Design Context Extractor
Design context extraction: analyzes screenshots, URLs, or live apps to extract color palettes, typography, spacing, and component patterns as structured design tokens.
- Type
- Subagent
- Repository
- yonatangross/orchestkit
- GitHub stars
- 284
- License
- MIT
- Repo last updated
- Sep 27, 2026
- Source file
- plugins/ork/agents/design-context-extractor.md
- Model
- sonnet
What Design Context Extractor is
Design Context Extractor is a subagent published in the yonatangross/orchestkit repository on GitHub, which has about 284 stars. The repository describes itself as: “The Complete AI Development Toolkit for Claude Code. 106 skills, 36 agents, 171 hooks. Install `ork` for stable (v9.x), or `ork-alpha` for the v10 line, which ships daily.”
A subagent is a specialist assistant that Claude can hand part of a task to. It is a markdown file whose frontmatter sets a name, a description that tells Claude when to delegate, and optionally the tools and model it may use; the body becomes the subagent's own system prompt.
Because a subagent works in its own context, it keeps the main conversation focused: Claude can send a narrow job, such as a review or a specialised analysis, to Design Context Extractor and get back a compact result.
How to install Design Context Extractor
Claude Code
- Download design-context-extractor.md from the repository.
- Save it to ~/.claude/agents/ to use it in every project, or to .claude/agents/ inside one project to share it through version control.
- Claude Code watches these folders, so the subagent is usually available right away. Ask Claude to use it by name, or @-mention it to make sure it runs.
Claude Cowork
- Cowork loads subagents through plugins. If the repository is packaged as a plugin marketplace, add it under Customize → Plugins → Add marketplace and install the plugin that contains this subagent.
- Otherwise, bundle the file into your own plugin's agents/ folder and upload it from Customize → Plugins.
New to extending Cowork? Our plugins guide and Customize guide explain how skills, plugins, and connectors fit together.
Inside the source file
An excerpt from plugins/ork/agents/design-context-extractor.md, shared under the repository's MIT license. Read the full file on GitHub.
Directive
Extract design context ("Design DNA") from visual sources — screenshots, URLs, or existing project styles — and produce structured design tokens. Focus on precision: exact color values, measured spacing, identified typography.
Consult project memory for existing design decisions and token structures before starting. Persist extracted design context to project memory for future sessions.
Read existing token files, Tailwind config, and CSS custom properties before extracting. Compare extracted values against project's existing design system to identify gaps and conflicts.
When gathering context, run independent reads in parallel:
- Read existing token files → all in parallel
- Read Tailwind/CSS config → all in parallel
- Read component style patterns → all in parallel
Extract what's visually present. Don't invent token tiers or semantic aliases that aren't supported by the source material. A screenshot of a landing page doesn't need a full enterprise token architecture.
Agent Teams (CC 2.1.33+)
When running as a teammate:
- Send extracted tokens to design-system-architect for integration into the project's token hierarchy.
- Send component inventory to frontend-ui-developer for implementation planning.
- Use SendMessage to share color palettes and typography specs with teammates.
- Before any SendMessage to a peer outside your team, call ListAgents and address a listed name — never send to a guessed session name.
- A reply to any message you send to another session is delivered to your PARENT session's conversation, not to you; send and move on, never wait for an answer. Cross-session messaging works on Bedrock, Vertex and Foundry and with telemetry disabled, so a provider or DISABLE_TELEMETRY=1 is not a reason to fall back to polling files.
MCP Tools
- stitch — get_screen, get_project, generate_screen_from_text for AI-powered design extraction (official Google Stitch MCP)
- context7 — Tailwind CSS and Style Dictionary documentation
Without stitch: Falls back to multimodal image analysis (Claude can read screenshots directly) and CSS parsing for URLs.
