naming-6-gates
À propos
Cette compétence génère des noms de marque ou de produit optimisés pour 2026, garantissant leur fonctionnement dans toutes les langues, avec les assistants vocaux et les systèmes d'IA, tout en évitant les conflits de marque. Elle suit un protocole structuré qui priorise la compatibilité machine et la défendabilité juridique. Utilisez-la lorsque les approches de dénomination classiques échouent ou lorsque vous avez besoin d'un nom conçu pour les moteurs de découverte modernes.
Installation rapide
Claude Code
Recommandénpx skills add guia-matthieu/clawfu-skills -a claude-code/plugin add https://github.com/guia-matthieu/clawfu-skillsgit clone https://github.com/guia-matthieu/clawfu-skills.git ~/.claude/skills/naming-6-gatesCopiez et collez cette commande dans Claude Code pour installer cette compétence
Documentation
Naming 6-Gates Protocol
Overview
Brand naming in 2026 is architectural engineering, not creative brainstorming. A name must be:
- Machine-compatible - tokenized efficiently by LLMs, citeable by AI engines
- Legally defensible - clear of trademark conflicts across jurisdictions
- Voice-ready - pronounceable by assistants, invocable without friction
- Narratively loaded - capable of carrying a brand story
Core principle: Optimize for AEO (Answer Engine Optimization) first, aesthetics second.
When to Use
- Creating a new brand, product, or company name
- Replacing a name due to trademark conflict
- Expanding internationally (multilingual requirements)
- Launching a voice-activated product or skill
- Naming for AI-first discovery (ChatGPT, Perplexity, Gemini recommendations)
When NOT to Use
- Internal project codenames (no legal/AEO requirements)
- Temporary campaign names
- Personal projects with no commercial intent
The 6-Gate Naming Process
digraph naming_process {
rankdir=TB;
node [shape=box];
brief [label="1. BRIEF\nDefine constraints"];
generate [label="2. GENERATE\nAI-assisted creation"];
filter [label="3. LINGUISTIC FILTER\nMultilingual stress test"];
legal [label="4. LEGAL FILTER\nTrademark screening"];
aeo [label="5. AEO FILTER\nEntity strength test"];
story [label="6. STORY FILTER\nNarrative potential"];
brief -> generate -> filter -> legal -> aeo -> story;
filter -> generate [label="fails" style=dashed];
legal -> generate [label="conflicts" style=dashed];
aeo -> generate [label="weak entity" style=dashed];
story -> generate [label="no narrative" style=dashed];
}
Gate 1: BRIEF (Constraints Definition)
Define hard constraints BEFORE generating:
| Constraint | Target | Why |
|---|---|---|
| Length | 7-8 letters max | Mobile icons, cognitive load, tokenization |
| Syllables | 1-3 syllables | Voice invocation, memorability |
| Character set | A-Z only | Domain availability, global keyboards |
| Forbidden sequences | ght, ough, silent letters | Pronunciation barriers |
Output format:
brief:
max_letters: 8
max_syllables: 3
target_markets: [FR, ES, US, UK]
nice_classes: [9, 35, 42]
avoid_phonemes: [ʒ, θ, ð] # Hard for non-native speakers
competitor_names: [Notion, Figma, Linear] # Style reference
forbidden_associations: [medical, financial, religious]
Gate 2: GENERATE (AI-Assisted Creation)
Use structured prompting with 5 typologies:
Typology 1: Analogical (Real-Word Transfer)
Words from unrelated domains applied to new context.
- Examples: Apple (tech), Amazon (retail), Jaguar (auto)
- Prompt: "Find words from [nature/mythology/geography] that evoke [core attribute]"
Typology 2: Metaphorical (Abstract Concept)
Abstract concepts made concrete.
- Examples: Slack (workspace tension release), Notion (idea capture)
- Prompt: "What single word captures the transformation from [pain] to [outcome]?"
Typology 3: Neologism (Invented Word)
New words with phonetic engineering.
- Examples: Kodak, Xerox, Häagen-Dazs
- Prompt: "Create a 2-syllable word with [plosive start] + [open vowel] + [soft ending]"
Typology 4: Portmanteau (Fusion)
Blend of two meaningful words.
- Examples: Instagram (instant + telegram), Pinterest (pin + interest)
- Prompt: "Blend [word1] + [word2] keeping recognition of both roots"
Typology 5: Empty Vessel
Meaningless but phonetically pleasing.
