AI Power Tools for Sparx EA Docs Release 3.6.4

AI Power Tools for Sparx EA documentation

AI Power Tools for Sparx EA lets an AI assistant work directly on a live Sparx Enterprise Architect repository. It has two parts: the AI Power Tools MCP Server for Sparx EA, which gives the assistant 149 operations on your model and comes from a Claude Desktop extension, and 24 skills that tell the assistant how to do each kind of job well and come from a Claude plugin on your account. This documentation describes release 3.6.4.

Put it to work

AI-augmented architecture puts AI to work on four parts of an architect's job. Pick the one in front of you. Each group of AI Augmentation Scenarios walks through real pieces of that work, step by step: what to ask for, where you decide, and how to check the result.

  • Modeling. Get the architecture into the repository and keep it there: build a model from source documents, load a hierarchy, compose and tidy diagrams, and move content in and out of Enterprise Architect.
  • Analysis. Answer questions from the model itself: what depends on an application before you change it, how a requirement reaches the systems that realize it, and what changed, when and by whom.
  • Governance. Keep the model conformant and under control: validate against your modeling standard, catch a non-conforming change before it is promoted, clean up decay, and own your modeling language.
  • Stakeholder Engagement. Turn the repository into answers people can act on: portfolio answers in business terms, a reporting database, Power BI, and Prolaborate dashboards.

What you can do with it maps the same four uses to the tools and skills behind them.

Get set up

  1. Choose your packages: the Claude Desktop extension for the server and the Claude plugin for the skills, or the VS Code extension for GitHub Copilot.
  2. Check capabilities and surfaces to see what your AI client can do, and the requirements.
  3. Install and verify, then start a session safely before your first piece of work.

How to trust the result

The AI plans and drives the work. AI Power Tools does the computing and checking, and gives the same answer for the same question on the same model state. Where the AI decides and where AI Power Tools decides explains the split. Every scenario ends with Checking the result: how to tell the answer is accurate, that it is complete, and that nothing changed that should not have. How to trust the result gives the six ways to check and a recipe for any task. Before a bulk or destructive change, take a baseline.

Reading the steps

Each step in a skill or scenario says who does it:

  • AI: the AI does it, using its judgment.
  • AI Power Tools: an operation of the AI Power Tools MCP Server for Sparx EA does it, with a fixed, repeatable result.
  • Engine: a code engine computes it, which needs local code execution.
  • AI with computer use; otherwise you: the AI does it where your client gives it computer use, and hands the step to you where it does not.
  • You: a decision or approval that is yours.

Choose your AI client in the profile selector at the top of the page, and the steps and pages show how they run for you. Where the AI decides and where AI Power Tools decides explains the difference.

Getting help

For AI assistants

The whole reference is also published as a single Markdown file with each release, for AI assistants to read: AI-Power-Tools-for-Sparx-EA-Reference.md.

AI Power Tools for Sparx EA · Release 3.6.4 Full reference for AI assistants (Markdown)