All skills
jwynia avatar

/character-naming

@99a8797
by J Wyniajwynia/agent-skills160 stars
20

Break LLM name defaults with external entropy. Use when character names cluster around statistical medians (Chen, Patel, Maya, Marcus), when cast has collision risks, or when fantasy cultures need phonologically consistent naming.

Use this Skill: https://skilld.dev/gh/jwynia/agent-skills/character-naming

This session only. Nothing lands on disk.

naming-framework.md

≈5.5k tokens on demand. Your agent reads this file only when SKILL.md points to it.

Naming Framework

A diagnostic framework for creating names that work—brands, products, characters, places, titles.

Problem Statement

Names fail in predictable ways: they don't feel right, they don't fit together, they're forgettable, they send wrong signals. The difference between a name that works and one that doesn't often comes down to invisible patterns—sound, meaning, cultural resonance—that most people can sense but can't articulate.

Core Insight: Names Operate on Multiple Layers

Every name communicates through layers that work together or against each other:

  1. Sound Layer — How it sounds, feels in the mouth, and what sounds evoke
  2. Meaning Layer — Denotation, connotation, metaphor, reference
  3. Cultural Layer — What associations it triggers in specific audiences
  4. Functional Layer — How it works in use (typing, speaking, remembering)

When layers align, names feel inevitable. When they conflict, names feel wrong even if no one can say why.


The Naming States

State N1: Name Doesn't "Feel Right"

Symptoms: Stakeholders reject names but can't articulate why. Names feel "off" without clear reason. Gut reactions are negative despite meeting requirements.

Key Questions:

  • Do the sounds match the intended emotional tone?
  • Is there a meaning conflict between layers?
  • Does it violate cultural expectations for this category?

Common Causes:

  • Sound-meaning mismatch (harsh sounds for gentle brand)
  • Cultural layer conflict (sounds like something negative in context)
  • Functional friction (hard to say, spell, or type)

Interventions:

  • Analyze sound layer independently (see Sound Patterns below)
  • Check for unintended associations
  • Test pronunciation and typing flow

State N2: Names Don't Belong Together

Symptoms: Product family feels disjointed. Character names could be from different worlds. Place names lack coherent cultural identity.

Key Questions:

  • Is there a consistent sound palette?
  • Do syllable structures match?
  • Is there a unifying pattern (prefix, suffix, rhythm)?

Common Causes:

  • No phoneme inventory defined
  • Mixed linguistic origins without intent
  • Ad-hoc naming without system

Interventions:

  • Define phoneme inventory (which sounds are "in" this system)
  • Establish syllable templates (CV, CVC, CVCV, etc.)
  • Create naming conventions document
  • Regenerate outliers to fit system

State N3: Name Is Forgettable

Symptoms: People can't recall the name. It blends into category. No distinctive hook.

Key Questions:

  • Is there a memorable sound pattern (alliteration, rhythm)?
  • Does it have a clear meaning hook?
  • Is it too similar to competitors/alternatives?

Common Causes:

  • Too generic (follows category conventions too closely)
  • No distinctive sound feature
  • Meaning is abstract without concrete anchor

Interventions:

  • Add sound distinctiveness (unusual but pronounceable combination)
  • Create meaning hook (metaphor, unexpected reference)
  • Test against competitors for differentiation

State N4: Name Sends Wrong Signals

Symptoms: Audience interprets name differently than intended. Wrong category assumptions. Unintended associations.

Key Questions:

  • What does this name sound like it should be?
  • What category conventions is it following/breaking?
  • Are there unfortunate associations in target markets?

Common Causes:

  • Sound patterns associated with different category
  • Cultural reference not understood by audience
  • Unintended meaning in other languages/contexts

Interventions:

  • Audit category sound conventions
  • Test with target audience for associations
  • Check international/cross-cultural meanings

State N5: Name Doesn't Work in Practice

Symptoms: People misspell it. They mispronounce it. It's hard to type. Domain isn't available.

Key Questions:

  • Is spelling intuitive from pronunciation?
  • Are there common mistypings?
  • Does it work in all required contexts (URL, voice search, etc.)?

