Image turns already get a vision-model description injected into the prompt,
but the agent kept calling read_image to "confirm" the pictures. Those tool
results carry base64 image parts that Goose persists, so every later turn
against the text-only chat provider failed with `unknown variant image_url`
before the agent could write the page. WeChat page requests therefore fell
through to the fail-closed delivery message.
Drop read_image for the turn whenever a vision model handles the images, say
so explicitly in the injected prompt, and teach the poison scan to recognise
tool image parts so already-polluted sessions rotate instead of failing again.
Co-authored-by: Cursor <cursoragent@cursor.com>
Goose cannot persist historical image scrub (PUT 405), so DeepSeek rejects
image_url on the next turn. Reattach the recent report image and rotate to a
fresh session instead of treating the follow-up as a missing page source.
Co-authored-by: Cursor <cursoragent@cursor.com>
Strip image_url from persisted session payloads, keep image turns scoped to the active request, and align proxy/vision/page-data paths so WeChat image history does not leak into fresh sessions.
Co-authored-by: Cursor <cursoragent@cursor.com>
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>
john