Problem
Today we can add selected workspace file text to the prompt as a snippet, but there is no comparable one-click way to capture the current preview/browser page content and attach it to the chat composer.
That makes UI/debugging workflows slower because users have to manually copy/paste page text or describe what they are seeing instead of attaching the actual rendered page content as context.
Proposal
Add a composer action that fetches the current page content and attaches it as a snippet/text attachment in the chat input flow.
Suggested behavior:
- Read the active page's URL, title, and visible text/content
- Package that content into the existing snippet/text-attachment pipeline
- Show it in the composer the same way other attached snippets are surfaced
- Include lightweight source metadata (for example URL/title) in the snippet body or filename
- Apply sensible truncation/size limits so very large pages do not overwhelm the prompt
Acceptance criteria
- Users can attach the current page content with one action from the preview/browser flow
- The captured content is added to the chat composer as a snippet/text attachment
- The attachment includes enough source metadata to identify the page
- Large pages are trimmed to a safe size instead of failing silently
- The flow works without requiring manual text selection or copy/paste
Notes
This seems complementary to the existing snippet flow for selected workspace-file text: same destination in the composer, but the source is the active rendered page instead of the code editor/file viewer
Activity
0No activity yet.