fix(context_management): count the system prompt through compact's shared counter

Finishes the main merge's conflict in clear_tool_uses.py. The input_tokens trigger now adds the system prompt through count_system_tokens, the same counter compact uses, while the count reused for cleared_input_tokens stays messages and tools only, so the reported savings are just the cleared tool results

litellm/llms/anthropic/messages/context_management/editors/clear_tool_uses.py

@@ -31,7 +31,7 @@ from ..constants import (
 )
 from ..errors import AnthropicContextManagementError
 from ..placeholders import build_cleared_tool_result_content
-from .compact import system_to_text
+from .compact import count_system_tokens
 
 if TYPE_CHECKING:
     from litellm.types.llms.openai import ChatCompletionToolParam
@@ -199,14 +199,6 @@ def _select_in_batches(candidates: Sequence[str], tokens: Mapping[str, int], cle
     return tuple(candidates[:batch_end])
 
 
-def _with_system(
-    messages: Sequence[Mapping[str, object]], system: str | Sequence[Mapping[str, object]] | None
-) -> Sequence[Mapping[str, object]]:
-    """The messages to count: like the API, an ``input_tokens`` trigger counts the system prompt too."""
-    system_text: Final = system_to_text(system)
-    return ({"role": "system", "content": system_text}, *messages) if system_text else messages
-
-
 def _trigger_met(
     trigger: dict[str, object],
     model: str,
@@ -215,7 +207,7 @@ def _trigger_met(
     system: str | Sequence[Mapping[str, object]] | None,
     counted_tool_uses: int,
 ) -> tuple[bool, int | None]:
-    """Return (trigger_met, input_tokens if counted for reuse)."""
+    """Return (trigger_met, message and tool tokens for reuse); like the API, input_tokens adds the system prompt."""
     trigger_type: Final = trigger.get("type", "input_tokens")
     threshold = trigger.get("value")
 
@@ -226,14 +218,15 @@ def _trigger_met(
 
     if not isinstance(threshold, int):
         threshold = DEFAULT_INPUT_TOKENS_TRIGGER
-    current_tokens: Final = litellm.token_counter(
+    message_tokens: Final = litellm.token_counter(
         model=model,
         messages=messages,
         tools=cast("list[ChatCompletionToolParam] | None", tools),
     )
+    current_tokens: Final = message_tokens + count_system_tokens(model, system)
     verbose_logger.debug("context_management polyfill: current_tokens: %s", current_tokens)
     verbose_logger.debug("context_management polyfill: threshold: %s", threshold)
-    return current_tokens > threshold, current_tokens
+    return current_tokens > threshold, message_tokens
 
 
 def _resolve_keep_count(keep: dict[str, object]) -> int:

litellm/llms/anthropic/messages/context_management/editors/compact.py

@@ -320,13 +320,15 @@ def _count_effective_tokens(
         content: Final = compaction_block.get("content") or ""
         if content:
             total += litellm.token_counter(model=model, text=content)
-    system_text: Final = system_to_text(system)
-    if system_text:
-        total += litellm.token_counter(model=model, text=system_text)
-    return total
+    return total + count_system_tokens(model, system)
 
 
-def system_to_text(
+def count_system_tokens(model: str, system: str | Sequence[Mapping[str, object]] | None) -> int:
+    system_text: Final = _system_to_text(system)
+    return litellm.token_counter(model=model, text=system_text) if system_text else 0
+
+
+def _system_to_text(
     system: str | Sequence[Mapping[str, object]] | None,
 ) -> str:
     """Flatten an Anthropic-style ``system`` value into a single string for

tests/unit/llms/anthropic/messages/context_management/test_clear_tool_uses.py

@@ -709,10 +709,18 @@ def test_input_tokens_trigger_counts_the_system_prompt(system: str | list[dict[s
     _, with_system = apply_clear_tool_uses_20250919(
         model=MODEL, messages=deepcopy(messages), tools=None, system=system, edit_spec=edit_spec
     )
+    _, met_without_system = apply_clear_tool_uses_20250919(
+        model=MODEL,
+        messages=deepcopy(messages),
+        tools=None,
+        system=None,
+        edit_spec={**edit_spec, "trigger": {"type": "input_tokens", "value": 0}},
+    )
     assert without_system is None
     assert with_system is not None
+    assert met_without_system is not None
     assert with_system["cleared_tool_uses"] == 2
-    assert with_system["cleared_input_tokens"] < history_tokens
+    assert with_system["cleared_input_tokens"] == met_without_system["cleared_input_tokens"] > 0
 
 
 @pytest.mark.parametrize(("clear_tool_inputs", "triggers"), [(False, False), (True, True)])