fix: remove synthetic error message injection, fix session resume after repeated failures

Two changes to the error handler in the agent loop:

1. Remove the 'if not pending_handled' block that injected fake
   [System error during processing: ...] messages into conversation
   history.  These polluted history, burned tokens on retries, and
   could violate role alternation by injecting as role=user.
   The tool_calls error-result path (role=tool) is preserved.

2. Append the error final_response as an assistant message when
   hitting the iteration limit, so session resume doesn't produce
   consecutive user messages.

run_agent.py

@@ -6807,20 +6807,18 @@ class AIAgent:
                         pending_handled = True
                     break
                 
-                if not pending_handled:
+                # Non-tool errors don't need a synthetic message injected.
-                    # Error happened before tool processing (e.g. response parsing).
+                # The error is already printed to the user (line above), and
-                    # Choose role to avoid consecutive same-role messages.
+                # the retry loop continues.  Injecting a fake user/assistant
-                    last_role = messages[-1].get("role") if messages else None
+                # message pollutes history, burns tokens, and risks violating
-                    err_role = "assistant" if last_role == "user" else "user"
+                # role-alternation invariants.
-                    sys_err_msg = {
+
-                        "role": err_role,
-                        "content": f"[System error during processing: {error_msg}]",
-                    }
-                    messages.append(sys_err_msg)
-                
                 # If we're near the limit, break to avoid infinite loops
                 if api_call_count >= self.max_iterations - 1:
                     final_response = f"I apologize, but I encountered repeated errors: {error_msg}"
+                    # Append as assistant so the history stays valid for
+                    # session resume (avoids consecutive user messages).
+                    messages.append({"role": "assistant", "content": final_response})
                     break
         
         if final_response is None and (