Concrete Objectives
- Analyze visual source (screenshot, URL, or project styles)
- Extract precise color values in hex and oklch
- Identify typography hierarchy (families, weights, scale)
- Measure spacing patterns and establish scale
- Catalog component types present in the design
- Output structured tokens in requested format
Output Format
{
"source": "screenshot|url|project",
"input": "/path/or/url",
"extraction": {
"colors": { "count": 8, "space": "oklch" },
"typography": { "families": 2, "scale_steps": 8 },
"spacing": { "base": 4, "scale_steps": 8 },
"components": ["navbar", "hero", "card"]
},
"output_format": "tailwind|w3c|css|markdown",
"output_path": "path/to/generated/file",
"confidence": "high|medium|low"
}Task Boundaries
DO:
- Extract colors, typography, spacing from visual sources
- Convert colors to oklch for perceptual uniformity
- Detect existing project tokens and compare
- Output in multiple formats (Tailwind, W3C, CSS, Markdown)
- Use stitch MCP tools when available
DON'T:
- Implement components (that's frontend-ui-developer)
- Redesign existing UIs (that's design-system-architect)
- Modify backend code
- Create full design systems from a single screenshot
Standards
- OKLCH color space for all extracted colors
- W3C Design Token Community Group format for .tokens.json
- Three-tier token hierarchy when output includes semantic mapping
- WCAG 2.1 AA contrast verification for extracted color pairs
Integration
- Provides to: design-system-architect (token specs), frontend-ui-developer (component inventory)
- Receives from: screenshots, URLs, project style files
- Skill references: design-context-extract, design-system-tokens
Status Protocol
Report using the standardized status protocol. Load: Read("${CLAUDE_PLUGIN_ROOT}/shared/status-protocol.md").
Your final output MUST include a status field: DONE, DONE_WITH_CONCERNS, BLOCKED, or NEEDS_CONTEXT. Never report DONE if you have concerns. Never silently produce work you are unsure about.
Before you install
- Read the whole file first. Skills, commands, and subagents are instructions Claude will follow, so make sure they match what you want.
- Check which tools, scripts, or MCP servers it uses. Local servers and scripts run with your permissions.
- Try it in a test project or a copy of your files before pointing it at real work.
- Pin the version you tested, and review changes before updating.
- Watch for instructions that fetch web content or run shell commands; those are where prompt injection risks start. See our prompt injection guide.
FAQ
What is Design Context Extractor?
Design Context Extractor is a subagent for Claude Code and Claude Cowork from the yonatangross/orchestkit repository on GitHub. Design context extraction: analyzes screenshots, URLs, or live apps to extract color palettes, typography, spacing, and component patterns as structured design tokens.
How do I install Design Context Extractor in Claude Code?
Download design-context-extractor.md from the repository. Save it to ~/.claude/agents/ to use it in every project, or to .claude/agents/ inside one project to share it through version control. Claude Code watches these folders, so the subagent is usually available right away. Ask Claude to use it by name, or @-mention it to make sure it runs.
Can I use Design Context Extractor in Claude Cowork?
Cowork loads subagents through plugins. If the repository is packaged as a plugin marketplace, add it under Customize → Plugins → Add marketplace and install the plugin that contains this subagent. Otherwise, bundle the file into your own plugin's agents/ folder and upload it from Customize → Plugins.
Is Design Context Extractor safe to install?
It is a third-party community resource, not reviewed by Anthropic or this site. Read the source file first, check which tools and connectors it uses, and install only from sources you trust.
Similar resources
- Monitoring Engineer Observability and monitoring specialist. Prometheus metrics, Grafana dashboards, alerting rules, distributed tracing, log aggregation, and SLOs/SLIs. Subagent · yonatangross/orchestkit
- Ork The Complete AI Development Toolkit — 107 skills, 36 agents, 171 hooks for full-stack development. Plugin · yonatangross/orchestkit
- Database Engineer PostgreSQL specialist: schema design, migrations, query optimization, pgvector/full-text search, Alembic migrations. Subagent · yonatangross/orchestkit
- Multimodal Specialist Vision, audio, image and video generation, and multimodal processing specialist. Integrates Claude Opus 5.5, GPT-5, Gemini 2.5/3, GPT Image 2, Nano Banana Pro, Kling 3.0, Sora 2 and Veo 3.1 for analysis, generation, transcription and multimodal RAG. Subagent · yonatangross/orchestkit
- Design System Architect Design system architect: token hierarchies, theming strategies, component library design, Figma-to-code pipelines, and design governance. Subagent · yonatangross/orchestkit
- Deployment Manager Deployment specialist: production releases, rollback procedures, feature flags, blue-green/zero-downtime deployments. Subagent · yonatangross/orchestkit
- Emulate Engineer Stateful API emulation via Vercel emulate. Seeds GitHub/Vercel/Google/Slack/Apple/Entra/AWS/MongoDB/Okta/Resend/Stripe/Clerk/Linear, webhooks, port isolation, Next.js adapter. Use to replace flaky API mocks. Subagent · yonatangross/orchestkit
- Demo Producer Universal demo video producer that creates polished marketing videos for any content - skills, agents, plugins, tutorials, CLI tools, or code walkthroughs. Uses VHS terminal recording and Remotion composition. Subagent · yonatangross/orchestkit