feat: chat uploads, vision turn isolation, and MindSpace agent improvements

Add chat file/image upload UX, attachment proxying, vision thumbnails, and per-turn image scoping so agents only use the current upload. Extend MindSpace asset context, billing token state, OA/scenario verify scripts, and related runtime config.

Co-authored-by: Cursor <cursoragent@cursor.com>

server.mjs

@@ -76,6 +76,7 @@ import {
 import {
   buildMindSpacePublicRoutePath,
   buildMindSpacePublicUrlForUser,
+  isPublicWorkspaceHtmlRelativePath,
   resolveMindSpaceRuntimeConfig,
   resolveMindSpacePublishRoot,
   resolveMindSpaceServerRuntimeOptions,
@@ -5673,9 +5674,24 @@ async function resolvePublishedRoute(req, res, password = null) {
         referrer: req.get('referer'),
       },
     );
+    const embed = isPlazaEmbedRequest(req.query);
+    const raw = String(req.query.view ?? '').toLowerCase() === 'raw';
+    if (
+      !password &&
+      !embed &&
+      !raw &&
+      result.publication?.accessMode === 'public' &&
+      isPublicWorkspaceHtmlRelativePath(result.workspaceRelativePath)
+    ) {
+      const redirectPath = buildMindSpacePublicRoutePath(
+        result.ownerId,
+        String(result.workspaceRelativePath).split('/'),
+      );
+      return res.redirect(301, redirectPath);
+    }
     return await sendPublishedPage(req, res, result, {
-      embed: isPlazaEmbedRequest(req.query),
-      raw: String(req.query.view ?? '').toLowerCase() === 'raw',
+      embed,
+      raw,
       ownerSlug: req.params.ownerSlug,
     });
   } catch (error) {