fix: litellm_models.py to use correct endpoint from config and handle SSL for private IPs

- Modified litellm_models.py to read endpoint URL from environment variables or config file instead of hardcoded IP
- Added automatic SSL verification bypass for private/internal IPs to handle self-signed certificates
- Updated EndpointManager to properly pass environment variables to internal modules
- Fixed the script getting stuck by using the correct Litellm server endpoint (10.189.8.10 instead of 192.168.1.100)

Author: James Zhu

code_assistant_manager/endpoints.py

@@ -374,7 +374,7 @@ def _execute_model_command(
 
         # Try internal module execution
         if self._is_internal_module_command(tokens):
-            return self._execute_internal_module(tokens[2])
+            return self._execute_internal_module(tokens[2], env)
 
         # Try literal model list (command not found on PATH)
         if shutil.which(tokens[0]) is None:
@@ -391,14 +391,32 @@ def _is_internal_module_command(self, tokens: List[str]) -> bool:
             and tokens[2] in self.INTERNAL_MODEL_MODULES
         )
 
-    def _execute_internal_module(self, module_name: str) -> Optional[str]:
+    def _execute_internal_module(self, module_name: str, env: Dict[str, str] = None) -> Optional[str]:
         """Execute an internal Python module for model listing."""
         try:
             mod = importlib.import_module(module_name)
-            buf = io.StringIO()
-            with contextlib.redirect_stdout(buf):
-                mod.list_models()
-            return buf.getvalue().strip()
+
+            # Temporarily set environment variables if provided
+            old_env = {}
+            if env:
+                for key, value in env.items():
+                    if key in os.environ:
+                        old_env[key] = os.environ[key]
+                    os.environ[key] = value
+
+            try:
+                buf = io.StringIO()
+                with contextlib.redirect_stdout(buf):
+                    mod.list_models()
+                return buf.getvalue().strip()
+            finally:
+                # Restore original environment variables
+                for key, value in old_env.items():
+                    os.environ[key] = value
+                for key in env.keys():
+                    if key not in old_env and key in os.environ:
+                        del os.environ[key]
+
         except Exception as e:
             print(f"Warning: Failed to load module {module_name}: {e}")
             return None

code_assistant_manager/litellm_models.py

@@ -7,6 +7,7 @@
 import requests
 
 from .env_loader import load_env
+from .config import ConfigManager
 
 logger = logging.getLogger(__name__)
 
@@ -24,8 +25,30 @@ def fetch_litellm_models(api_key: str, base_url: str = "https://192.168.1.100:41
         "accept": "application/json",
         "x-litellm-api-key": api_key,
     }
-    # Security: Enable SSL certificate verification for secure connections
-    r = requests.get(url, params=params, headers=headers, timeout=30, verify=True)
+
+    # Determine if SSL verification should be enabled
+    # For internal/private IPs, skip verification as they often use self-signed certificates
+    import ipaddress
+    from urllib.parse import urlparse
+
+    verify_ssl = True
+    try:
+        parsed = urlparse(base_url)
+        if parsed.hostname:
+            try:
+                ip = ipaddress.ip_address(parsed.hostname)
+                # Skip verification for private IPs (10.0.0.0/8, 172.16.0.0/12, 192.168.0.0/16, 127.0.0.0/8)
+                if ip.is_private or ip.is_loopback:
+                    verify_ssl = False
+            except ValueError:
+                # Not an IP address, keep verification enabled
+                pass
+    except Exception:
+        # If parsing fails, keep verification enabled
+        pass
+
+    # Make the request
+    r = requests.get(url, params=params, headers=headers, timeout=30, verify=verify_ssl)
     r.raise_for_status()
     return r.json()
 
@@ -42,9 +65,21 @@ def list_models():
         logger.error("API_KEY_LITELLM environment variable is required but not found")
         raise SystemExit("API_KEY_LITELLM environment variable is required")
 
+    # Get base URL from environment (set by EndpointManager) or from config
+    base_url = os.environ.get("endpoint")
+    if not base_url:
+        # Try to get from config
+        try:
+            config = ConfigManager()
+            litellm_config = config.get_endpoint_config("litellm")
+            base_url = litellm_config.get("endpoint", "https://192.168.1.100:4142")
+        except Exception as e:
+            logger.debug(f"Could not load config, using default URL: {e}")
+            base_url = "https://192.168.1.100:4142"
+
     logger.debug("Fetching Litellm models")
     try:
-        models_data = fetch_litellm_models(api_key)
+        models_data = fetch_litellm_models(api_key, base_url)
         model_count = len(models_data.get("data", []))
         logger.debug(f"Found {model_count} models")