From 7da6fe54b5201dc42c2c55baac943fc46ec5e097 Mon Sep 17 00:00:00 2001 From: tin-berri Date: Mon, 7 Sep 2026 18:03:43 -0700 Subject: [PATCH] fix: skip one-shot Claude Code cache injection (#40175) --- .../anthropic_cache_control_hook.py | 82 +++++- litellm/llms/anthropic/common_utils.py | 90 +++++++ litellm/proxy/litellm_pre_call_utils.py | 8 +- .../prompt_caching_deployment_check.py | 1 + .../test_anthropic_cache_control_hook.py | 237 ++++++++++++++++++ .../anthropic/test_anthropic_common_utils.py | 24 ++ .../test_prompt_caching_deployment_check.py | 61 +++++ 7 files changed, 492 insertions(+), 11 deletions(-) diff --git a/litellm/integrations/anthropic_cache_control_hook.py b/litellm/integrations/anthropic_cache_control_hook.py index 3519240dda9..4f9b18713d0 100644 --- a/litellm/integrations/anthropic_cache_control_hook.py +++ b/litellm/integrations/anthropic_cache_control_hook.py @@ -16,6 +16,8 @@ from collections.abc import Iterable, Mapping, Sequence from typing import TYPE_CHECKING, Any, Final, cast from urllib.parse import urlparse +from pydantic import TypeAdapter, ValidationError + from litellm._logging import verbose_logger from litellm.integrations.custom_logger import CustomLogger from litellm.integrations.custom_prompt_management import CustomPromptManagement @@ -23,6 +25,7 @@ from litellm.integrations.prompt_management_base import PromptManagementClient from litellm.litellm_core_utils.prompt_templates.common_utils import ( with_prompt_cache_breakpoint, ) +from litellm.llms.anthropic.common_utils import is_claude_code_one_shot_subagent_request from litellm.types.integrations.anthropic_cache_control_hook import ( GATEWAY_INJECTED_CACHE_METADATA_KEY, GATEWAY_INJECTED_FOR_EVERY_DEPLOYMENT, @@ -62,10 +65,26 @@ OPENAI_PROMPT_CACHE_BREAKPOINT_BLOCK_TYPES: Final = frozenset( ) OPENAI_API_HOST: Final = "api.openai.com" OPENAI_API_BASE_ENV_VARS: Final = ("OPENAI_BASE_URL", "OPENAI_API_BASE") +_OBJECT_MAPPING_ADAPTER: Final = TypeAdapter(dict[object, object]) +_OBJECT_LIST_ADAPTER: Final = TypeAdapter(list[object]) AllToolParamValues = ChatCompletionToolParam | AllAnthropicToolsValues +def _validated_object_mapping(value: object) -> dict[object, object] | None: + try: + return _OBJECT_MAPPING_ADAPTER.validate_python(value) + except ValidationError: + return None + + +def _validated_object_list(value: object) -> list[object] | None: + try: + return _OBJECT_LIST_ADAPTER.validate_python(value) + except ValidationError: + return None + + def supports_openai_prompt_cache_breakpoint(model: str) -> bool: model_map_flag: Final = _model_map_prompt_cache_breakpoint_flag(model) if model_map_flag is not None: @@ -114,6 +133,36 @@ CARRY_UNMATCHED_MESSAGE_POINTS: Final = "_litellm_carry_unmatched_cache_control_ class AnthropicCacheControlHook(CustomPromptManagement): + @staticmethod + def _request_value(request_kwargs: object, key: str) -> object: + request_mapping: Final = _validated_object_mapping(request_kwargs) + if request_mapping is None: + return None + return request_mapping.get(key) + + @staticmethod + def _request_user_agent(request_kwargs: object) -> str | None: + proxy_server_request: Final = AnthropicCacheControlHook._request_value(request_kwargs, "proxy_server_request") + proxy_server_request_mapping: Final = _validated_object_mapping(proxy_server_request) + if proxy_server_request_mapping is None: + return None + headers: Final = proxy_server_request_mapping.get("headers") + headers_mapping: Final = _validated_object_mapping(headers) + if headers_mapping is None: + return None + user_agent: Final = next( + (value for key, value in headers_mapping.items() if