菜单
Turn a room photo, scan, or point cloud into a structured layout — walls, openings, furniture as typed objects with dimensions — using SpatialLM-class models, ready for blueprint-validate and re-design. Trigger: "here's a photo of my room", "measure this space", "redesign from this picture".
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/wjlgatech/design-anything --skill scene-to-layout命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Turns any design intent — a text brief or a picture — into an execution-ready, gate-verified blueprint: 3D-printable models (STL), construction floor plans, garment sewing patterns, game/sim scenes. Discovers what the user wants to design, retrieves survival-tested principles and constraint tables, emits deterministic parametric source, and verifies with machine gates (watertight, clearances, egress, zero-waste, fit). Use when the user wants to design, redesign, or validate a physical or spatial thing: "design me a…", "is this printable / buildable / cuttable", "turn this sketch or photo into a plan", "redesign my room / house / garden / apron / game level", "make it 3D-print ready". NOT for graphic/UI/logo design, photoreal rendering, or engineering certification (a gate is never a permit or PE stamp).
Drive blender-mcp for design-anything work — Drive Blender from any LLM via MCP — the natural execution backbone from agent to printable geometry. Use when a design task routes to this tool (agent-infrastructure). GENERATED satellite skill: facts fetched 2026-07-19; treat as a pointer, not gospel.
Drive GarmentCode (ETH) for design-anything work — Sewing patterns as programs — the substrate the AI pattern-generation wave targets. Use when a design task routes to this tool (garment-design). GENERATED satellite skill: facts fetched 2026-07-19; treat as a pointer, not gospel.
正在显示 SKILL.md
基于 SOC 职业分类
| name | scene-to-layout |
| description | Turn a room photo, scan, or point cloud into a structured layout — walls, openings, furniture as typed objects with dimensions — using SpatialLM-class models, ready for blueprint-validate and re-design. Trigger: "here's a photo of my room", "measure this space", "redesign from this picture". |
| kind | skill |
| license | CC-BY-4.0 |
| runtimes | ["claude-code","codex","hermes"] |
The input is a picture (the repo's second input modality) rather than text — an existing room/site that must become structured data before redesign.
"Photo of a 4x5m living room + 'the door is 80cm' reference" → layout JSON with scaled dimensions → "now add a reading corner with daylight" → redesign loop.
Output JSON must round-trip: rendered back to a 2D plan, wall topology is closed (rooms are bounded polygons) and openings lie on walls. An unclosed room polygon is a failed extraction — report it, don't patch it silently.
Never claim measured accuracy from an unreferenced photo; renovation decisions need a tape measure.
Model access varies by runtime; degrade to "ask user for measurements" when no vision/SpatialLM backend is available (skip-not-fail).
