{hjemmesidekongen}/ai

våbenskjold

Brand strategy toolkit - create brands from scratch, audit existing brands, evolve them over time. Structured YAML guidelines that every other plugin can consume.

våbenskjold

The brand strategy layer of the hjemmesidekongen/ai methodology. Brand-before-code is the principle — våbenskjold makes sure every project starts with a codified brand that downstream tools can consume. Not colors and fonts (that's segl), but the strategic foundation: positioning, voice, values, and audience.

Version: 0.1.0 · 4 skills · 5 commands

What It Does

våbenskjold turns brand strategy into structured YAML files through interactive, multi-phase processes. Create a brand from scratch with market research, audit an existing brand from content samples, or evolve one that's drifted. Every output follows a strict schema — guideline.yml, voice.yml, values.yml — so other plugins can load and consume them without interpretation.

Commands

CommandUsageWhat It Does
/brand:create/brand:create acmeCreate a new brand from scratch through 7 interactive phases
/brand:audit/brand:audit acmeCodify an existing brand by analyzing content samples and materials
/brand:evolve/brand:evolve acmeGuided reinvention — preserve what works, change what doesn't
/brand:apply/brand:apply acme [--to PATH]Load brand into context or wire into a project
/brand:status/brand:status [acme]Show brand status and guideline summary

Create Workflow

/brand:create my-saas           # 7-phase interactive process
# Phase 1: intake — name, domain, vision
# Phase 2: market research — Perplexity-powered competitive analysis
# Phase 3: positioning & audience — differentiation, segments, content pillars
# Phase 4: voice discovery — archetype, scales, vocabulary, writing patterns
# Phase 5: values & beliefs — core values with in-practice behaviors
# Phase 6: dos and donts — generated from voice + values
# Phase 7: final review — complete summary, user approval required

Audit → Evolve Workflow

/brand:audit existing-brand     # Analyze content samples, extract patterns
# ... time passes, brand drifts ...
/brand:evolve existing-brand    # Diagnose what's off, scope changes, evolve

Key Concepts

Brand as Structured Data

Every brand produces the same four files at .ai/brand/{name}/:

guideline.yml — positioning and strategy:

  • Name, tagline, elevator pitch
  • Positioning: category, target audience, differentiator, promise
  • Audience segments with pain points, motivations, channels
  • Content pillars with percentage allocation (must sum to 100)
  • Optional: competitive positioning matrix, origin story

voice.yml — personality and writing style:

  • Personality archetype (Expert, Friend, Innovator, Guide, Motivator)
  • Five voice scales (1-10): formal↔casual, serious↔playful, technical↔simple, reserved↔expressive, humble↔confident
  • Writing patterns: sentence structure, paragraph organization, hooks
  • Vocabulary: signature phrases, power words, never-use words
  • Platform adaptations: tone shifts per channel (social, docs, marketing)
  • Anti-patterns: what the brand never sounds like

values.yml — decision framework:

  • 3-5 core values with definitions, in-practice behaviors (3-5 examples each), and violation examples
  • Distinctive beliefs with implications
  • Non-negotiables — lines the brand never crosses
  • Decision framework — priority order and explicit tradeoffs (e.g., "clarity over cleverness")

dos-and-donts.md — human-readable rules derived from voice + values. At least 5 dos and 5 donts with rationale.

Market Research (brand-strategy)

When creating a brand from scratch, Phase 2 runs 2-3 Perplexity searches to map the competitive landscape. The output goes to market-research.md:

  • Key competitors with positioning, strengths, weaknesses
  • Positioning gaps — underserved audience segments
  • Audience signals — who buys and why

This isn't generic market analysis — it's focused on finding the open positioning space for your brand.

Voice Quantitative Analysis (brand-audit)

When auditing an existing brand, våbenskjold analyzes 5+ content samples across the five voice scales. Each sample gets scored individually, then averaged. High variance across samples indicates inconsistent voice — the audit flags this explicitly.

Progressive Disclosure Loading

brand-loader serves brand data to downstream skills at three levels:

LevelWhat's LoadedWhen Used
L1 (discovery)Check if brand exists, list available brandsQuick existence checks
L2 (summary)Name, tagline, positioning summary from guideline.ymlContext-setting
L3 (full)Specific files based on consumer needActive consumption

L3 adapts to the consumer: content skills get voice.yml + dos-and-donts.md, strategy skills get guideline.yml + values.yml, design skills get guideline.yml for aesthetic alignment.

Evolution Tracking

brand-evolve produces an evolution-log.md documenting every change:

  • Before → after for each element
  • What was preserved and why
  • Rationale for each evolution decision

This creates a paper trail for brand decisions — useful when someone asks "why did we change the voice?"

Component Reference

Skills

SkillModelInteractivePurpose
brand-strategyPrincipalYesCreate brand from scratch via market research and sparring
brand-auditPrincipalYesCodify existing brand from content analysis
brand-evolvePrincipalYesGuided reinvention preserving brand equity
brand-loaderJuniorNoLoad brand with progressive disclosure (L1/L2/L3)

All interactive skills use Principal tier — brand strategy requires judgment, not just pattern matching.

Resource Schemas

Three YAML schemas in resources/ define the structure of brand outputs:

SchemaGoverns
guideline-schema.ymlPositioning, audience, content pillars
voice-schema.ymlArchetype, scales, vocabulary, writing patterns
values-schema.ymlCore values, beliefs, non-negotiables, decision framework

Output Files

.ai/brand/{name}/
├── guideline.yml          # Positioning and strategy
├── voice.yml              # Personality and writing style
├── values.yml             # Decision framework and beliefs
├── dos-and-donts.md       # Human-readable brand rules
├── market-research.md     # Competitive landscape (brand-strategy only)
├── audit-sources.md       # Source-to-finding citations (brand-audit only)
└── evolution-log.md       # Change tracking (brand-evolve only)

All YAML, all local, all version-controllable.

See also

  • segl — takes brand guidelines and creates visual identity + design tokens
  • kronen — foundation plugin that våbenskjold depends on
  • smedjen — content-writer agent consumes brand voice for copy generation
  • Architecture — how the five plugins relate

On this page