For debugging the target language is irrelevant. As described in the readme the extension uses the parser + lexer interpreters from the (Typescript) runtime. All you need to set up is your Java installation (to generate the interpreter data) and a Visual Studio debugging task. The readme assumes you are familiar with creating such a configuration.
Once you opened your grammar go to the Debug section and from the top bar open the configuration (that wheel icon).
It will create a launch.json file for you if you haven one already, which is stored in the .vcode subfolder of the folder you opened in vscode.
Then use the „Add Configuration…“ button to create a new configuration. That will open a selection list for the debug type. Use „ANTLR: Parse“:
After that you will get a prefilled set of settings. Change the „input“ option to point to your input file (where the code is, you wanna parse). Now select that launch config in the drop down from the first screenshot. Switch to your grammar tab and click the green triangle. That should run your entire grammar and come up with a parse tree at the end (or an error if something was wrong).
Note: when you open a grammar in vscode (with my extension installed), it will automatically do an initial generation run in the background (unless you have that disabled, in which case debugging is not possible). So, after a few seconds you should be able to debug without any further step than the launch setup from above.