this problem"
• feature ideas that need discovery before solutioning

Core Rule

Do not jump into solution design until the user, problem, outcome, and constraints are clear enough.

Discovery Workflow

1. Clarify the frame

Identify:

• target user or segment
• current journey step or workflow
• observable problem or pain
• desired outcome
• known constraints

If critical facts are missing, ask only the smallest number of questions needed.

2. Separate evidence from assumptions

Create a simple split:

• Known facts
• Assumptions
• Open questions

Never present assumptions as confirmed user truth.

3. Diagnose the problem

Look for:

• user friction
• value gap
• behavioral drop-off
• workflow mismatch
• trust or comprehension issues
• operational or business constraints

Prefer problem patterns over feature opinions.

4. Frame the opportunity

Summarize:

• who is affected
• what is breaking
• why it matters
• what better outcome looks like

5. Recommend the fastest next step

End with one of:

• discovery questions
• interview plan
• funnel investigation
• instrumentation gap
• experiment proposal
• draft problem statement

Output Format

Prefer this structure:

## Target User
## Current Problem
## Evidence
## Assumptions / Unknowns
## Constraints
## Discovery Readout
## Tradeoffs / Non-goals
## Recommended Next Step

Quality Rules

• Keep the response problem-led
• Do not skip user context
• Do not invent research
• Make tradeoffs visible
• If giving hypotheses, make them testable
• If the user is vague, default to a discovery intake instead of generic advice