fix(lite): keep pi's system role on reasoning model groups

pi switches its system prompt to the developer role for any model marked reasoning, so an operator's
cache_control_injection_points entry for role system stopped matching on Claude groups. lite pi now sets
compat.supportsDeveloperRole false on those entries, so pi sends system as it did before.

litellm/proxy/client/cli/commands/pi.py

@@ -193,7 +193,13 @@ def _reasoning_fields(limit: ModelLimits | None) -> Mapping[str, JsonValue]:
     efforts: Final = limit.supported_reasoning_efforts if limit else None
     if not efforts:
         return MappingProxyType({})
-    return MappingProxyType({"reasoning": True, "thinkingLevelMap": _thinking_level_map(efforts)})
+    return MappingProxyType(
+        {
+            "reasoning": True,
+            "thinkingLevelMap": _thinking_level_map(efforts),
+            "compat": {"supportsDeveloperRole": False},
+        }
+    )
 
 
 def _model_entry(

tests/unit/proxy/client/cli/test_pi.py

@@ -257,7 +257,7 @@ class TestProviderBlock:
             ),
         ],
     )
-    def test_reasoning_efforts_become_pi_thinking_levels(self, efforts, thinking_level_map):
+    def test_reasoning_efforts_become_pi_thinking_levels_and_keep_the_system_role(self, efforts, thinking_level_map):
         block = provider_block(
             "http://localhost:4000",
             ("m-1",),
@@ -270,6 +270,7 @@ class TestProviderBlock:
                 "maxTokens": 8192,
                 "reasoning": True,
                 "thinkingLevelMap": thinking_level_map,
+                "compat": {"supportsDeveloperRole": False},
             }
         ]