Common Causes:

  • Spelling-pronunciation mismatch
  • Unusual letter combinations
  • Conflicting with common words for autocorrect

Interventions:

  • Test spelling from dictation
  • Check typo patterns
  • Verify domain and social handle availability
  • Test voice search recognition

The Four Layers

Sound Layer

How the name sounds and what sounds communicate.

Sound-Meaning Connections

Certain sounds carry consistent associations across languages (sound symbolism):

Sound Pattern Association Examples
Depth sounds (ɑ, o, u, m, n) Depth, weight, seriousness "Om", "profound", "doom"
Light sounds (i, e, l, s) Speed, lightness, precision "swift", "sleek", "elite"
Power sounds (k, t, p, x) Strength, hardness, impact "strike", "crack", "apex"
Flow sounds (l, r, w, vowel clusters) Movement, continuity "flow", "glide", "aurora"
Tech sounds (x, z, -ix, -ex) Technical, modern, digital "Linux", "Xerox", "codex"
Phoneme Selection

Which sounds to include or exclude shapes the name's character:

High-frequency sounds (feel natural, trustworthy):

  • Consonants: t, n, s, k, m, p, l, r
  • Vowels: a, i, e, o, u

Low-frequency sounds (feel distinctive, exotic):

  • Consonants: x, z, q, zh
  • Vowels: ü, ø, æ (in English contexts)

Principle: Use common sounds for accessibility, rare sounds for distinctiveness. Too many rare sounds = unpronounceable.

Syllable Structure

Syllable patterns affect pronounceability and feel:

Pattern Feel Example
CV (consonant-vowel) Open, flowing "Sora", "Kano"
CVC Solid, complete "Mark", "Bond"
CVCV Balanced, memorable "Toyota", "Coca"
CVCC Weighty, serious "Ernst", "Kraft"
CCV Dynamic, energetic "Sprite", "Slack"

Principle: Simpler syllable structures = easier pronunciation across languages.


Meaning Layer

What the name denotes, connotes, and references.

Meaning Types
Type Description Example
Descriptive Says what it is "General Motors", "Toys R Us"
Metaphorical Implies qualities through comparison "Amazon" (vast), "Apple" (simple, fresh)
Abstract Coined word, meaning assigned "Xerox", "Kodak"
Founder/Place Named after person or location "Ford", "Brooklyn Brewery"
Acronym Initials of longer name "IBM", "BMW"
Portmanteau Blend of words "Pinterest" (pin+interest)
Meaning Layers

Names can carry multiple meaning levels:

  1. Literal — Direct denotation
  2. Metaphorical — What it implies about qualities
  3. Cultural — References to shared knowledge
  4. Personal — Individual associations

Best names work on multiple levels, rewarding deeper attention.


Cultural Layer

What associations the name triggers in specific audiences.

Category Conventions

Every category has naming conventions that signal membership:

Category Convention Examples
Luxury fashion French/Italian sounds "Hermès", "Versace"
Tech startups Dropped vowels, -ly, -ify "Tumblr", "Spotify"
Law firms Partner surnames "Sullivan & Cromwell"
Pharmaceuticals X, Z, scientific suffixes "Xanax", "Prozac"
Fantasy characters Apostrophes, unusual clusters "Drizzt", "Kvothe"

Strategic choice: Follow conventions to signal belonging, or break them to differentiate.

Cultural Resonance

Names carry cultural weight:

  • References — Allusions to mythology, literature, history
  • Language origin — Latin (authority), Greek (science), Japanese (precision)
  • Sound associations — What this sounds like in the culture

Functional Layer

How the name works in practical use.

