diff --git a/strix/interface/main.py b/strix/interface/main.py
index 9ec17e5..c6ebe2b 100644
--- a/strix/interface/main.py
+++ b/strix/interface/main.py
@@ -212,7 +212,7 @@ async def warm_up_llm() -> None:
             {"role": "user", "content": "Reply with just 'OK'."},
         ]
 
-        llm_timeout = int(Config.get("llm_timeout"))  # type: ignore[arg-type]
+        llm_timeout = int(Config.get("llm_timeout") or "300")
 
         completion_kwargs: dict[str, Any] = {
             "model": model_name,
diff --git a/strix/llm/config.py b/strix/llm/config.py
index 609b628..3426327 100644
--- a/strix/llm/config.py
+++ b/strix/llm/config.py
@@ -18,6 +18,6 @@ class LLMConfig:
         self.enable_prompt_caching = enable_prompt_caching
         self.skills = skills or []
 
-        self.timeout = timeout or int(Config.get("llm_timeout"))  # type: ignore[arg-type]
+        self.timeout = timeout or int(Config.get("llm_timeout") or "300")
 
         self.scan_mode = scan_mode if scan_mode in ["quick", "standard", "deep"] else "deep"
diff --git a/strix/llm/request_queue.py b/strix/llm/request_queue.py
index c3ddc37..de249e2 100644
--- a/strix/llm/request_queue.py
+++ b/strix/llm/request_queue.py
@@ -12,8 +12,8 @@ from strix.config import Config
 
 class LLMRequestQueue:
     def __init__(self, max_concurrent: int = 1, delay_between_requests: float = 4.0):
-        delay_between_requests = float(Config.get("llm_rate_limit_delay"))  # type: ignore[arg-type]
-        max_concurrent = int(Config.get("llm_rate_limit_concurrent"))  # type: ignore[arg-type]
+        delay_between_requests = float(Config.get("llm_rate_limit_delay") or "4.0")
+        max_concurrent = int(Config.get("llm_rate_limit_concurrent") or "1")
 
         self.max_concurrent = max_concurrent
         self.delay_between_requests = delay_between_requests
diff --git a/strix/runtime/docker_runtime.py b/strix/runtime/docker_runtime.py
index 4719833..a870160 100644
--- a/strix/runtime/docker_runtime.py
+++ b/strix/runtime/docker_runtime.py
@@ -122,7 +122,7 @@ class DockerRuntime(AbstractRuntime):
 
         for attempt in range(max_retries):
             try:
-                self._verify_image_available(Config.get("strix_image"))  # type: ignore[arg-type]
+                self._verify_image_available(Config.get("strix_image") or "")
 
                 try:
                     existing_container = self.client.containers.get(container_name)
diff --git a/strix/tools/executor.py b/strix/tools/executor.py
index 1ec0375..9dbc74e 100644
--- a/strix/tools/executor.py
+++ b/strix/tools/executor.py
@@ -19,8 +19,8 @@ from .registry import (
 )
 
 
-SANDBOX_EXECUTION_TIMEOUT = float(Config.get("strix_sandbox_execution_timeout"))  # type: ignore[arg-type]
-SANDBOX_CONNECT_TIMEOUT = float(Config.get("strix_sandbox_connect_timeout"))  # type: ignore[arg-type]
+SANDBOX_EXECUTION_TIMEOUT = float(Config.get("strix_sandbox_execution_timeout") or "500")
+SANDBOX_CONNECT_TIMEOUT = float(Config.get("strix_sandbox_connect_timeout") or "10")
 
 
 async def execute_tool(tool_name: str, agent_state: Any | None = None, **kwargs: Any) -> Any: