Introduce `outputSchema` (was: Should output also have a schema?)

[bokand]

The explainer includes an example returning some structured output:

return {
        content: [
            {
                type: "text",
                text: `Stamp "${name}" added successfully! The collection now contains ${stamps.length} stamps.`, 
            },
        ]
    };

Should we include an outputSchema as MCP recently added?

I'm not sure the schema itself helps (the return is already a JavaScript object...) but descriptions might help the agent understand what the output means?

[bwalderman]

The motivation for outputSchema in MCP is discussed here modelcontextprotocol/modelcontextprotocol#356

Seems like if it makes sense for classic MCP, it would make sense for WebMCP too.

[khushalsagar]

This will need to be supported for the declarative API as well. The case where JS preventDefaults the submit event, the result is similar to the imperative tools case. So it could just be an attribute on the form element.

The cross-document case where we're consider ld-json for the result needs more thought. @domfarolino FYI.

[anssiko]

RESOLUTION: Add outputSchema to the imperative API and investigate how to associate outputSchema with declarative WebMCP (issue #9)