feat(add Inspect AI):

agentic_security/probe_data/modules/inspect_ai_task.py

@@ -0,0 +1,13 @@
+from inspect_ai import Task, eval, task
+from inspect_ai.dataset import example_dataset
+from inspect_ai.scorer import model_graded_fact
+from inspect_ai.solver import chain_of_thought, generate, self_critique
+
+
+@task
+def theory_of_mind():
+    return Task(
+        dataset=example_dataset("theory_of_mind"),
+        plan=[chain_of_thought(), generate(), self_critique()],
+        scorer=model_graded_fact(),
+    )