isinstance(key, str) and key.lower() == "user-agent"), + None, + ) + return user_agent if isinstance(user_agent, str) else None + + @staticmethod + def _request_system(request_kwargs: object) -> str | list[object] | None: + system: Final = AnthropicCacheControlHook._request_value(request_kwargs, "system") + if isinstance(system, str): + return system + return _validated_object_list(system) + def get_chat_completion_prompt( self, model: str, @@ -520,12 +569,13 @@ class AnthropicCacheControlHook(CustomPromptManagement): points: Sequence[CacheControlInjectionPoint], messages: list[AllMessageValues], tools: list[object] | None, + cache_control: object, model: str, custom_llm_provider: str | None, api_base: object, prompt_cache_options: object, ) -> Sequence[Mapping[str, object]] | None: - if AnthropicCacheControlHook._should_stand_down(points, messages, None, tools): + if AnthropicCacheControlHook._should_stand_down(points, messages, None, tools, cache_control): return None return AnthropicCacheControlHook._stamped_with_dialect( points, model, custom_llm_provider, api_base, prompt_cache_options @@ -561,6 +611,7 @@ class AnthropicCacheControlHook(CustomPromptManagement): messages: list[AllMessageValues], system: str | list | None, tools: list | None, + cache_control: object = None, ) -> bool: """Whether configured injection points must yield to client-set cache_control. @@ -573,13 +624,14 @@ class AnthropicCacheControlHook(CustomPromptManagement): """ if all(point.get("_litellm_judged") for point in points): return False - return AnthropicCacheControlHook._request_has_cache_control(messages, system, tools) + return AnthropicCacheControlHook._request_has_cache_control(messages, system, tools, cache_control) @staticmethod def _request_has_cache_control( messages: list[AllMessageValues], system: str | list | None, tools: list | None = None, + cache_control: object = None, ) -> bool: """Return True if the request already carries any client-supplied cache_control. @@ -591,6 +643,8 @@ class AnthropicCacheControlHook(CustomPromptManagement): carry the mark either at the top level (Anthropic shape) or nested under ``function`` (OpenAI shape); the Anthropic chat transform accepts both. """ + if cache_control is not None: + return True if AnthropicCacheControlHook.count_request_cache_breakpoints(messages, system) > 0: return True if tools is not None: @@ -612,6 +666,8 @@ class AnthropicCacheControlHook(CustomPromptManagement): custom_llm_provider: str | None, tools: list | None = None, enable_prompt_caching: bool | None = None, + cache_control: object = None, + request_kwargs: object = None, ) -> list[CacheControlInjectionPoint]: """Default breakpoints when ``litellm.enable_anthropic_prompt_caching`` is on. @@ -649,7 +705,12 @@ class AnthropicCacheControlHook(CustomPromptManagement): if not supports_prompt_caching(model=model, custom_llm_provider=provider): return [] - if AnthropicCacheControlHook._request_has_cache_control(messages, system, tools): + if AnthropicCacheControlHook._request_has_cache_control(messages, system, tools, cache_control): + return [] + + if is_claude_code_one_shot_subagent_request( + messages, system, tools, AnthropicCacheControlHook._request_user_agent(request_kwargs) + ): return [] control: Final = AnthropicCacheControlHook._default_control() @@ -665,6 +726,7 @@ class AnthropicCacheControlHook(CustomPromptManagement): models: Iterable[str], tools: list[AllToolParamValues] | None = None, enable_prompt_caching: bool | None = None, + request_kwargs: object = None, ) -> list[AllMessageValues]: """Return the messages auto prompt caching will send, default