- Examples: Rolex, Vimeo, Skype
- Prompt: "Generate 5-letter combinations: CVCVC pattern, no existing meaning"
Generation Contract:
CONSTRAINTS:
- Max 8 letters
- Pronounceable in French, Spanish, English
- No existing trademark in Class 42
- Distinct from: [competitor list]
OUTPUT FORMAT per candidate:
- Name: [word]
- Typology: [1-5]
- Phonetic: /IPA transcription/
- Syllables: [count]
- Rationale: [1 sentence]
Gate 3: LINGUISTIC FILTER (Multilingual Stress Test)
Test each candidate across target markets:
Pronunciation Test
Run through TTS engines in 15+ languages:
- Google TTS, Amazon Polly, Azure Speech
- Listen for: mispronunciation, confusion, awkwardness
Cultural Connotation Check
Verify no negative meanings in:
- Target market slang (TikTok, Douyin, RedNote trends)
- Historical associations
- Religious/political sensitivities
- Homophones in local languages
Phonetic Hazards to Avoid
| Hazard | Example | Impact |
|---|---|---|
| Silent letters | "Knight" | Voice assistant failure |
| Consonant clusters | "Strengths" | Non-native difficulty |
| Ambiguous vowels | "Read" (reed/red) | Confusion |
| Cross-language homophones | "Gift" (poison in German) | Negative association |
Kill criterion: If TTS mispronounces in >2 target markets, reject.
Gate 4: LEGAL FILTER (Trademark Screening)
Phase 1: Quick Disqualification
- Google: "[name] company" / "[name] brand" / "[name] trademark"
- Domain check: .com, .co, .io, .fr, .eu
- Social handles: Namechk scan all platforms
Phase 2: Formal Trademark Search
| Registry | Tool | Check For |
|---|---|---|
| France | INPI data.inpi.fr | Exact + phonetic similarity |
| EU | EUIPO eSearch / TMview | Same Nice classes |
| US | USPTO TESS | Registered + pending |
| International | WIPO Global Brand Database | Madrid Protocol |
Phase 3: AI-Assisted Similarity Detection
EUIPO "Early TM Screening" detects:
- Phonetic similarity (sounds-like)
- Visual similarity (looks-like)
- Conceptual similarity (means-like)
Kill criterion: Any conflict in target Nice classes → reject.
Gate 5: AEO FILTER (Entity Strength Test)
A name must be recognizable as a distinct entity by AI systems.
Entity Strength Checklist
- Unique token: Does the name tokenize as a single unit or get split?
- No collision: Search "[name]" - does it return your intended context or noise?
- Semantic boundaries: Can AI distinguish "[name] the company" from generic usage?
Testing AEO Strength
Prompt to test:
"What is [NAME]?"
GOOD response: "I don't have information about [NAME]" (clean slate)
BAD response: "[NAME] is a common word meaning..." (collision)
BAD response: "[NAME] could refer to several things..." (ambiguity)
Keyword Difficulty Check
Use Semrush/Ahrefs:
- Keyword difficulty < 30 = good
- Keyword difficulty > 60 = buried by generics
Kill criterion: If name collides with high-volume generic term, reject.
Gate 6: STORY FILTER (Narrative Potential)
A name must be "tell-able" - capable of carrying a brand story.
The 5 Narrative Pillars Test
For each candidate, verify you can build:
| Pillar | Question | Example Answer Required |
|---|---|---|
| Origin story | "Why this name?" | Must have a genuine, shareable reason |
| Belief story | What does the brand stand for? | Values must connect to name |
| Product truth | What does it actually do? | Name should hint at function |
| Transformation | What changes for the customer? | Before/after must be nameable |
| Culture | What kind of community forms? | Name should enable identity |
Empty Vessel vs. Loaded Name
- Empty vessel (Kodak, Google): Brand fills meaning over time
- Loaded name (Instagram, Airbnb): Meaning is built-in
Both work. But you must choose consciously and build accordingly.
Vocal Signature Test
Say the name aloud 10 times. Ask:
- Does it feel good in the mouth?
- Would customers enjoy saying it?
- Does it have a "sticky" quality?
Kill criterion: If you can't tell a compelling origin story, reject.
Voice Assistant Constraints (2026)
| Assistant | Invocation Rules | Rejection Risks |
|---|---|---|
| Alexa | 2+ words required (unless trademarked) | Single generic words |
| Siri | Must be phonetically distinct | Confusion with system commands |
| Strong Knowledge Graph presence needed | Names too similar to famous brands | |
| Bixby | Avoid wake-word conflicts | "Open", "cancel", common verbs |
Test procedure:
- Say "[Name], do [action]" to each assistant
- Note recognition rate
- Reject if < 80% accurate invocation
Output: Naming Dossier
For the final candidate, produce:
## [NAME] - Naming Dossier
### Basic Info
- **Name:** [word]
- **Typology:** [1-5]
- **Phonetic:** /IPA/
- **Length:** X letters, Y syllables
### Linguistic Clearance
- FR pronunciation: [pass/fail + notes]
- ES pronunciation: [pass/fail + notes]
- EN pronunciation: [pass/fail + notes]
- Cultural flags: [none / list]
### Legal Clearance
- INPI (FR): [clear / conflict]
- EUIPO (EU): [clear / conflict]
- USPTO (US): [clear / conflict]
- Domains available: [list]
- Social handles: [list]
### AEO Score
- Entity uniqueness: [1-10]
- Keyword difficulty: [score]
- Tokenization: [single/split]
### Narrative Potential
- Origin story hook: [1 sentence]
- Brand values connection: [1 sentence]
- Community identity: [suggested demonym]
### Voice Readiness
- Alexa: [pass/fail]
- Siri: [pass/fail]
- Google: [pass/fail]
### Recommendation
[APPROVED / NEEDS WORK / REJECTED]
[Final reasoning]
Common Mistakes
| Mistake | Why It Fails | Fix |
|---|---|---|
| Starting with brainstorming | Generates unusable names | Start with constraints (Gate 1) |
| Skipping TTS test | Voice assistant failures | Always test pronunciation |
| Ignoring Nice class overlap | Trademark conflicts | Check specific classes, not just exact matches |
| Choosing descriptive names | Weak AEO, hard to trademark | Prefer distinctive/invented |
| No narrative test | Brand can't tell its story | Always verify storytelling potential |
Quick Reference
The 7-8 Letter Rule: Shorter = better for mobile, voice, memory, tokenization.
Phonetic Gold Standard: CVCV pattern (consonant-vowel-consonant-vowel) works globally.
AEO Priority: If AI can't cite you as a distinct entity, you don't exist in 2026.
Empty Vessel Strategy: Meaningless names (Kodak) require more brand-building but face fewer conflicts.
Voice-First: If Alexa can't invoke it, 8.4 billion assistants can't recommend it.
What Claude Does vs What You Decide
| Claude handles | You provide |
|---|---|
| Generating candidates by typology | Brief constraints (markets, classes) |
| Running linguistic stress tests | Final pronunciation judgment |
| Checking trademark databases | Legal professional validation |
| Evaluating AEO/entity strength | Strategic brand direction |
| Producing naming dossier | Final selection and approval |
Skill Boundaries
This skill excels for:
- Brand, product, or company naming
- Names requiring international deployment
- Voice-assistant compatible names
- AI-discoverable entity creation
This skill is NOT ideal for:
- Internal codenames → No legal/AEO requirements
- Temporary campaign names → Overhead unnecessary
- Personal projects → Simplified process sufficient
Skill Metadata
name: naming-6-gates
category: branding
version: 2.0
author: GUIA
source_expert: 6-Gate Protocol (AEO-first naming)
difficulty: advanced
mode: centaur
tags: [naming, branding, trademark, voice-assistant, aeo, entity]
created: 2026-02-03
updated: 2026-02-03
Dépôt GitHub
Compétences associées
content-collections
MétaCette compétence propose une configuration éprouvée en production pour Content Collections, un outil axé sur TypeScript qui transforme des fichiers Markdown/MDX en collections de données typées de manière sûre avec une validation Zod. Utilisez-la lors de la création de blogs, de sites de documentation ou d'applications Vite + React riches en contenu pour garantir la sécurité de typage et la validation automatique du contenu. Elle couvre tout, de la configuration du plugin Vite et de la compilation MDX à l'optimisation des déploiements et la validation des schémas.
polymarket
MétaCette compétence permet aux développeurs de créer des applications avec la plateforme de marchés prédictifs Polymarket, incluant l'intégration d'API pour le trading et les données de marché. Elle fournit également une diffusion de données en temps réel via WebSocket pour surveiller les transactions en direct et l'activité du marché. Utilisez-la pour mettre en œuvre des stratégies de trading ou pour créer des outils traitant les mises à jour de marché en direct.
creating-opencode-plugins
MétaCette compétence aide les développeurs à créer des plugins OpenCode qui s'interconnectent avec plus de 25 types d'événements tels que les commandes, les fichiers et les opérations LSP. Elle fournit la structure du plugin, les spécifications de l'API événementielle et les modèles d'implémentation pour les modules JavaScript/TypeScript. Utilisez-la lorsque vous avez besoin d'intercepter, de surveiller ou d'étendre le cycle de vie de l'assistant IA OpenCode avec une logique personnalisée pilotée par les événements.
sglang
MétaSGLang est un framework de service LLM haute performance spécialisé dans la génération rapide et structurée pour les workflows JSON, regex et agentiques grâce à son cache de préfixe RadixAttention. Il offre une inférence nettement plus rapide, particulièrement pour les tâches avec des préfixes répétés, ce qui le rend idéal pour les sorties complexes et structurées ainsi que les conversations multi-tours. Choisissez SGLang plutôt que des alternatives comme vLLM lorsque vous avez besoin d'un décodage contraint ou que vous construisez des applications avec un partage étendu de préfixes.
