fix(backend/copilot): yield final StreamError after transient retry exhaustion for _HandledStreamError

When _run_stream_attempt raises a _HandledStreamError and all transient
retries are exhausted, the outer retry loop sets ended_with_stream_error
but stream_err remains None.  The post-loop code only emits a StreamError
when stream_err is not None, so the SSE stream closes silently and the
frontend never learns the request failed.

Yield a StreamError with the attempt's error message and code just before
breaking out of the retry loop, ensuring clients always receive an error
notification.

autogpt_platform/backend/backend/copilot/sdk/service.py

@@ -2274,6 +2274,14 @@ async def stream_chat_completion_sdk(
                     retryable=True,
                 )
                 ended_with_stream_error = True
+                # After exhausting all transient retries the stream is about
+                # to end.  Yield a final StreamError so the client knows the
+                # request failed — without this the SSE stream would close
+                # silently and the frontend would never show an error.
+                yield StreamError(
+                    errorText=exc.error_msg or FRIENDLY_TRANSIENT_MSG,
+                    code=exc.code or "transient_api_error",
+                )
                 break
             except Exception as e:
                 stream_err = e