What problem does it solve?
It turns a brand’s website copy and visual system into a reusable, on-brand kit so other agents can generate accurate branded outputs without guessing.
Core Features & Use Cases
- Computed-DOM visual extraction: Uses live browser computed styles (not raw stylesheet variables) plus screenshots to confirm palette, typography, and UI values.
- Multi-mode inputs: Builds from a URL/hostname, from existing brand files, or from scratch with guided seed prompts.
- Actionable deliverables: Produces a durable markdown source of truth and a brand-styled HTML preview for easy review and downstream use.
- Auto-activation: After the first build, appends an idempotent instruction to /agent/user.md so future branded generations consult the saved kit automatically.
Quick Start
Run the skill with the argument URL, for example: Build a brand kit for https://example.com, and then use the generated markdown kit and HTML preview to guide all future branded outputs.