fix(extension): distinguish unreadable pages from missing access

Author: waleedlatif1

apps/extension/src/capture.ts

@@ -4,7 +4,7 @@ import {
   pageSnapshotSchema,
 } from '@sim/browser-extension-protocol'
 import { generateId } from '@sim/utils/id'
-import { SIM_ORIGIN } from '@/config'
+import { PAGE_ACCESS_ORIGINS, SIM_ORIGIN } from '@/config'
 import { extractPageText } from '@/extract'
 import { digestUrl } from '@/utils'
 
@@ -51,7 +51,12 @@ export async function captureActivePage(
   if (expected && expected.tabId !== tab.id) {
     throw new PageCaptureError('The page changed before reading. Attach it again.')
   }
-  if (!tab.url) throw new SiteAccessRequiredError(tab.id, tab.windowId)
+  if (!tab.url) {
+    const allowed = await chrome.permissions.contains({ origins: PAGE_ACCESS_ORIGINS })
+    signal?.throwIfAborted()
+    if (allowed) throw new PageCaptureError('This page cannot be attached.')
+    throw new SiteAccessRequiredError(tab.id, tab.windowId)
+  }
   let changed = false
   const activated = (active: chrome.tabs.OnActivatedInfo) => {
     if (active.windowId === tab.windowId && active.tabId !== tab.id) changed = true

apps/extension/src/config.ts

@@ -1,5 +1,6 @@
 declare const __SIM_ORIGIN__: string
 export const SIM_ORIGIN = __SIM_ORIGIN__
+export const PAGE_ACCESS_ORIGINS = ['https://*/*', 'http://*/*']
 export const REFERENCE_PAGE_KEY = 'referenceCurrentPage'
 export const OPEN_SIM_COMMAND = 'open-sim'
 export const CAPTURE_INTENT_KEY = 'captureIntent'

apps/extension/src/site-access.tsx

@@ -1,7 +1,6 @@
 import { useEffect, useRef, useState } from 'react'
 import { Chip, ChipModalBody, ChipModalHeader, ChipModalSurface } from '@sim/emcn'
-
-const PAGE_ACCESS_ORIGINS = ['https://*/*', 'http://*/*']
+import { PAGE_ACCESS_ORIGINS } from '@/config'
 
 interface PageAccessButtonProps {
   allowed?: boolean

apps/sim/app/(landing)/privacy/privacy-content.tsx

@@ -842,7 +842,7 @@ export const PRIVACY_CONFIG: LegalPageConfig = {
         {
           kind: 'paragraph',
           content: richText(
-            'Browser shortcuts can optionally recover an explicit sim/ shortcut that Chrome sends to a supported search engine. After You enable a provider and approve access to its listed sites, Chrome applies request rules that send only the matching shortcut name and path values to Sim’s normal link resolver. Other search parameters and ordinary search terms are not sent to Sim by this feature. The extension stores Your provider choices locally and does not change Your default search engine. These site permissions can also allow page attachment when You explicitly request it; enabling shortcut recovery does not enable Reference current page or background capture. Turning recovery off stops these redirects. Site access remains until You revoke it in the extension or Chrome settings.'
+            'Browser shortcuts can optionally recover an explicit sim/ shortcut that Chrome sends to a supported search engine. After You enable a provider and approve access to its listed sites, Chrome applies request rules that send only the matching shortcut name and path values to Sim’s normal link resolver. Other search parameters and ordinary search terms are not sent to Sim by this feature. The extension stores Your provider choices locally and does not change Your default search engine. These site permissions can also allow page attachment when You explicitly request it; enabling shortcut recovery does not enable Reference current page or background capture. Turning recovery off stops these redirects. Site access remains until You revoke or restrict it in Chrome settings.'
           ),
         },
         {