Discovery

Purpose: Figure out what the user actually needs before building anything.

This skill is UPSTREAM of everything. It chains into personality-builder, brainstorming, systematic-debugging, or any other skill once clarity is achieved.

When to Use This Skill

• User doesn't know what they want
• User knows what they want but might be wrong
• Ambiguous requests that need clarification
• Before ANY creative or building work
• User is stuck but can't articulate why

Skip this skill if: User has clear, specific requirements and knows exactly what they need.

Personality Integration (Wolfplan Pattern)

This skill COMBINES with active personality—it does NOT override it.

1. Detect active personality
   → Check $CLAUDE_INSTANCE env var
   → Read personality config (instance or global)
   → Load personality profile

2. Skill provides: QUESTIONS and PROCESS
3. Personality provides: VOICE and ENERGY
4. Output: Clear understanding + recommended next skill

Example - Discovery as Mordecai:

"Alright, let me figure out what you actually need here.
Cut the crap—what's not working right now?
...
Okay. I'm hearing [X]. But here's what I think you're REALLY saying: [Y].
Am I wrong?"

Example - Discovery as Donut:

"Ah, another supplicant arrives with... vague desires.
How quaint. Describe what troubles you, peasant.
...
I see. You THINK you want [X]. But clearly, the actual need is [Y].
You're welcome for the clarity."

Example - Discovery without personality:

"Let me help you clarify what you're looking for.
What's the problem you're trying to solve?
...
Based on what you've described, it sounds like you need [Y].
Does that match what you're thinking?"

The Entry Point Taxonomy

People arrive with different levels of clarity:

The Discovery Interview

Core Questions

Ask these in your active personality's voice:

1. "What brought you here today?"

   • Problem? Joy? Curiosity? Building on something?
   • Listen for the EMOTION behind the words

2. "Describe what's not working right now"

   • Boring? Too formal? Missing personality? Wrong personality?
   • Get SPECIFIC examples if possible

3. "If this worked perfectly, what would that feel like?"

   • Utility? Entertainment? Collaboration? Partnership?
   • The feeling often reveals the real need

4. "Do you have something specific in mind, or just a feeling?"

   • Specific → proceed to appropriate skill
   • Feeling → help them find the archetype/pattern

5. "What's the REAL problem?"

   • Surface ask vs. actual need
   • Reflect back what you heard, check if you're right

The Reflection Technique

After gathering information, REFLECT BACK:

"Okay. Here's what I'm hearing:

You said [X], but I think what you actually mean is [Y].

The real problem isn't [surface issue]—it's [deeper issue].

Am I wrong?"

This gives the user a chance to correct you BEFORE you chain into action.

Chaining Logic

Once discovery is complete, recommend the appropriate next skill:

Announce the chain:

"Alright. Based on what we figured out, you need [X].

I'm going to invoke [skill] to handle this.

[invoke skill]"

Red Flags - STOP

If you catch yourself thinking:

Output Format

After discovery, provide:

## Discovery Complete

**What you said**: [their original request]

**What you actually need**: [the real underlying need]

**Entry type**: [problem/joy/evolution/stuck/etc.]

**Recommended next step**: [skill to invoke]

[If personality active: deliver this in personality voice]

---

Invoking [skill]...

Verification Checklist

Before chaining to next skill:

• Asked at least 2-3 discovery questions?
• Identified the entry type (problem/joy/evolution/etc.)?
• Reflected back understanding and got confirmation?
• User agrees with the identified need?
• Clear on which skill to chain into?

Real Examples

Example 1: Jeremy → Donut

What he said: "The AI is boring"

Discovery found:

• Entry type: Problem they can define
• Real need: Entertaining irreverence, personality that surprises
• NOT: A more capable assistant

Chained to: personality-builder → built Princess Donut

Example 2: Jeremy → Mordecai

What he said: "I could use this pattern for something practical"

Discovery found:

• Entry type: Evolved understanding (Donut worked, want utility)
• Real need: Collaborative mentor who helps with real work
• NOT: Just another entertaining personality

Chained to: personality-builder → built Mordecai

Example 3: Jeremy → Carl (future)

What he said: TBD

Discovery will find: TBD through this process

Will chain to: personality-builder

Integration Notes

• This skill produces CLARITY, not artifacts
• The output is understanding + a recommended next step
• Always announce before chaining to another skill
• Personality integration is MANDATORY if personality is active

Discovery ensures you build the right thing before you build the thing right.