Hey everyone,
Since Cursor Composer already allows us to run Bash commands and inserts their output directly into the context window, I’d love to see this extended to support images. Right now, if we run a command that generates an image (e.g., a screenshot from Puppeteer or Playwright), there doesn’t seem to be a way to pass that output into Cursor’s multimodal input channels.
Would it be possible to handle image output in some way—either by detecting when a command returns an image file and embedding it, or by supporting base64-encoded images and automatically rendering them in the chat? This would be a huge improvement for workflows involving automated screenshots, visual debugging, and other multimodal tasks. Think, ‘Check if our changes went through by running ‘screenshot localhost:/homepage’ to check the output’
Let me know if there’s anything missing on my end to make this work!
Thanks!