The single entry point to Brand Skills. The other 16 skills are hidden from the agent's skill list
(disable-model-invocation: true) so they cost no context until needed; this skill routes to them.
How to run a brand skill
Read ../<skill>/SKILL.md (a sibling of this skill's directory) and follow it as if it had been
invoked. In that file, ${CLAUDE_SKILL_DIR} means its own directory. When a skill hands off to
another by name ("run brand-context"), load that one the same way — never through the Skill tool.
The user can still call any of them directly as /<skill>.
Start every brand with brand-init (no brand.yaml yet), then brand-context. Everything else reads
that package.
Routing
| Need | Skill |
|---|---|
New brand, no brand.yaml yet |
brand-init |
| Capture or update brand DNA | brand-context |
| Name a product, brand, app or bot | naming |
| Full strategy report (heart, archetype, values) | brand-strategy |
| Standalone vs sub-brand vs house of brands | brand-architecture |
| Competitive map and positioning statement | brand-positioning |
| ICP, personas, audience language | target-audience |
| How competitors brand themselves | competitor-branding |
| Logo direction, color, type, imagery brief | brand-identity |
| Tone, voice, vocabulary, writing rules | brand-voice |
| Taglines, value prop, key messages | brand-messaging |
| Origin and founder story | brand-story |
| The brand book | brand-guidelines |
| Brand health check | brand-audit |
| Reposition and relaunch an existing brand | rebranding |
| The brand inside Claude Code (spinner) | brand-mod |
Several needs at once → run them in pipeline order: context → naming → architecture → identity → strategy/positioning/audience/competitors → voice/messaging/story → guidelines.