Add 120s timeout to image analysis streaming

The vision model request was hanging indefinitely, freezing the bot.
The streaming loop had no timeout so if the model never returned
chunks, the bot would wait forever. Now times out after 2 minutes.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

utils/llm_client.py

@@ -373,31 +373,39 @@ class LLMClient:
 
         async with self._semaphore:
             try:
-                stream = await self._client.chat.completions.create(
-                    model=self.model,
-                    messages=[
-                        {"role": "system", "content": system_prompt},
-                        {"role": "user", "content": user_content},
-                    ],
-                    temperature=0.8,
-                    max_tokens=2048,
-                    stream=True,
-                )
+                async def _stream_image():
+                    stream = await self._client.chat.completions.create(
+                        model=self.model,
+                        messages=[
+                            {"role": "system", "content": system_prompt},
+                            {"role": "user", "content": user_content},
+                        ],
+                        temperature=0.8,
+                        max_tokens=2048,
+                        stream=True,
+                    )
 
-                chunks: list[str] = []
-                notified = False
-                async for chunk in stream:
-                    delta = chunk.choices[0].delta if chunk.choices else None
-                    if delta and delta.content:
-                        if not notified and on_first_token:
-                            await on_first_token()
-                            notified = True
-                        chunks.append(delta.content)
+                    chunks: list[str] = []
+                    notified = False
+                    async for chunk in stream:
+                        delta = chunk.choices[0].delta if chunk.choices else None
+                        if delta and delta.content:
+                            if not notified and on_first_token:
+                                await on_first_token()
+                                notified = True
+                            chunks.append(delta.content)
 
-                content = "".join(chunks).strip()
+                    return "".join(chunks).strip()
+
+                content = await asyncio.wait_for(_stream_image(), timeout=120)
                 elapsed = int((time.monotonic() - t0) * 1000)
                 self._log_llm("image", elapsed, bool(content), req_json, content or None)
                 return content if content else None
+            except asyncio.TimeoutError:
+                elapsed = int((time.monotonic() - t0) * 1000)
+                logger.error("LLM image analysis timed out after %ds", elapsed // 1000)
+                self._log_llm("image", elapsed, False, req_json, error="Timeout")
+                return None
             except Exception as e:
                 elapsed = int((time.monotonic() - t0) * 1000)
                 logger.error("LLM image analysis error: %s", e)