Functional Tests
Test Question Failure Mode
Spelling Can people spell it from hearing? Constant correction needed
Pronunciation Can people say it from reading? Name avoidance
Typing Is it easy to type quickly? Typos, frustration
Searchability Does it return relevant results? Lost traffic
Domain Is .com (or relevant TLD) available? Compromised identity
Voice Does voice search recognize it? Missed commands
Domain Strategy

For brand names requiring web presence:

Priority TLDs:

  1. .com — Still default, highest trust
  2. .io/.ai — Acceptable for tech
  3. Country codes — For regional brands
  4. New TLDs — Lower recognition but increasing

If exact .com unavailable:

  • Add word: "get[name].com", "[name]app.com"
  • Spelling variant (if intuitive)
  • Consider whether name is worth the limitation

The Naming Process

Phase 1: Requirements

Before generating names, establish:

  1. Purpose — What is being named? What does it do?
  2. Audience — Who will use/encounter this name?
  3. Tone — What emotional register? (Serious, playful, technical, warm)
  4. Constraints — Length limits, required sounds, forbidden associations
  5. Context — Where will the name appear? What's around it?

Phase 2: Pattern Definition

Define the naming system:

  1. Sound palette — Which phonemes to favor/avoid
  2. Syllable templates — Allowed syllable structures
  3. Meaning direction — Descriptive, metaphorical, abstract, etc.
  4. Cultural register — What conventions to follow/break

Phase 3: Generation

Generate candidates systematically:

  1. Sound-first — Start with sound patterns, find meanings
  2. Meaning-first — Start with concepts, find sounds
  3. Reference-first — Start with cultural touchstones, adapt
  4. Combination — Blend words, roots, morphemes

Generate quantity: 50-100 raw candidates before filtering.

Phase 4: Evaluation

Filter candidates through layers:

Layer Pass Criterion
Sound Pronounceable, appropriate feel, memorable
Meaning Clear enough, positive associations, layered
Cultural No negative associations, appropriate register
Functional Spellable, typeable, available (if required)

Phase 5: Validation

For finalists:

  1. Audience testing — Do target users respond appropriately?
  2. Competitive check — Distinct from alternatives?
  3. Legal check — Trademark availability
  4. Technical check — Domain, social handles, etc.
  5. International check — No negative meanings in key markets

Professional Naming Process (Sequential)

For software products, companies, and brands, the naming process works significantly better when phases are separated. Do not combine phases—complete each fully before moving to the next.

Why Sequential Matters

When phases are mixed:

  • Pattern discovery is constrained by premature evaluation
  • Synthesis is rushed before patterns are fully explored
  • Evaluation criteria aren't established before judging begins
  • Results are mediocre—names that "work" but don't resonate

When phases are separated:

  • Discovery explores freely without judgment
  • Synthesis builds on complete pattern understanding
  • Evaluation applies consistent criteria
  • Results are names that feel inevitable

Rule: Complete each phase. Document the output. Only then proceed.


Phase 1: Pattern Discovery

Goal: Explore the naming space without generating names yet.

Duration: Separate session. Don't rush.

Process:

  1. Base Pattern Exploration

    • Core concepts and values of what's being named
    • Key activities and actions it performs
    • Target audience characteristics
    • Emotional resonance desired
    • Industry terminology and conventions
    • Related metaphors and analogies
    • Sensory associations
    • Cultural reference points
  2. Sound-Structure Exploration

    • Sound-meaning connections relevant to this project
    • Structural patterns that might work (length, syllables)
    • Cultural-linguistic resonance for target audience
    • Formation principles (what makes names in this space work)
  3. Cultural-Linguistic Mapping

    • Group-specific language patterns
    • Industry expressions and jargon
    • Identity markers for target audience
    • Shared experiences to reference
  4. Multi-Layer Meaning Exploration

    • Visual-verbal bridge opportunities
    • Multiple interpretation possibilities
    • Reference integration options

Output: Pattern documentation. No names yet—just the landscape.

Completion Check:

  • Base patterns documented
  • Sound-structure patterns identified
  • Cultural context mapped
  • Meaning layer opportunities noted
  • Ready for synthesis (not before)

Phase 2: Pattern Synthesis

Goal: Generate name candidates by combining discovered patterns.

Prerequisite: Phase 1 complete and documented.

