kta-design-spec

Generate per-screen design specs from feature PRDs into plans/{slug}/design.

4|Updated May 7, 2026
One-click install
npx skills add https://github.com/dantech0xff/dreams --skill kta-design-spec
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: kta-design-spec
Source: https://github.com/dantech0xff/dreams/tree/main/.agents/skills/kta-design-spec
Command: npx skills add https://github.com/dantech0xff/dreams --skill kta-design-spec

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

Automates the generation of per-screen design specs from per-feature PRDs, enabling consistent, tool-agnostic briefs for AI design tooling.

Core Features & Use Cases

  • Generates one spec per screen from the PRD set and groups them by feature.
  • Enforces a mandatory six-state component model in every spec and validates token references to style foundation.
  • Produces an INDEX with feature-to-screen mapping and a final design-bundle MD suitable for Stitch / Figma AI / Claude.
  • Integrates with plans/{slug}/prd/ inputs and outputs into plans/{slug}/design/.

Quick Start

Invoke the kta-prd-pipeline to supply per-feature PRDs and run this design-spec workflow to emit per-screen specs.

Frequently Asked Questions about kta-design-spec

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate per-screen design specs from a PRD for AI tooling?

To generate per-screen design specs from a PRD, you need a tool that converts feature documents into structured, tool-agnostic briefs. This process loops across features and screens to enforce a mandatory six-state component model and validates token references for AI design tooling.

What is a six-state component model in AI design specs?

A six-state component model in AI design specs mandates that every UI component lists six distinct states. This validation ensures AI design tools receive complete interaction briefs, preventing incomplete component generation from your markdown PRD inputs.

How do I convert markdown PRDs into a design-bundle for Stitch or Figma AI?

Converting markdown PRDs into a design-bundle for Stitch or Figma AI requires parsing feature documents and grouping screens by feature. The output includes an INDEX with feature-to-screen mapping and a final design-bundle markdown file.

Does this design spec workflow output Kotlin or Compose code?

No, this design spec workflow explicitly enforces no Kotlin or Compose code output. It strictly produces markdown design specs and a design-bundle to serve as tool-agnostic briefs for AI design tooling.

What is the best way to structure design tokens for automated spec generation?

The best way to structure design tokens for automated spec generation is using a style-foundation markdown file. The spec generation process validates references against these style-foundation tokens to ensure consistent design application across all screens.

Can I use this per-screen spec workflow without a dedicated PRD pipeline?

You can use this workflow by manually supplying per-feature PRDs in the plans directory, but it is designed to integrate with a PRD pipeline. Running a pipeline ensures the required markdown PRD inputs are correctly structured before generating the design-spec set.