heritage-blueprint skill
A skill that makes an offline page with a blue-style 3D blueprint of a monument from photos and sources or merges models into one
- 82 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The skill runs local build and verification scripts and opens sources and photos you point to
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/heritage-blueprintDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The skill runs local build and verification scripts and opens sources and photos you point to
This is third-party code. Review the repository files before installing.
What it does
heritage-blueprint is a skill for making a 3D web page with a blueprint look of an architectural monument or heritage object. The agent gathers evidence from real photos, historical sources and available surveys, records the source of every dimension, builds geometry in plain WebGL and delivers one HTML file where rotation, pan, zoom, exploded view and reset work offline. Scripts, styles and fonts are embedded in the file. The skill separates recorded, reported, photo-estimated and unknown dimensions and does not pass a schematic off as a survey. A second workflow merges several finished models into one page, for example six Hangzhou landmarks in a three by two grid, keeping your views and labels.
Who it is for. Authors of historical and tourist pages, studios and museums that need interactive monument models with clear dimension sources.
Good fit when
- You make a page about a tower or bridge and want a rotatable model with labels from sources
- Dimensions must be marked: what came from a document, what from a publication and what from a photo estimate
- You need one offline HTML file with no CDN dependency for an exhibition or museum kiosk
- Six finished models must merge into one landmarks page with your views
- One object model in a combined page needs a fix and its embedded copy updating
Not a fit when
- You need an engineering survey, CAD or BIM: without survey data the skill does not call the result exact
- You expect a reconstruction from one description with no photos or sources
- You need a model of a modern building with working drawings
- The reference photos cannot be used: the repository does not distribute others photos without a license
Example request
Use $heritage-blueprint to make a Leifeng Pagoda model in Hangzhou, check photos and sources, mark the dimensions and build an offline pageLimitations
The documentation and SKILL.md are bilingual, with Chinese text first. Building a single object needs Python and Node: `scripts/build.py` builds the page and `scripts/verify.cjs` checks program invariants. Photo match, real WebGL and label layout are inspected by eye in a browser. The builder takes only single models, and a many-model page is assembled with a project-specific combiner. A model is not called an exact survey when no survey data exists. Rights to the examples and demo recordings are described in `COPYRIGHT.md`, and photos without redistribution permission are not in the repository.
How to disable. Delete the `heritage-blueprint` folder from the agent skills directory. Finished HTML files and the source ledger sit in your working folder and are deleted separately.
FAQ
How do I install heritage-blueprint?
Download the whole repository, name the folder `heritage-blueprint` and put it in your agent skills directory, for example `~/.claude/skills/`. One SKILL.md file is not enough: scripts and templates are required.
What is the output?
One HTML file that opens offline: rotate, pan, zoom, exploded view and reset. Scripts, styles and fonts are embedded with no CDN. Sources sit in embedded JSON and the ledger.
Can I merge several models?
Yes. Ask the agent to merge finished HTML files into one page, for example three by two, with your views and labels. The workflow is in `references/gallery.md`.
How does the skill handle unknown dimensions?
It separates recorded, reported, photo-estimated and unknown values and marks them with words like "about" and "reported". If data is missing the field stays empty rather than being filled with equal layers.
How is the result verified?
With `node <skill>/scripts/verify.cjs` for program invariants and by viewing the page in a browser: photo match, WebGL and labels are checked by eye.
Discussion
No questions or reviews yet.
Related
SkillsA skill rebuilds the object in a reference image as a procedural Three.js model written in code, not downloaded art
A skill that draws interactive isometric illustrations in SVG, with WebGL shaders and React output, explaining an idea as a working machine
A skill for Live2D projects: from layers and Cubism rigging to browser acceptance and rollback, with code and a sample model.
A skill that turns photos of buildings, streets and local culture into art-directed editorial posters via an image model
The scan-to-threejs skill refines a 3D scene in a loop, checking renders against room photos from matching camera angles.
A Claude Code skill that builds an offline 3D page of any machine you can take apart, with a recorded video tour.
README in short
The README states the goal in two sentences, shows an online Hangzhou landmarks example and a video. It then explains installation by copying the folder, two typical prompts, the output format and license terms for code, font and examples.
README badge
Original SKILL.md
--- name: heritage-blueprint description: "按实景照片、史料和测绘资料制作蓝图风格古建或文物三维 HTML，或将多个已有模型合并成单页。Build offline heritage models from photos and records, or combine existing models into one interactive page." --- # 古建与文物蓝图 · Heritage Blueprint 作者 / Author：[@木渡川（tanglele）](https://x.com/tanglele0318) · [MIT License](LICENSE) · [版权范围 / Scope](COPYRIGHT.md) 交付可离线打开的 HTML，支持旋转、平移、缩放、分解和复位。模型用原生 WebGL 渲染；脚本、样式和字体全部内嵌，不依赖 CDN 或外部资源。资料台账另存。 Build a self-contained HTML from the object's own evidence. Keep approved geometry and manual views. Mark estimates and unknowns; inspect the final page before delivery. ## 从哪里开始 | 任务 | 做法 | |---|---| | 新建或修改单体 | 按下面四步制作 | | 多个成品合成一页 | 读 [合页流程](references/gallery.md)，保留原模型，适配布局、视角和标注 | | 修改合页中的某个建筑 | 先改该单体，再更新嵌入副本 | 用户已调好的视角，先保存再刷新。已确认的对象、年代和样式直接沿用。 ## 1. 查资料 读 [证据与尺寸](references/evidence.md)。确定地点、年代和表现范围，打开实景照片、文保记录及可得的测绘或修缮图纸。照片要看原图，图纸要记页码、单位和基准。 每项形制和尺寸记录来源，分清一手记录、报道值、照片估计和未知项。资料冲突时核对年代、修缮和测量口径。缺数据就留空，不能用等分层高或通用塔身补成“已核准”。 核准塔类尺寸时，逐项填写 [尺寸核对表](references/evidence.md#dimension-register)：层高、收分、檐挑、台基、塔刹。区分内外层级、对边与对角尺寸、地坪与构件标高。示意图保留“约”“报道”“估计”等标识；用户要求精细核准的部位不能擅自简化。 ## 2. 建模 读 [场景接口](references/runtime.md)。用 `assets/blueprint.template.html` 保留蓝图界面和交互，在工作目录编写： - `scene.json`：标题、文件路径和动画时长。 - `geometry.js`：本对象的几何、构件层级、分解动作和标注锚点。 - `evidence.json`：来源、尺寸和未验证项。 - `inset.html`：需要时生成的静态形制示意。 先比对整体轮廓、层级和孔洞，再做砖缝、瓦垄等细节。示意图与主体共用参数；整体和局部放大分别标明。现有构建器只接收静态示意，单个竖直标尺也不能代替逐层或水平尺寸绘制。 孔洞须有真实空腔和厚度；构件接触、屋面穿插、柱脚与台阶问题按 [几何检查](references/geometry-qa.md) 处理。平滑曲面保留真实转折，隐藏计算网格，不加暗色光晕。 分解通过构件节点变换实现，默认时长 2.85 秒。展示位移不表示原建筑存在折叠机构。无测绘数据时，不称精确复原、数字孪生、CAD 或 BIM。