Skill 详情
open-design
Useful for generating design prototypes and artifacts, but is tool/workflow-specific.
使用前先检查
自动化审核只检查相关性,不代表安全审查或推荐。使用前请阅读来源中的说明。
SKILL.md
这段内容是审核时保存的快照。外部来源才是完整且最新的版本。
--- name: open-design description: > Route design-generation work into the right Open Design workflow mode — prototype, deck, document, or media artifact. Use when the user needs local-first UI prototype generation, presentation deck creation, or design-system-aware HTML/PDF/PPTX artifacts via locally-installed coding agents. Supports 72 built-in design systems, 5 visual directions, 93 media prompt templates, and multi-format export. Triggers on: open-design, local design tool, prototype generation, design deck, design artifact, design agent workflow, open design prototype, design system artifact, UI prototype generation. allowed-tools: Bash Read Write Edit Glob Grep WebFetch compatibility: Requires Node.js ~24 and pnpm 10.33.x. Run `pnpm tools-dev run web` to launch the local UI. Auto-detects 13+ coding agent CLIs (Claude Code, Cursor, Gemini CLI, GitHub Copilot, etc.). Extend with file-based skills in the `skills/` directory. Plugin install via `claude plugin marketplace add nexu-io/open-design`. license: MIT metadata: tags: open-design, design-tool, local-first, prototype, deck, ui-generation, multi-format, design-system, media-generation, html-export, pptx-export version: "1.0" source: https://github.com/nexu-io/open-design platforms: Claude, Gemini, Codex, ChatGPT --- # open-design — local-first design artifact generation > **Keyword**: `open-design` · `local design tool` · `prototype generation` · `design artifact` > > Open-source alternative to Anthropic's Claude Design — generates web, mobile, and desktop prototypes, decks, and media artifacts using coding agents already installed on your machine. Open Design is most useful when you choose the **smallest generation mode that fits the job**: - start with a prototype before reaching for a full deck - pick one design system before layering custom styles - choose one output format (HTML review before PDF/PPTX export) - use the Turn 1 discovery form to reduce redirect cycles before generation begins ## When to use this skill - The user needs a local-first UI prototype, presentation deck, or document artifact from a single prompt - The user wants to generate design artifacts using coding agents they already have installed (Claude Code, Cursor, Gemini CLI, etc.) - The user wants to apply one of 72 built-in design systems (Linear, Stripe, Vercel, Notion, Apple, etc.) to their output - The user needs multi-format export: HTML review → PDF, PPTX, ZIP, or Markdown - The user wants AI-generated images, video, or audio media embedded in design artifacts - The user wants to extend Open Design with custom file-based skills or install it as a Claude Code plugin ## Do not use this skill when - The task is live UI component authoring in a real codebase → route to `stitch-skills` or `design-system` - The task is video production or Remotion animation → route to `video-production` - The task is a data-driven presentation deck without design-system application → route to `presentation-builder` - The task is compressing or optimizing existing media files → route to `compresso` - The task is a broad web UI critique or audit → route to `web-accessibility` ## Instructions ### Step 1: Capture the generation mode Collect the minimum packet before choosing a workflow: - **Output type**: UI prototype, presentation deck, landing page, document, or media artifact - **Design system**: one of the 72 built-ins (Linear, Stripe, Vercel, Notion, Tesla, Apple…) or custom - **Visual direction**: one of 5 directions (minimal, bold, playful, enterprise, dark) - **Export format**: HTML review, PDF, PPTX, ZIP, or Markdown - **Agent**: which installed CLI to delegate to (auto-detected or explicit) Use this rule set: 1. If output is a UI prototype or landing page → choose `prototype` mode 2. If output is a presentation or investor deck → choose `deck` mode 3. If output is a long-form document or spec → choose `document` mode 4. If media (image/video/audio) generation is the primar在 GitHub 阅读完整来源 (打开外部页面)