feat(v2): Web 界面（对话/协作过程/人工检验/指标）+ worker 优雅降级

router_system/worker.py

@@ -190,14 +190,19 @@ def _make_llama_generate(cfg: Dict[str, Any]) -> Callable[[str], Awaitable[str]]
     timeout_s = float(cfg.get("per_step_timeout_s", 300))
 
     async def _gen(prompt: str) -> str:
-        import httpx
-        async with httpx.AsyncClient(timeout=timeout_s) as client:
-            resp = await client.post(
-                f"{base_url}/chat/completions",
-                json={"model": model, "messages": [{"role": "user", "content": prompt}],
-                      "temperature": temperature, "max_tokens": 4096},
-            )
-            resp.raise_for_status()
-            return resp.json()["choices"][0]["message"]["content"]
+        try:
+            import httpx
+            async with httpx.AsyncClient(timeout=timeout_s) as client:
+                resp = await client.post(
+                    f"{base_url}/chat/completions",
+                    json={"model": model, "messages": [{"role": "user", "content": prompt}],
+                          "temperature": temperature, "max_tokens": 4096},
+                )
+                resp.raise_for_status()
+                return resp.json()["choices"][0]["message"]["content"]
+        except Exception as e:  # noqa: BLE001
+            # 连不上本地模型 -> 优雅降级（不抛 500），提示用户准备运行时
+            return ("（本地降级）无法连接本地 llama-server，未能生成该步骤内容。"
+                    f"请先运行 scripts/setup_runtime.py 并启动模型。错误：{type(e).__name__}")
 
     return _gen