Process:

  1. Level-Based Exploration

    Work through distinct levels, tracking evolution:

    Level Focus Examples
    Meta Vision, purpose, identity What this represents at highest level
    Macro Domain, category, positioning Where it sits in the landscape
    Core Function, activity, interaction What it actually does
    Micro Expression, implementation Specific word choices
  2. Sound-Structure Synthesis

    • Combine sound patterns with meaning
    • Test rhythm and flow
    • Check pronunciation across contexts
    • Verify cultural fit
  3. Layer Integration

    • Build meaning layers into candidates
    • Test multiple readings
    • Assess accessibility at different levels
  4. Candidate Generation

    • Sound-first approach (patterns → meanings)
    • Meaning-first approach (concepts → sounds)
    • Reference-first approach (touchstones → adaptations)
    • Combination approach (blends, portmanteaus)

Output: 50-100 raw candidates with notes on pattern origins.

Completion Check:

  • All levels explored
  • Multiple synthesis approaches used
  • Candidates documented with pattern rationale
  • Quantity sufficient for meaningful filtering
  • Ready for evaluation (not before)

Phase 3: Pattern Evaluation

Goal: Filter candidates through systematic criteria.

Prerequisite: Phase 2 complete. Candidates exist.

Process:

  1. Evidence Classification

    Separate what you can verify internally from what requires external checking:

    Internally Verifiable:

    • Pattern alignment with project goals
    • Linguistic analysis (structure, pronunciation)
    • Meaning analysis (layers, resonance)
    • Cultural-linguistic effectiveness

    Externally Verifiable (Phase 5):

    • Domain availability
    • Trademark status
    • Social handle availability
    • Market conflicts
  2. Layer-by-Layer Evaluation

    Layer Criteria Pass/Fail
    Sound Pronounceable, appropriate feel, memorable
    Meaning Clear, positive associations, layered
    Cultural No negative associations, right register
    Functional Spellable, typeable, searchable
  3. Pattern Strength Assessment

    • How well does candidate express discovered patterns?
    • Is the pattern-to-name translation clear?
    • Does it feel inevitable or forced?
  4. Ranking Development

    • Tier 1: Strong across all layers
    • Tier 2: Strong in most, fixable weaknesses
    • Tier 3: Interesting but significant issues
    • Cut: Fatal flaws

Output: Ranked shortlist with documented rationale.

Completion Check:

  • All candidates evaluated against layers
  • Internal verification complete
  • Rankings established with reasoning
  • External verification needs documented
  • Ready for validation (not before)

Phase 4: Validation

Goal: Verify finalists against external reality.

Prerequisite: Phase 3 complete. Shortlist exists.

Process:

  1. Technical Validation

    • Domain availability (see domain-naming.md)
    • Social handle availability
    • App store name availability (if relevant)
  2. Legal Validation

    • Trademark search
    • Business name registration check
    • Existing brand conflicts
  3. Market Validation

    • Similar names in use
    • Competitor analysis
    • Industry pattern fit
  4. Audience Validation

    • Target user response testing
    • First impression testing
    • Pronunciation testing
    • Recall testing
  5. International Validation (if relevant)

    • Meaning in other languages
    • Pronunciation across accents
    • Cultural associations in key markets

Output: Validated finalist(s) with complete assessment.

Completion Check:

  • Technical availability confirmed
  • Legal risks assessed
  • Market positioning validated
  • Audience response tested
  • International issues checked
  • Ready for decision

Phase 5: Documentation & Handoff

Goal: Document the decision and rationale for future reference.

Process:

  1. Decision Narrative

    • Why this name emerged from the process
    • Key pattern connections
    • How it expresses project values
    • What makes it work
  2. Implementation Guidance

    • Pronunciation guide
    • Usage guidelines
    • Variant handling (nicknames, abbreviations)
    • What to avoid
  3. Asset Documentation

    • Secured domains
    • Registered handles
    • Trademark status
    • Style guidance

Output: Complete naming documentation package.