breakpoints included. @@ -681,11 +743,13 @@ class AnthropicCacheControlHook(CustomPromptManagement): for candidate in ( AnthropicCacheControlHook.get_default_injection_points( messages=messages, - system=None, model=model, custom_llm_provider=None, tools=tools, enable_prompt_caching=enable_prompt_caching, + system=AnthropicCacheControlHook._request_system(request_kwargs), + cache_control=AnthropicCacheControlHook._request_value(request_kwargs, "cache_control"), + request_kwargs=request_kwargs, ) for model in models ) @@ -730,6 +794,7 @@ class AnthropicCacheControlHook(CustomPromptManagement): non_default_params["cache_control_injection_points"], messages, tools, + non_default_params.get("cache_control"), model, custom_llm_provider, api_base, @@ -747,6 +812,8 @@ class AnthropicCacheControlHook(CustomPromptManagement): custom_llm_provider=custom_llm_provider, tools=tools, enable_prompt_caching=enable_prompt_caching, + cache_control=non_default_params.get("cache_control"), + request_kwargs=non_default_params, ) if points: non_default_params["cache_control_injection_points"] = points @@ -853,10 +920,13 @@ class AnthropicCacheControlHook(CustomPromptManagement): enable_prompt_caching: Final = cast( # cast-ok: kwargs is untyped; key stamped as bool by the proxy bool | None, kwargs.pop("enable_prompt_caching", None) ) + cache_control: Final = kwargs.get("cache_control") configured: Final = cast( # cast-ok: kwargs is untyped; this key only holds the documented injection-point list list[CacheControlInjectionPoint] | None, kwargs.pop("cache_control_injection_points", None) ) - if configured and AnthropicCacheControlHook._should_stand_down(configured, typed_messages, system, tools): + if configured and AnthropicCacheControlHook._should_stand_down( + configured, typed_messages, system, tools, cache_control + ): return messages, system injection_points: list[CacheControlInjectionPoint] = configured or [] if not injection_points and model is not None: @@ -867,6 +937,8 @@ class AnthropicCacheControlHook(CustomPromptManagement): model=model, custom_llm_provider=custom_llm_provider, enable_prompt_caching=enable_prompt_caching, + cache_control=cache_control, + request_kwargs=kwargs, ) if not injection_points: return messages, system diff --git a/litellm/llms/anthropic/common_utils.py b/litellm/llms/anthropic/common_utils.py index d9424d6a243..2b57883cc13 100644 --- a/litellm/llms/anthropic/common_utils.py +++ b/litellm/llms/anthropic/common_utils.py @@ -67,6 +67,96 @@ _BEDROCK_VERSION_SUFFIX_RE: Final = re.compile(r"-v\d+(?::\d+)?$") _INFERENCE_PROFILE_MINOR_RE: Final = re.compile(r":\d+$") _DATED_RELEASE_SUFFIX_RE: Final = re.compile(r"-\d{8}$") _DOTTED_VERSION_RE: Final = re.compile(r"(\d)\.(\d)") +_CLAUDE_CODE_BILLING_HEADER_PREFIX: Final = "x-anthropic-billing-header:" +_CLAUDE_CODE_OBJECT_MAPPING_ADAPTER: Final = TypeAdapter(dict[object, object]) +_CLAUDE_CODE_OBJECT_LIST_ADAPTER: Final = TypeAdapter(list[object]) + + +def is_claude_code_user_agent(user_agent: str) -> bool: + return user_agent.startswith("claude-cli/") + + +def _validated_claude_code_mapping(value: object) -> dict[object, object] | None: + try: + return _CLAUDE_CODE_OBJECT_MAPPING_ADAPTER.validate_python(value) + except ValidationError: + return None + + +def _validated_claude_code_list(value: object) -> list[object] | None: + try: + return _CLAUDE_CODE_OBJECT_LIST_ADAPTER.validate_python(value) + except ValidationError: + return None + + +def _claude_code_billing_fields(text: str) -> tuple[tuple[str, str], ...] | None: + stripped: Final = text.strip() + if "\n" in stripped or "\r" in stripped or not