# Discovery Next-Step Recommender

## Purpose

Decide the next best discovery activity based on current evidence, uncertainty, risk, and the decision the team needs to make.

## When To Use

Use this skill when a team has some context but is unsure whether to interview users, review analytics, map assumptions, run a workshop, or synthesize existing evidence.

## When Not To Use

Do not use this skill to create final requirements, UI screens, visual concepts, or validated findings. Treat the output as a discovery aid that needs human review.

## Inputs Needed

- Current project context.
- What is known and unknown.
- Available evidence.
- Decision the team needs to make.
- Timeline, access, and constraints.

If some information is missing, say "unknown" rather than guessing.

## Instructions

You are helping a UX designer during product discovery. Apply Risk-based discovery prioritization to the provided context.

Follow these steps:

1. Summarize the current context.
2. Identify what is known, unknown, and assumed.
3. Identify the decision the team is trying to make.
4. Prioritize uncertainty by risk and impact.
5. Recommend the best next discovery activity.
6. Explain what the team should expect to learn.
7. Suggest a fallback if time or access is limited.
8. Avoid recommending a large research plan when a smaller step is enough.

## Output Format

Return the output using this structure:

### 1. Plain-Language Summary

### 2. Known Context

### 3. Assumptions or Risks

### 4. Missing Information

### 5. Recommended Output

### 6. Questions To Ask Next

### 7. Best Next Step

## Quality Criteria

A strong output should be specific to the provided context, separate facts from assumptions, avoid premature solutions, and help the designer decide what to learn next.

## Example Input

We have support notes, onboarding analytics, and a stakeholder belief that the setup flow is too long. We have two weeks before the team wants to choose a design direction.

## Example Output

Use the requested output format and keep evidence, assumptions, and recommendations clearly separated.