Agents365-ai/drawio-skill

From text & real sources to maintainable .drawio architecture models: Diagram IR with source-kind profiles, incremental sync preserving manual layout, multi-view projection, architecture-as-test with a CI action, query/review, what-if, accessible Story Mode, and a built-in MCP server

What it solves

It eliminates the manual effort of creating and maintaining professional architecture diagrams. Instead of hand-drawing boxes and arrows, users can generate editable .drawio files from natural language prompts, existing source code, infrastructure-as-code (IaC) configurations, or database schemas. It also solves the problem of "diagram drift," where documentation becomes outdated as the system evolves, by providing tools for incremental synchronization and automated architecture testing.

How it works

The project acts as a bridge between structured data (or AI prompts) and the draw.io XML format. It uses a pipeline of extractors to pull data from sources like Python, JS/TS, Go, Rust, Terraform, Kubernetes, and SQL DDL. This data is then processed through an auto-layout engine (utilizing Graphviz and ELK) to determine optimal node placement and edge routing. For AI agents, it provides an MCP (Model Context Protocol) server and an "Agent Skills" package, allowing LLMs to call specific tools to build, sync, or query diagrams. It can also convert Mermaid text into native draw.io elements.

Who it’s for

  • Software Architects and Developers who need to visualize complex codebases, class hierarchies, or system dependencies.
  • DevOps Engineers wanting to turn Terraform or Kubernetes manifests into visual infrastructure maps with official cloud icons.
  • Technical Writers who need to maintain bilingual diagrams or convert diagrams into PowerPoint decks and interactive HTML walkthroughs.
  • AI Agent Users (Claude Code, Cursor, etc.) who want their AI assistant to generate and refine professional diagrams.

Highlights

  • Multi-Source Import: Generates diagrams from code (Python, JS, Go, Rust), IaC (Terraform, K8s, Docker Compose), and SQL schemas.
  • Incremental Sync: Updates diagrams without erasing manual layout adjustments or annotations.
  • Diagram-as-Test: Enforces architecture contracts via YAML/JSON rules (e.g., flagging unauthorized database access or cycles).
  • C4 Model Support: Automatically creates multi-page C4 diagrams with clickable drill-down navigation.
  • Rich Asset Library: Includes 10,000+ official shapes and 300+ AI/LLM brand logos not found in standard draw.io.
  • Advanced Visuals: Supports "Tube-Map Mode" for pipelines, animated data-flow SVGs, and architecture time-lapse videos.
  • Agent Integration: Built-in MCP server for seamless use with Claude Desktop, Cursor, and other MCP-compatible hosts.

Related

  • Project
    DayuanJiang/next-ai-draw-io

    An AI-powered tool that generates and edits draw.io diagrams using natural language, images, or documents, supporting multiple LLM providers and MCP integration.

  • Project
    yizhiyanhua-ai/fireworks-tech-graph

    An AI agent skill that converts natural language descriptions into professional, geometry-verified technical diagrams, SVG/PNG exports, and semantic animations.

  • Project
    nicobailon/visual-explainer

    An agent skill that converts complex terminal output, such as diagrams and tables, into styled, interactive HTML pages for better readability.

  • Project
    markdown-viewer/skills

    A collection of standardized skills for AI agents to generate professional diagrams, data visualizations, and technical documentation in Markdown.

  • Project
    sparklabx/drawio-ai-kit

    An orchestration and validation framework that enables AI agents to generate structurally precise and standardized draw.io diagrams for cloud architectures.