stripped.startswith(_CLAUDE_CODE_BILLING_HEADER_PREFIX): + return None + fields: Final = tuple( + field + for raw_field in stripped.removeprefix(_CLAUDE_CODE_BILLING_HEADER_PREFIX).split(";") + if (field := raw_field.strip()) + ) + if not fields or any("=" not in field for field in fields): + return None + parsed_fields: Final = tuple( + (parts[0].strip(), parts[1].strip()) for field in fields for parts in (field.split("=", 1),) + ) + if any(not key or not value for key, value in parsed_fields): + return None + return parsed_fields + + +def _claude_code_billing_texts(system: object) -> tuple[str, ...] | None: + if isinstance(system, str): + return (system,) + blocks: Final = _validated_claude_code_list(system) + if blocks is None: + return None + block_mappings: Final = tuple(_validated_claude_code_mapping(block) for block in blocks) + if any(block is None for block in block_mappings): + return None + text_values: Final = tuple( + block.get("text") for block in block_mappings if block is not None and block.get("type") == "text" + ) + if len(text_values) != len(blocks) or any(not isinstance(text, str) for text in text_values): + return None + meaningful_text: Final = tuple(text for text in text_values if isinstance(text, str) and text.strip()) + return meaningful_text or None + + +def _is_claude_code_subagent_billing_system(system: object) -> bool: + billing_texts: Final = _claude_code_billing_texts(system) + if billing_texts is None: + return False + billing_fields: Final = tuple( + fields for text in billing_texts if (fields := _claude_code_billing_fields(text)) is not None + ) + if len(billing_fields) != len(billing_texts): + return False + subagent_values: Final = tuple( + value for fields in billing_fields for key, value in fields if key == "cc_is_subagent" + ) + return subagent_values == ("true",) + + +def is_claude_code_one_shot_subagent_request( + messages: list[AllMessageValues], + system: object, + tools: object, + user_agent: str | None, +) -> bool: + only_message: Final = _validated_claude_code_mapping(messages[0]) if len(messages) == 1 else None + return ( + user_agent is not None + and is_claude_code_user_agent(user_agent) + and not tools + and only_message is not None + and only_message.get("role") == "user" + and _is_claude_code_subagent_billing_system(system) + ) def _strip_bedrock_id_suffixes(model: str) -> str: diff --git a/litellm/proxy/litellm_pre_call_utils.py b/litellm/proxy/litellm_pre_call_utils.py index 56512570448..3c186e19829 100644 --- a/litellm/proxy/litellm_pre_call_utils.py +++ b/litellm/proxy/litellm_pre_call_utils.py @@ -789,12 +789,6 @@ def apply_missing_session_id_policy( ) -def is_claude_code_user_agent(user_agent: str) -> bool: - """Claude Code identifies itself as ``claude-cli/ ...``; the IDE - extensions and the Agent SDK run through the same CLI and share that prefix.""" - return user_agent.startswith("claude-cli/") - - def is_codex_user_agent(user_agent: str) -> bool: """Codex builds its user agent as ``/ ...`` and ships several first-party originators: ``codex-tui``, ``codex_cli_rs``, @@ -811,6 +805,8 @@ def should_auto_drop_params_for_agentic_cli(user_agent: str, data: dict, proxy_c requests routed to providers that reject them. An explicit drop_params from the caller or in the operator's ``litellm_settings`` always wins over this default.""" + from litellm.llms.anthropic.common_utils import is_claude_code_user_agent + if not (is_claude_code_user_agent(user_agent) or is_codex_user_agent(user_agent)): return False if "drop_params" in data: diff --git a/litellm/router_utils/pre_call_checks/prompt_caching_deployment_check.py