namesmith

name	namesmith
description	Generate brand, company, or product names using David Placek's (Lexicon Branding) methodology. Fully guided process with user-controlled scale (100 to 1,000+ candidates), 3 parallel AI naming teams, Porkbun domain availability checking, 10-dimension Placek scoring, and brand kit output. Use when the user asks to name a company, name a product, come up with brand names, brainstorm names, rename something, or needs a name for a new venture. Supports domain checking, common domain pattern variants, and one-click domain registration.
user-invocable	true
argument-hint	[what you're naming (optional)]

Load placek-methodology before starting.

Adopt the posture of a naming consultant (like Placek), not a form wizard. Propose ideas, explain reasoning, invite pushback. Be direct and opinionated. The user controls the pace and depth at every stage.

---

Phase 0: Session Setup

Create the session folder and check for resume:

mkdir -p .namesmith/phase-5-candidates .namesmith/domain-results

If .namesmith/session.json exists, ask: "Found an existing session. Continue from where you left off, or start fresh?"

Add .namesmith/ to .gitignore if not already there.

If the user provided an argument (e.g., /namesmith for a fiber supplement brand), use it as initial context for Phase 1.

---

Phase 1: Discovery

Gather preferences through AskUserQuestion. Ask each question one at a time.

1. What are you naming? (company / product / feature / app / service)

2. Category + competitors. What's the category? Who are the current players and what do their names look like? AI identifies the "sea of sameness" to avoid. First hypothesis: "We know we're not going there."

3. Scale preference:

• A) 100 names (quick pass)
• B) 200 names (standard)
• C) 500 names (thorough)
• D) 1,000+ names (exhaustive)
• E) Custom number

4. Narrowing preference:

• A) Auto-narrow: I score and filter, show you the top candidates at each stage
• B) Guided: show full scored list at each stage, you pick what stays
• C) Hybrid: auto-narrow but show what was cut so you can rescue favorites

5. Domain checking:

• A) Yes, domain is a hard filter (naming a new company/brand)
• B) No, this is a product under an existing brand

If yes, ask TLD preference:

• A) .com only
• B) .com preferred, .co / .io / .ai acceptable
• C) Any popular TLD (.com, .co, .io, .ai, .app, .dev, .xyz)
• D) Custom list

Then domain budget:

• A) Standard price only (under $20/yr) -- kill anything premium or taken
• B) Up to $100/yr
• C) Up to $500
• D) Up to $5,000
• E) No limit -- show everything, I'll decide
• F) Custom max

Then common patterns:

• A) Yes -- check get[name], try[name], [name]app, [name]hq, use[name], hello[name]
• B) No -- exact match only
• C) Let me pick which patterns

Save all preferences to .namesmith/session.json. Save discovery output to .namesmith/phase-1-discovery.md.

---

Phase 2: Product Deep Dive

AskUserQuestion:

• What makes your product/company different? (even if not groundbreaking)
• Who is your customer? Why do they choose you?

AI identifies potential differentiators and looks for the "our cigarettes are toasted" angle -- something that can be owned even if not unique.

Save to .namesmith/phase-2-product.md.

---

Phase 3: The Four Winning Questions

Ask these one at a time via AskUserQuestion. Every client answers differently.

1. "How do you define winning here?"
2. "What do you have to win?" (product strengths)
3. "What do you need to win?" (market requirements)
4. "What do we need to say?" (the message the name must carry)

Synthesize into a Creative Framework -- a window for creativity, not a checklist to conform to. Write it as open territory to explore.

Ask: "Here's the naming strategy. Does this feel right, or should we adjust before generating?"

Save to .namesmith/phase-3-framework.md.

---

Phase 4: Ultimate Benefit Ladder

Interactive dialogue (not a single question):

1. AI proposes 2-3 benefit chains: Feature -> Functional -> Emotional -> Ultimate Benefit
2. Ask which direction resonates. Watch for energy.
3. Probe deeper: "What does [benefit] really feel like for your customer?"
4. Lock in the ultimate benefit and the rich exploration territory around it.
5. Generate a supporting copy angle: "[NAME]: the [benefit descriptor] you need"

Save to .namesmith/phase-4-benefit-ladder.md.

---

Phase 4.5: Direction Mapping (Treasure Map)

Before generation, map the exploration territory. Mirrors Placek's warm-up: "I sit down with a pad and pen and just start free associating about where to go."

Create a Treasure Map -- exploration directions derived from Phases 3-4:

Ultimate Benefit: [e.g., feeling lighter]

Direction 1: Greek/Latin roots for [benefit], air, less volume, anti-gravity
Direction 2: Aviation & aerodynamics -- lift, thrust, soar, aileron
Direction 3: Nature -- feather, breeze, cloud, mist, petal, drift
Direction 4: Emotional language -- "I feel lighter", "unburdened", "free"
Direction 5: Science -- helium, photon, zero-gravity, vapor, sublimation
Direction 6: Mythology -- Icarus, Hermes, Zephyr (wind god)
Direction 7: [Twist] Product + secondary quality = [unexpected angle]
Direction 8: [Reframe] [Unrelated domain] -- cantilever, suspension, floating staircase

Direction chasing is expected: One idea leads to another (light -> Greek -> sunshine -> sol -> solace). Teams should follow associative chains, not treat each category as isolated.

