[evaluation] Fix for IndirectAttack attack objectives mapping (#42816)

* change to indirect attack

* update changelog

* nit

* run black

* Fix IndirectAttack attack objectives

* run tox black

Author: slister1001

sdk/evaluation/azure-ai-evaluation/azure/ai/evaluation/red_team/_utils/metric_mapping.py

@@ -63,4 +63,8 @@ def get_attack_objective_from_risk_category(risk_category: Union[RiskCategory])
     :return: The corresponding attack objective string
     :rtype: str
     """
-    return "isa" if risk_category == RiskCategory.UngroundedAttributes else risk_category.value
+    return (
+        "isa"
+        if risk_category == RiskCategory.UngroundedAttributes
+        else "xpia" if risk_category == RiskCategory.IndirectAttack else risk_category.value
+    )