## Summary Add OpenAI-compatible response_format: json_object as opt-in via .cora.yaml config to reduce JSON parsing errors. ## Accept Criteria - [ ] Add review.response_format field to config schema (values: json_object, none) - [ ] Pass response_format to LLM API when enabled - [ ] Graceful fallback if provider returns error with response_format - [ ] Default behavior unchanged (backward compatible) - [ ] Unit test: ChatRequest serialization with response_format ## Scope Config schema + LLM request building + fallback handling
Summary
Add OpenAI-compatible response_format: json_object as opt-in via .cora.yaml config to reduce JSON parsing errors.
Accept Criteria
Scope
Config schema + LLM request building + fallback handling