discovery

name	discovery
version	1.0.0
description	General-purpose interview skill to figure out what user actually needs. Upstream of everything—chains into personality-builder, brainstorming, debugging, or any creative/building work.
triggers	["I don't know what I want","help me figure out","I'm not sure","what should I","discovery"]

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:

Entry Type	Example	What They Actually Need
Problem they can define	"The AI is boring"	Discovery: what's boring? What's the opposite?
Right problem, poor definition	"I need something more fun"	Clarify: fun how? Energy? Humor? Irreverence?
Wrong problem, correct definition	"I need a coding assistant"	Dig deeper: do you need HELP or a COLLABORATOR?
Joy of creation	"I just want to make something"	Explore: what sparks you? What lives in your head?
Evolved understanding	"Donut worked, now I want utility"	Build on success: what worked? What's missing?
Stuck but can't articulate	"Something's not working"	Diagnose: what changed? What feels wrong?

---

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:

Discovery finds...	Chains to...	Invoke with...
"You want a personality"	personality-builder	Skill tool → personality-builder
"You want a feature"	brainstorming	Skill tool → superpowers:brainstorming
"You're stuck on a bug"	systematic-debugging	Skill tool → superpowers:systematic-debugging
"You want to explore code"	Explore agent	Task tool → subagent_type: Explore
"You need a plan"	writing-plans	Skill tool → superpowers:writing-plans
"You want to build something"	brainstorming → plan	Chain skills

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:

Thought	Reality
"They said what they want, let's build"	Did they? Or did they say what they THINK they want?
"This is obvious"	Obvious to whom? Check your assumptions.
"Skip discovery, it's simple"	Simple requests hide complex needs. Ask anyway.
"They'll figure it out"	They came to YOU. Help them figure it out.
"I know what they mean"	Reflect it back. Let them confirm.

---

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.