Fix writing assist: disable thinking mode, drop stuck-buffer 409

- stream_chat: add think=False parameter passed through to Ollama payload.
  qwen3 models have thinking enabled by default; without this flag the model
  spends minutes generating internal thinking tokens that stream_chat silently
  discards, leaving the frontend spinner blank until the SSE connection times
  out and the widget disappears.

- create_assist_buffer: orphan (overwrite) a still-running buffer instead of
  raising. The old asyncio task holds a direct reference and completes
  harmlessly against the stale buffer. New requests always win.

- assist_route: remove the 409 guard that blocked new requests when a previous
  generation got stuck. create_assist_buffer now handles this transparently.

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

src/fabledassistant/services/llm.py

@@ -108,12 +108,18 @@ async def stream_chat(
     messages: list[dict],
     model: str,
     options: dict | None = None,
+    think: bool = False,
 ) -> AsyncGenerator[str, None]:
-    """Stream chat completion from Ollama, yielding content chunks."""
+    """Stream chat completion from Ollama, yielding content chunks.
+
+    Set think=False (default) to disable chain-of-thought on qwen3+ models.
+    Thinking tokens are silently discarded anyway, but disabling avoids the
+    multi-minute delay before the first content token arrives.
+    """
     merged_options = {"num_ctx": Config.OLLAMA_NUM_CTX}
     if options:
         merged_options.update(options)
-    payload: dict = {"model": model, "messages": messages, "stream": True, "options": merged_options}
+    payload: dict = {"model": model, "messages": messages, "stream": True, "options": merged_options, "think": think}
     # read=None: no per-chunk timeout — Ollama may pause for any duration while
     # processing a large input context before the first token arrives.
     async with httpx.AsyncClient(timeout=httpx.Timeout(connect=30.0, read=None, write=None, pool=30.0)) as client: