En

Agentcad Skill Release 0.4.0

A CAD modeling tool for AI agents, generating 3D models from code.

开发流程 设计3D模型生成3D模型用Python脚本生成3D模型 通用 ★ 1.4k Updated 2026-08-02

Install & Use

Copy this prompt and send it to your AI assistant (Claude / Cursor / TRAE / Codex / WorkBuddy etc.) to auto-install:

Help me install this AI Skill: Agentcad Skill Release 0.4.0.
It is used for: A CAD modeling tool for AI agents, generating 3D models from code.
Full Skill content: https://321skill.com/skills/agentcad-skill-release-0-4-0-x-7/raw/index.md
Read that page and install it.

The prompt includes a link to the full Skill content. You can also view the full content.

This Skill addresses the lack of direct tools for AI agents when 3D model design is required. Traditional CAD software relies on manual operation and cannot be controlled automatically via code. Agentcad enables AI agents to directly write Python scripts (based on build123d or CadQuery) and execute them to rapidly generate 3D geometry, outputting formats like STEP, STL, and GLB, along with multi-view preview images and geometric measurements.

Usage is straightforward: simply describe your design requirements in natural language, and the AI agent will generate a compliant Python script, executed via the agentcad run command. It supports dry-run to preview volume, dimensions, and other metrics. After a formal run, you receive a four-view preview (front, right, top, isometric) and an interactive 3D viewer. You can also use the measure command to measure feature dimensions, check-spec to validate requirements, and diff to compare effects between versions.

It is ideal for AI agent developers, mechanical engineers, 3D printing enthusiasts, and teams looking to integrate CAD modeling into automated workflows—especially users already familiar with Python programming who prefer designing components with code rather than mouse operations, significantly improving iteration efficiency.

It is recommended to perform a dry-run check before generating scripts to avoid wasting versions. Note that scripts must include a show_object(result) call; otherwise, no output will be produced. Output files are saved to a folder named after the output label for later viewing or iteration. Currently, only build123d and CadQuery libraries are supported; other CAD engines or direct modeling operations are not supported.

Key Features

Unlike graphical CAD tools like Fusion 360, Agentcad is entirely code-script-based, making it suitable for automatic generation and iteration by AI agents. Compared to OpenSCAD, it supports more modern libraries like build123d and CadQuery, with syntax closer to standard Python programming conventions, and includes built-in helper functions such as `show_object`, eliminating the need for manual imports.

Limitations

Requires Python 3.10–3.12 environment and must have the agentcad library installed (`pip install agentcad`). The size and complexity of generated models are limited by memory and script execution time; real-time interactive modeling is not supported.

FAQ

Which Python versions does Agentcad support?

Officially requires Python 3.10, 3.11, or 3.12; other versions may not work correctly.

How can I view the generated model?

Each successful run generates a `viewer.html` file—open it in a browser for interactive 3D model preview. A composite four-view preview image, `preview.png`, is also output.

Installation guide for AI assistants

If your AI coding assistant (Claude Code, Cursor, TRAE etc.) can see this page, send it this message to auto-install:

Visit https://321skill.com/skills/agentcad-skill-release-0-4-0-x-7/raw/index.md to read the original Skill definition (Markdown format) for Agentcad Skill Release 0.4.0, and install it according to the instructions.