code exec

Summary:

Test Plan:

tests/integration/agents/test_agents.py

@@ -187,7 +187,7 @@ def test_builtin_tool_web_search(llama_stack_client_with_mocked_inference, agent
         messages=[
             {
                 "role": "user",
-                "content": "Search the web and tell me who the current CEO of Meta is.",
+                "content": "Search the web and tell me who the founder of Meta is.",
             }
         ],
         session_id=session_id,

tests/integration/conftest.py

@@ -6,12 +6,17 @@
 import inspect
 import itertools
 import os
+import platform
 import textwrap
 
 from dotenv import load_dotenv
 
+from llama_stack.log import get_logger
+
 from .report import Report
 
+logger = get_logger(__name__, category="tests")
+
 
 def pytest_configure(config):
     config.option.tbstyle = "short"
@@ -24,6 +29,10 @@ def pytest_configure(config):
         key, value = env_var.split("=", 1)
         os.environ[key] = value
 
+    if platform.system() == "Darwin":  # Darwin is the system name for macOS
+        os.environ["DISABLE_CODE_SANDBOX"] = "1"
+        logger.info("Setting DISABLE_CODE_SANDBOX=1 for macOS")
+
     if config.getoption("--report"):
         config.pluginmanager.register(Report(config))