Add handling for empty response (#808)

In some cases, the llm may fail to generate any response (maybe timeout
or connection issues), it will result in None response and that creates
an error when it is passed in for regular expression tag parsing. This
PR fixes that by ensuring response is not None before passing in to
process.

---------

Signed-off-by: Arthur Chan <[email protected]>

agent/prototyper.py

@@ -404,18 +404,20 @@ def _container_tool_reaction(self, cur_round: int, response: str,
                                build_result: BuildResult) -> Optional[Prompt]:
     """Validates LLM conclusion or executes its command."""
     prompt = prompt_builder.DefaultTemplateBuilder(self.llm, None).build([])
-    prompt = self._container_handle_bash_commands(response, self.inspect_tool,
-                                                  prompt)
 
-    # Then build fuzz target.
-    prompt = self._container_handle_conclusion(cur_round, response,
-                                               build_result, prompt)
-    if prompt is None:
-      # Succeeded.
-      return None
+    if response:
+      prompt = self._container_handle_bash_commands(response, self.inspect_tool,
+                                                    prompt)
+
+      # Then build fuzz target.
+      prompt = self._container_handle_conclusion(cur_round, response,
+                                                 build_result, prompt)
+      if prompt is None:
+        # Succeeded.
+        return None
 
     # Finally check invalid responses.
-    if not prompt.get():
+    if not response or not prompt.get():
       prompt = self._container_handle_invalid_tool_usage(
           self.inspect_tool, cur_round, response, prompt)