Application: Brand Names

Sound Patterns for Brand Attributes

Attribute Sound Approach
Trustworthy Common phonemes, simple syllables, familiar patterns
Innovative Unusual combinations, invented words, tech sounds
Luxurious French/Italian sounds, flowing patterns, soft consonants
Powerful Hard consonants, short syllables, strong stops
Friendly Open vowels, soft consonants, nickname-able

Naming Conventions by Sector

Documented in domain-naming.md for:

  • Technology
  • Consumer products
  • Financial services
  • Healthcare
  • Professional services

Application: Character Names

Consistency Within Culture

Characters from the same culture should share:

  • Phoneme inventory — Same sounds appear
  • Syllable patterns — Similar structures
  • Naming conventions — Patronymics, honorifics, etc.

Use the conlang skill for systematic character naming when building fictional cultures.

Character-Sound Alignment

Match sounds to character attributes:

  • Heroic names often have strong consonants, open vowels
  • Villainous names may have harsh sounds, unusual clusters
  • Gentle characters benefit from soft consonants, flowing patterns

Subversion: Breaking these patterns can create interesting dissonance (soft-named villain, harsh-named hero).


Application: Fiction Titles

See fiction-titles.md for:

  • Genre-specific title patterns
  • Sound-emotion integration
  • Symbolic layering techniques

Anti-Patterns

The Kitchen Sink

Pattern: Trying to communicate everything in one name. Problem: Overcomplicated, hard to remember, no clear signal. Fix: Pick one primary message; let other layers be subtle.

The Inside Joke

Pattern: Name meaningful only to creators. Problem: Audience doesn't get it; seems random. Fix: Test with naive users; meaning should be discoverable.

The Sound-Alike

Pattern: Too similar to existing name in category. Problem: Confusion, legal issues, no differentiation. Fix: Check competitors thoroughly; verify distinctiveness.

The Forced Meaning

Pattern: Name that "means" something only via tortured explanation. Problem: No one gets it without the explanation. Fix: If meaning requires explanation, it doesn't work as meaning.

The Unpronounceable

Pattern: Looks interesting but no one can say it. Problem: Word of mouth dies; people avoid saying it. Fix: Test pronunciation with naive readers; simplify clusters.

The Apostrophe Catastrophe

Pattern: Apostrophes used for exotic feel with no linguistic logic. Problem: Inconsistent, confusing, hard to type/search. Fix: If using apostrophes, define what they mean; use sparingly.


Integration with Other Frameworks

Framework Connection
conlang Use phoneme inventories for consistent character/place naming
cliche-transcendence Apply to naming patterns—avoid default names for roles
worldbuilding Names should reflect cultural evolution and history
sensitivity-check Audit names for unintended associations

Sources

This framework synthesizes concepts from:

  • Sound symbolism research — Klink, Sapir, and others on sound-meaning connections
  • PHOIBLE database — Cross-linguistic phoneme frequency data
  • Brand naming practice — Landor, Lexicon, professional naming methodology
  • Constructed language design — Conlang community principles
  • Cognitive psychology — Memory, recognition, and processing fluency research

Completion Indicators

A name is ready when:

  • All four layers (sound, meaning, cultural, functional) are aligned
  • Stakeholders respond positively without explanation needed
  • Target audience testing shows appropriate associations
  • Functional requirements met (domain, legal, etc.)
  • No fatal conflicts in key markets
  • Fits within naming system (if part of family)

Source: SKILL.md on GitHub

1 warning14d5 checks · Risk SAFE
  • Gen Agent Trust Hub14d

    The character-naming skill provides a robust framework for generating character names using curated data pools and phoneme patterns to avoid common LLM biases. It utilizes local Deno scripts to perform sound analysis and manage cast lists for collision detection. The skill adheres to security best practices, utilizing trusted sources for its minimal external dependencies and maintaining operations within its local environment.

  • Socket14d

    No alerts

  • Snyk14d

    Risk: LOW · No issues

  • Runlayer7mo

    58/58 files flagged

  • ZeroLeaks5mo

    Score: 93/100 · 2 sections analyzed

Signed by skilld at 99a8797. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 2 months ago.

Dormantupdated 8 months ago
Other metadata
metadata
{
  "author": "jwynia",
  "version": "1.0",
  "type": "generator",
  "mode": "generative",
  "domain": "fiction"
}

README badge

README badge for jwynia/agent-skills/character-naming