fix(compact_20260112): carry prior compaction summary into Phase C summary call

When a request already contains a compaction block, Phase A slices
`effective_messages` to the turns since that block. Previously Phase C
passed the original `system` to the summary model, so multi-round
compaction silently dropped accumulated history each time the polyfill
fired. Pass `augmented_system` (original system + prior summary
prefix) so the summary model can produce a comprehensive summary that
incorporates both the prior round's context and the current slice.
`summarized_system` for the downstream call stays built from the
original `system` + new `summary_text`.

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

@@ -620,10 +620,12 @@ async def apply_compact_20260112(
             applied_edits=[applied],
         )
 
-    # Phase C: summarize.
+    # Phase C: summarize. ``augmented_system`` carries any prior compaction
+    # summary so multi-round compaction does not lose accumulated history —
+    # ``effective_messages`` only contains turns since the last compaction.
     prompt = _build_summary_prompt(edit_spec, tools)
     summary_messages = _build_summary_messages(
-        effective_messages, prompt, system=system
+        effective_messages, prompt, system=augmented_system
     )
     propagated_metadata = _propagate_metadata(metadata)
 

tests/test_litellm/llms/anthropic/experimental_pass_through/context_management/test_compact.py

@@ -896,6 +896,49 @@ async def test_system_prompt_forwarded_to_summary_call_as_content_blocks():
     assert "Initial task: review PR #123." in content
 
 
+async def test_summary_call_carries_prior_compaction_summary_into_system():
+    """Multi-round: when a prior compaction block is present, the summary
+    model receives the augmented system (with ``Previous conversation
+    summary: <prior>``) so it can produce a comprehensive summary that
+    incorporates both the prior round's context and the current slice.
+    Without this, multi-round compaction would silently drop accumulated
+    history each time the polyfill fires.
+    """
+    messages = _messages_with_compaction(summary="ROUND_ONE_SUMMARY_TEXT")
+    mock_response = _make_mock_response("<summary>Round two</summary>")
+
+    captured_calls: list = []
+
+    async def _fake_call_summary_model(**kwargs):
+        captured_calls.append(kwargs)
+        return mock_response
+
+    with (
+        patch(
+            "litellm.llms.anthropic.experimental_pass_through.context_management.editors.compact._read_summary_model_setting",
+            return_value="claude-haiku-4-5",
+        ),
+        patch("litellm.token_counter", return_value=200_000),
+        patch(
+            "litellm.llms.anthropic.experimental_pass_through.context_management.editors.compact._call_summary_model",
+            side_effect=_fake_call_summary_model,
+        ),
+    ):
+        await apply_compact_20260112(
+            model=MODEL,
+            messages=messages,
+            tools=None,
+            system="Original agent role.",
+            edit_spec=_EDIT_SPEC_DEFAULT,
+        )
+
+    summary_messages = captured_calls[0]["summary_messages"]
+    assert summary_messages[0]["role"] == "system"
+    system_content = summary_messages[0]["content"]
+    assert "ROUND_ONE_SUMMARY_TEXT" in system_content
+    assert "Original agent role." in system_content
+
+
 async def test_summary_call_omits_system_message_when_system_is_none():
     """No system message is prepended when the caller did not provide one."""
     messages = _simple_messages()