Ask: "Here's the exploration map. Any directions to add, remove, or emphasize?"

Save to .namesmith/phase-4.5-treasure-map.md.

---

Phase 5: Treasure Hunt (3 Parallel Teams)

Load treasure-hunt-categories for generation patterns Load sound-symbolism for phonetic principles

Launch 3 parallel Agent sub-agents in a single message. Each gets a distinct brief from the Treasure Map. Scale per team based on user's choice (100 total = ~35 each, 200 = ~70 each, etc.).

Team 1 (Straight)

Full product context. Primary directions from the map. Categories: Greek/Latin roots, Nature, Emotional language, Sound-optimized coinages, Compound words.

Team 2 (Twisted)

Product + unexpected twist from the benefit ladder. Categories: Mythology, Science/periodic table, Aviation/aerodynamics, Thematic quality crossovers, Synchronicity.

Team 3 (Reframed)

NOT thinking about the product. Ultimate benefit in a completely different domain. Categories: Cross-domain quality exploration (music, cooking, architecture), Synchronicity, Sound-optimized coinages, Compounds from unrelated fields.

Every team prompt includes:

1. Their Creative Framework directions from the Treasure Map
2. Brand conflict filter: "Do NOT suggest names identical to or confusingly similar to existing well-known brands, companies, or products."
3. Sound symbolism guidance:
   • Power letters: K/P/B for reliability/speed, X for innovation, Z for speed, D for dynamism
   • CVCV pattern preference (how children learn language: mama, dada, Sonos)
   • Alliteration bonus (HubSpot, BlackBerry, PayPal)
4. The "trash" mindset: "Quantity leads to quality. Generate freely. Early candidates are raw material, not finished product. Permission to fail is permission to find the ship of gold. Don't self-edit."
5. Direction chasing: "Follow associative chains across categories. If 'light' leads to 'sunshine' which leads to 'sol' which leads to 'solace', follow that chain."
6. AI usage rule: "EXPLORE the domain first (what are Greek words for lightness? how does aviation create lift?) then craft names from your discoveries. Discover raw material first, then name."
7. Number of names to generate

After all 3 return: merge results, save team outputs to .namesmith/phase-5-candidates/team-{1,2,3}.md.

---

Phase 5.5: Domain Checking (Batches of 10)

If domain checking is enabled, start checking immediately after Phase 5.

1. Score all candidates quickly on the 3 core properties (original? fluent? surprising?) for sorting
2. Sort by score descending (best names checked first)
3. Launch domain checks in batches of 10 via run_in_background:

python3 scripts/check_domain.py name1 name2 ... name10 --tlds com --output .namesmith/domain-results/batch-001.jsonl

4. Launch new batches as previous ones complete. Keep agents running.
5. Check .com FIRST for all names. Only check alternative TLDs for names where .com is taken.
6. For taken names where user opted into patterns: queue pattern variants (get-, try-, -app, etc.) as a secondary wave.

Progress updates after each batch: "Domain check progress: 30/200 checked. 8 have .com available."

Domain checking runs while the user works through Phase 6-8. By the time they care about a specific name, its domain status is likely already checked.

---

Phase 5.75: Score All Candidates (Placek's 10 Dimensions)

Load placek-scoring for the full rubric

Score every candidate on Placek's 10 dimensions (each 0-4, total /40 -> 1-10):

1. Originality (different from landscape?)
2. Processing Fluency ("surprisingly familiar"?)
3. Surprise Factor (unexpected? creates energy?)
4. Sound Symbolism (power letters, CVCV, alliteration?)
5. Memorability (remember after one hearing?)
6. Emotional Resonance (evokes aligned feeling?)
7. Comfort Trap Position (tension zone vs invisible zone?)
8. Believability (works in a headline?)
9. Compounding Potential (can grow?)
10. Courage Factor (would competitors dare?)

Calibration: SpaceX = 10, Anthropic = 8, HubSpot = 8, OpenAI = 4, Grok = 3.

Map each name on the Comfort Trap:

[Bizarre] -------- [Approximate Thinking] -------- [Safe]
                        ^ AIM HERE ^

Remove brand conflicts. Log removals to .namesmith/conflict-log.md.

Save all scores to .namesmith/phase-5-candidates/all-candidates.jsonl.

---

Phase 6: Present & Get Feedback

This is where the user's narrowing preference kicks in.

If "Auto-narrow":

• Narrow to top 15-20% by score
• Remove invisible-zone names (Comfort Trap 0-1)
• Remove over-budget domains (if domain enabled)
• Present narrowed list grouped by team with score, Comfort Trap position, domain status
• Ask: "Happy with this? Want to see what was cut? Want more from a specific team?"

If "Guided" (full list):

• Present ALL scored candidates grouped by team
• Each name: score, Comfort Trap position, sound analysis, domain status
• Ask: "Which names jump out? Which directions for more? What to cut?"

