Attach Your Own Files When You Generate Tests

What the New File Attachments Add to Generate Mode

Generate mode now reads your files. Pass --files on the command line, or type @ inline to open a palette that groups matches into Files, Scenarios, and Test Cases. Kane CLI validates each attachment, uploads it, and maps it before submitting the request.

So instead of describing a flow from memory, you point Kane CLI at the document that already describes it. The generated cases follow your real scenarios, not a generic guess.

How to Attach Files in a Generation Session

Attach on the command line, or mention them inline while you refine.

kane-cli generate "cover the signup and checkout flows" --files ./spec.md ./wireframe.png

Inside the session, type @ to bring up the selector. It scrolls in a bounded window so it never takes over the screen. A mistyped path is surfaced with a warning instead of being silently dropped. The prompt locks while attachments process and shows a clear "Processing files" label, and Ctrl+C cancels an upload cleanly.

Note: Want to generate tests from your own specs? Start free. Try Kane CLI

Why Grounding Generation in Your Files Matters

A test is only as good as what it was based on. Generated from a vague objective, you get plausible cases. Generated from your spec and your data, you get cases that match how the product actually works.

From there the flow is the same. The session gives you two commands: /view inspects the generated scenarios and cases, and /save writes them to disk. View them as they form, drop any you do not want, then save the keepers to .testmd files you can run.

Note: Want the full generate command reference? Read the Kane CLI docs. Read the docs

Part of a Sharper Generate Experience

The same release tightened the rest of generate mode. It keeps its own input history, blocks accidental duplicate submits, gates each session behind a project and folder so context is always set, and writes generation errors into the scrollback instead of swallowing them.

Note: Want to see the code and examples? Open the repo. Open the repo

Your specs and designs already describe what the product should do. Now Kane CLI can read them directly, and turn them into tests you can run in a real browser.