[Fix] Fast mode loses large Brain query results

[roomote-roomote]

​Opened on behalf of @daniel-lxs. Follow up by mentioning @roomote-roomote or in the web UI.

What changed

• Default Fast-mode Brain query calls to detail: low, adaptive_return: true, and limit: 6 unless the caller explicitly provides those arguments.
• Pass gbrain query responses through unchanged, without field whitelisting, DTO mapping, or a Roomote-specific output byte cap.
• Cover the compact defaults, explicit overrides, and native result pass-through behavior.

Why this change was made

gbrain already exposes parameters designed to control query response size. Fast mode should use those controls rather than maintaining a separate Brain-specific payload format in Roomote.

Impact

Fast Brain queries request a compact result set by default while preserving gbrain's native response contract and caller control over query detail, adaptive return, and result count.

[roomote-community]

No code issues found. See task

_{Reviewed 84dd96b}