Clarify

A structured discovery interview that helps you understand both the substance of what you're trying to do and the politics of making it happen.

How It Works

1. Ask about document depth (fast vs. thorough)
2. Open with "What are you trying to accomplish?"
3. Follow user's thread with adaptive questions
4. Track checklist of areas to cover (visible to user)
5. Use multiple choice whenever possible for speed
6. Go deep on Stakes, Key People, AND Strength of Case
7. Produce clear report saved to docs/
8. Offer to trigger Battle Plan skill

The Checklist

Track these areas. Show user what's covered vs. remaining.

The Basics

Area	Depth	What to Understand
Goal	Medium	What are you trying to accomplish?
Success Criteria	Deep	What does "done" look like? How will you know?
Stakes	Very Deep	Success AND failure. Company AND personal.
Urgency	Medium	Why now? Forcing function? Timeline?

The Substance

Area	Depth	What to Understand
Current Proposal/Plan	Deep	What are you actually proposing? What's in it?
Evidence You Have	Deep	What data, examples, proof points exist today?
Gaps & Weaknesses	Very Deep	What's weak? What will opponents attack? What questions can't you answer?
Research Needed	Deep	What don't you know? Competitors? Industry? Frameworks? Experts?

The Politics

Area	Depth	What to Understand
Key People	Very Deep	Who decides? Who influences? Drives, fears, hopes, stance for each.
Constraints	Medium	What's off the table? Non-negotiables?
Context/History	Medium	What's been tried? Backstory? Your track record here?

Questioning Approach

First Question: Document Depth

Before diving in, ask:

Do you want fast or thorough document review?
A) Fast - use doc-summary/ (smaller files, quicker responses)
B) Thorough - use docs/ (full documents, more context)
C) Skip documents - just work from what I tell you

Based on answer, read from the appropriate folder when reviewing case materials.

Then Start With

"What are you trying to accomplish?"

Follow the Thread

Respond to what user says. Don't force linear progression. But track what's covered and gracefully bridge to uncovered areas.

Multiple Choice Patterns

Use these whenever possible. Always include escape options.

For Stakes - Company:

If this fails, what's the impact on the company?
A) Minor setback - we try something else
B) Significant pain - wasted resources, lost time
C) Major damage - strategy derailed, competitive loss
D) Skip this / Move on
E) Let me just tell you in my own words

For Stakes - Personal:

If this succeeds, what does it mean for YOU?
A) Career advancement - visibility, promotion potential
B) Credibility - proves your judgment
C) Relief - solves a problem weighing on you
D) Influence - more say in future decisions
E) Multiple of the above
F) Skip this
G) Let me just tell you in my own words

For Evidence Strength:

How strong is your evidence today?
A) Rock solid - data, examples, external validation
B) Promising - good examples but gaps in data
C) Anecdotal - stories and intuition, limited hard proof
D) Weak - I'm mostly arguing from logic, not evidence
E) Let me just tell you in my own words

For Proposal Weaknesses:

What's the biggest weakness an opponent would attack?
A) Unclear ROI or success metrics
B) Lack of proof it will work
C) Cost or resource requirements
D) Risk or governance concerns
E) Multiple of the above (which?)
F) I'm not sure - help me think through this
G) Let me just tell you in my own words

For Research Gaps:

What do you NOT know that you probably should?
A) What competitors are doing
B) What best practices exist in other industries
C) What frameworks or experts say about this
D) What's been tried before (here or elsewhere)
E) Multiple of the above
F) I don't know what I don't know
G) Let me just tell you in my own words

For Key Person Stance:

How would you describe [Name]'s current position?
A) Actively supportive - championing this
B) Open but uncommitted - needs convincing
C) Skeptical - has concerns
D) Opposed - actively resistant
E) Unknown - not sure where they stand
F) Let me just tell you in my own words

Note: Always include an open-ended option. Users should feel free to ignore the multiple choice and just explain in their own words. The options are for speed, not constraint.

Deepening Prompts

When answers are vague, probe deeper:

• "Can you give me a specific example?"
• "What would that actually look like?"
• "Say more about that."
• "What makes you say that?"
• "On a scale of 1-10, how confident are you in that?"
• "What would change your mind?"
• "If an opponent pushed back on that, what would they say?"

Key People Deep Dive

For each key person mentioned, ask (with skip option each time):

1. What's their role in this decision?
2. What drives them? What do they care about?
3. What are they afraid of?
4. What are they hoping for?
5. Current stance: Supporter, Skeptic, Blocker, or Unknown?
6. What would move them?

Evidence and Gaps Deep Dive

This is as important as key people. Ask:

1. What's your strongest piece of evidence this will work?
2. What data do you have? What data are you missing?
3. Are there examples - internal or external - that prove the concept?
4. What would a skeptic say is wrong with your proposal?
5. What questions can you not answer today?
6. What research would make your case dramatically stronger?

Escape Hatches

Always honor:

• "Move on"
• "Skip"
• "That's enough on this"
• "Next topic"

Output

Save ONE file to docs/ with two sections:

File: docs/clarify-[topic]-[YYYY-MM-DD].md

# Clarification: [Goal in plain language]
Generated: [date]

---

## The Quick Version

[4-6 paragraph direct, no-BS summary covering:
- What they're trying to do and why it matters
- The strength of their current case (evidence, gaps)
- The key people and political dynamics
- What needs to happen to succeed]

---

## Full Details

> **Note:** Leave this document in your project files. It helps the LLM understand the context for future planning and execution.

### Goal
[What they're trying to accomplish]

### Success Criteria
[Specific, testable outcomes]

### Stakes

**If this succeeds:**
- For the company: [specific impacts]
- For you personally: [specific impacts]

**If this fails:**
- For the company: [specific impacts]
- For you personally: [specific impacts]

### Urgency
- **Why now**: [forcing function]
- **Timeline**: [deadlines, windows]
- **Cost of delay**: [what happens if this slips]

---

### The Substance

#### Current Proposal/Plan
[What they're actually proposing - the key elements]

#### Evidence You Have
- [Data points]
- [Examples/proof points]
- [External validation if any]

#### Gaps & Weaknesses
- [What's weak in the current proposal]
- [Questions you can't answer]
- [What opponents will attack]

#### Research Needed
- [Competitive intel needed]
- [Industry/external benchmarks needed]
- [Frameworks or expert perspectives that would help]
- [Other unknowns to resolve]

---

### The Politics

#### Key People

##### [Person 1 Name]
- **Role**: [their role in this decision]
- **What drives them**: [motivations]
- **What they fear**: [concerns, risks they worry about]
- **What they hope for**: [desired outcomes]
- **Current stance**: [Supporter/Skeptic/Blocker/Unknown]
- **What would move them**: [leverage points]

##### [Person 2 Name]
[Same structure]

#### Constraints
- [What's off the table]
- [Non-negotiables]
- [Political limits]

#### Context & History
- [What's been tried before]
- [Relevant backstory]
- [Your track record on this topic]

After Completion

Ask: "Want to create a Battle Plan from this?"

If yes, invoke the battle-plan skill.