AI with MattLearn as you work.Member area

A useful skill starts
with a real task.

You don’t need to learn every AI tool or browse a whole library before starting. Bring one piece of work you want to improve.

  1. Choose something concrete.

    Understand an unfamiliar project. Review a change. Make a content plan you can deliver. Pick a task small enough that you can inspect the result, and gather the source material it needs.

  2. Check your tools.

    Choose your assistant and client. Codex, Claude Code and local-model clients handle instructions differently. The relevant setup guidance helps you supply the workflow in the right place.

  3. Give your assistant a workflow.

    A skill supplies reusable guidance. Your task and source material make it relevant. Read the prerequisites, provide the context and ask the assistant to keep missing information visible.

  4. Check what changed.

    Open cited files. Compare the output with the source. Run suitable tests when the work needs them. Keep a note of what helped and what still needs attention.

Use the tools you already work with.

Bring your own assistant or local-model client and any access it requires. AI with Matt provides workflows and guidance; this website doesn’t run an AI model or change your IDE settings for you.

Codex project-skill discovery has bounded verification. Claude Code setup is documented. For local models, supplying instructions and automatic discovery depend on the client. Each workflow explains what it needs.