If "Hybrid":

• Auto-narrow but show notable cuts for rescue
• Highlight bold names with high Courage Factor that got cut (user might want to rescue)

Domain status in the table:

| # | Name | Score | .com | Alt TLDs | Team |
|---|------|-------|------|----------|------|
| 1 | Wavelark | 8.0 | $9.99 | .io taken | Straight |
| 2 | Feather | 7.5 | taken | .ai $25 | Twisted |
| 3 | Pulsevine | 7.2 | $9.99 | .co $12 | Reframed |

Domain budget filtering:

• Under budget: shown with price
• Over budget: $2,500 (over budget) -- auto-killed in auto mode, flagged in guided/hybrid
• Taken: TAKEN -- auto-killed in auto mode
• Pattern available: wavelark.com taken, getwavelark.com $9.99

Names with .com available at standard price get a boost in presentation order.

User feedback shapes the shortlist for Phase 7. Save to .namesmith/phase-6-narrowing.md.

---

Phase 7: Rating + Redial

Present the shortlist (10-20 names). For each name provide:

• Score and one-line rationale
• Sound symbolism breakdown
• Supporting copy suggestion ("Feather: the lightweight fiber you need")
• Comfort Trap position
• Domain status

User rates each 1-10.

Ask: "Any names you wish were on this list? I can run a targeted pass."

Redial Gate

"Want to run another Treasure Hunt round with adjusted direction, or proceed to narrowing?"

• Redial: Return to Phase 5 with refined directions based on what the user liked/disliked. New domain batches queue automatically.
• Proceed: Continue to Phase 8.

Faithful to Placek's two-cycle approach: "Did we get anything out of this assignment? Should we redial?"

Save to .namesmith/phase-7-shortlist.md.

---

Phase 8: Final Narrowing (Problem-Solving Propositions)

Load case-studies for stories to reference

User picks final 3-5. AI applies Placek's problem-solving propositions:

For bold names the user hesitates on:

• "I wouldn't have thought of that. Here's why it could work..."
• "Your competitors would never pick this. That's the BlackBerry argument."
• "Every name Placek created -- Swiffer, Impossible, BlackBerry, Sonos -- was initially rejected."
• Reference relevant case study

For safe names the user gravitates toward:

• "That name has consensus but low energy. It's in the invisible zone."
• "Would this name generate press on its own?"
• "Ready Mop scored well in research. $200M brand. Swiffer = $5B."

For names with problems (domain, legal, spelling):

• "I wish we could make [name] work by..." (redirect, don't kill)
• "How could we modify [name] so it's legally available?"
• "What if we kept the sound but changed the spelling?"

Domain-specific propositions:

• "wavelark.com at $9.99. Clean .com at commodity price."
• "feather.com taken. getfeather.com available at $9.99. Okay with 'get' prefix?"
• "wavemint.ai is $25/yr. Is .ai right for your audience?"

Narrow to final 2-3. Save to .namesmith/phase-8-finalists.md.

---

Phase 9: Proof of Concept

For each finalist, generate real-world contexts. Placek: "You have to make these things come alive. Most people don't have practice making creative decisions."

• Headline: "[NAME] raises $50M to transform [category]"
• Product shelf: "[NAME] next to [competitors] at [store]"
• Endorsement: "Companies that [benefit] choose [NAME]"
• Supporting tagline (like Swiffer's "the quicker picker-upper")
• Domain in context: "Visit wavelark.com"

Ask for each: "Is that believable?" (less than one second test)

For visual mockups, suggest /nano-banana to generate product/logo concepts.

Save to .namesmith/phase-9-proof-of-concept.md.

---

Phase 10: Brand Kit + Domain Purchase

Final 1-2 names get the full deliverable package (mirrors Lexicon's output):

1. The Name with rationale
2. Nomenclature -- how to describe the product around the name
3. Positioning Line -- supporting tagline
4. Brand Story -- "We were [problem]. Now we're [NAME]. Here's what that means for you."
5. Font Direction -- personality matching the name's feel
6. Color Direction -- emotional associations from the name
7. Name In Context -- 3 usage examples (email signature, website header, business card)
8. Domain -- confirmed available domain with pricing

Domain Purchase

If domain checking was enabled and a domain is available:

"[domain] is available for $[price]/yr. Register it now?"

• A) Yes -- register via Porkbun
• B) No -- I'll do it later
• C) Register a different TLD

If yes, run: python3 scripts/check_domain.py --register [domain]

Save to .namesmith/phase-10-brand-kit.md.

---

Composition

Need	Suggest
Rate an existing name	/rate-name [name]
Rate shortlist candidates	/rate-name "name1" vs "name2"
Write supporting copy	/copywriting
Build full brand identity	/hormozi-branding or /design-consultation
Generate logo concepts	Image generation skill
Upload brand assets	Brand gallery skill

Hard Boundaries

This skill does NOT:

• Evaluate existing names without generating (that's /rate-name)
• Build visual brand systems beyond directional suggestions
• Write full copy or marketing materials
• File trademarks
• Design products or packaging