#34777 Cursor bridge endpoint (/cursor/chat/completions) doesn't translate image_url content blocks to Responses API format
## Bug
`/cursor/chat/completions` fails with a Responses API schema validation error whenever the request contains an `image_url` content block (i.e. any vision/multimodal request), for every model/provider tested. The identical request succeeds via `/v1/chat/completions`.
## Repro
POST a standard Chat Completions-format request to `/cursor/chat/completions` with a message containing an `image_url` content block, e.g.:
```json POST /cursor/chat/completions { "model": "<any vision-capable model, tested via OpenRouter>", "messages": [ { "role": "user", "content": [ {"type": "text", "text": "What's in this image?"}, {"type": "image_url", "image_url": {"url": "data:image/png;base64,..."}} ] } ] } ```
Result: fails with a Responses API schema validation error (`invalid_prompt`, discriminator mismatch — Responses API expects `input_text` / `input_image` / `encrypted_content`, not `text` / `image_url`).
The **same request** to `/v1/chat/completions` with the identical model/image succeeds and correctly identifies the image content — proving the bug is isolated to the `/cursor` bridge's content-block translation, not the underlying prov…