b/litellm/router_utils/pre_call_checks/prompt_caching_deployment_check.py index 0788c8db710..70362e60495 100644 --- a/litellm/router_utils/pre_call_checks/prompt_caching_deployment_check.py +++ b/litellm/router_utils/pre_call_checks/prompt_caching_deployment_check.py @@ -90,6 +90,7 @@ class PromptCachingDeploymentCheck(CustomLogger): enable_prompt_caching=( request_kwargs.get("enable_prompt_caching") is True if request_kwargs is not None else None ), + request_kwargs=request_kwargs, ) model_id_dict: Final = await prompt_cache.async_get_model_id( diff --git a/tests/test_litellm/integrations/test_anthropic_cache_control_hook.py b/tests/test_litellm/integrations/test_anthropic_cache_control_hook.py index de8b654987b..6b7780acd20 100644 --- a/tests/test_litellm/integrations/test_anthropic_cache_control_hook.py +++ b/tests/test_litellm/integrations/test_anthropic_cache_control_hook.py @@ -1777,6 +1777,224 @@ class TestEnableAnthropicPromptCaching: assert messages == before +class TestClaudeCodeOneShotAutoCaching: + BILLING_TEXT = "x-anthropic-billing-header: cc_version=2.1.263; cc_entrypoint=cli; cc_is_subagent=true;" + BILLING_SYSTEM = [{"type": "text", "text": BILLING_TEXT}] + MESSAGES = [{"role": "user", "content": [{"type": "text", "text": "unique fetched document"}]}] + + @staticmethod + def _kwargs(configured=None): + kwargs = { + "litellm_metadata": {}, + "proxy_server_request": { + "headers": { + "user-agent": "claude-cli/2.1.263 (external, cli)", + "x-app": "cli-bg", + } + }, + } + if configured is not None: + kwargs["cache_control_injection_points"] = configured + return kwargs + + @pytest.mark.parametrize( + "system", + [ + BILLING_TEXT, + BILLING_SYSTEM, + [*BILLING_SYSTEM, {"type": "text", "text": " "}], + [ + *BILLING_SYSTEM, + {"type": "text", "text": "x-anthropic-billing-header: cc_version=2.1.263; cc_entrypoint=cli;"}, + ], + ], + ids=["string", "text_block", "whitespace_block", "multiple_billing_blocks"], + ) + @pytest.mark.parametrize("tools", [None, []], ids=["absent_tools", "empty_tools"]) + def test_skips_defaults_and_attribution_for_one_shot_subagent(self, monkeypatch, system, tools): + monkeypatch.setattr(litellm, "enable_anthropic_prompt_caching", True) + messages = copy.deepcopy(self.MESSAGES) + kwargs = self._kwargs() + + result_messages, result_system = AnthropicCacheControlHook.maybe_inject_cache_control( + messages, + copy.deepcopy(system), + kwargs, + model="claude-sonnet-4-5", + custom_llm_provider="anthropic", + tools=tools, + ) + + assert result_messages == self.MESSAGES + assert result_system == system + assert "litellm_gateway_injected_cache" not in kwargs["litellm_metadata"] + + def test_user_agent_header_lookup_is_case_insensitive(self, monkeypatch): + monkeypatch.setattr(litellm, "enable_anthropic_prompt_caching", True) + kwargs = self._kwargs() + user_agent = kwargs["proxy_server_request"]["headers"].pop("user-agent") + kwargs["proxy_server_request"]["headers"]["User-Agent"] = user_agent + + result_messages, result_system = AnthropicCacheControlHook.maybe_inject_cache_control( + copy.deepcopy(self.MESSAGES), + copy.deepcopy(self.BILLING_SYSTEM), + kwargs, + model="claude-sonnet-4-5", + custom_llm_provider="anthropic", + ) + + assert result_messages == self.MESSAGES + assert result_system == self.BILLING_SYSTEM + + def test_router_affinity_skips_string_billing_system(self, monkeypatch): + monkeypatch.setattr(litellm, "enable_anthropic_prompt_caching", True) + messages = copy.deepcopy(self.MESSAGES) + kwargs = self._kwargs() + kwargs["system"] = self.BILLING_TEXT + + result = AnthropicCacheControlHook.messages_with_default_injections( + messages=messages, + models=("claude-sonnet-4-5",), + request_kwargs=kwargs, + ) + + assert result == messages + + @pytest.mark.parametrize( + "headers,system", + [ + ("not-a-mapping", BILLING_SYSTEM), + ( + {"user-agent": "claude-cli/2.1.263 (external, cli)"}, + [{"type": "text", "text": "x-anthropic-billing-header: malformed"}], + ), + ({"user-agent": "claude-cli/2.1.263 (external, cli)"}, None), + ({"user-agent": "claude-cli/2.1.263 (external, cli)"}, ["not-a-mapping"]), + ( + {"user-agent": "claude-cli/2.1.263 (external, cli)"}, + [{"type": "image", "text": BILLING_TEXT}], + ), + ], + ids=["malformed_headers", "malformed_billing", "missing_system", "malformed_block", "non_text_block"], + ) + def test_malformed_untrusted_context_keeps_defaults(self, monkeypatch, headers, system): + monkeypatch.setattr(litellm, "enable_anthropic_prompt_caching", True) + + points = AnthropicCacheControlHook.get_default_injection_points( + messages=copy.deepcopy(self.MESSAGES), + system=copy.deepcopy(system), + model="claude-sonnet-4-5", + custom_llm_provider="anthropic", + request_kwargs={"proxy_server_request": {"headers": headers}}, + ) + + assert len(points) == 2 + + def test_message_without_role_keeps_defaults(self, monkeypatch): + monkeypatch.setattr(litellm, "enable_anthropic_prompt_caching", True) + + points = AnthropicCacheControlHook.get_default_injection_points( + messages=[{"content": "missing role"}], + system=copy.deepcopy(self.BILLING_SYSTEM), + model="claude-sonnet-4-5", + custom_llm_provider="anthropic", + request_kwargs=self._kwargs(), + ) + + assert len(points) == 2 + + @pytest.mark.parametrize( + "messages,system,tools", + [ + ( + MESSAGES, + BILLING_SYSTEM, + [{"name": "WebFetch", "description": "fetch", "input_schema": {"type": "object"}}], + ), + (MESSAGES, [*BILLING_SYSTEM, {"type": "text", "text": "Explore the repository"}], None), + ( + [ + {"role": "user", "content": "first turn"}, + {"role": "assistant", "content": "reply"}, + *MESSAGES, + ], + BILLING_SYSTEM, + None, + ), + ], + ids=["tools", "real_system", "history"], + ) + def test_keeps_defaults_for_reusable_subagents(self, monkeypatch, messages, system, tools): + monkeypatch.setattr(litellm, "enable_anthropic_prompt_caching", True) + kwargs = self._kwargs() + + result_messages, result_system = AnthropicCacheControlHook.maybe_inject_cache_control( + copy.deepcopy(messages), + copy.deepcopy(system), + kwargs, + model="claude-sonnet-4-5", + custom_llm_provider="anthropic", + tools=copy.deepcopy(tools), + ) + + assert AnthropicCacheControlHook.count_request_cache_breakpoints(result_messages, result_system) == 2 + assert kwargs["litellm_metadata"]["litellm_gateway_injected_cache"] == "" + + @pytest.mark.parametrize( + "user_agent,system", + [ + ("anthropic-sdk-python/0.75.0", BILLING_SYSTEM), + ( + "claude-cli/2.1.263 (external, cli)", + [ + { + "type": "text", + "text": f"{BILLING_TEXT}\nadditional system instructions", + } + ], + ), + ( + "claude-cli/2.1.263 (external, cli)", + [ + { + "type": "text", + "text": "x-anthropic-billing-header: cc_version=2.1.263; cc_is_subagent=false;", + } + ], + ), + ], + ids=["different_client", "appended_instructions", "not_a_subagent"], + ) + def test_ambiguous_or_unmatched_signals_fail_open(self, monkeypatch, user_agent, system): + monkeypatch.setattr(litellm, "enable_anthropic_prompt_caching", True) + kwargs = self._kwargs() + kwargs["proxy_server_request"]["headers"]["user-agent"] = user_agent + + result_messages, result_system = AnthropicCacheControlHook.maybe_inject_cache_control( + copy.deepcopy(self.MESSAGES), + copy.deepcopy(system), + kwargs, + model="claude-sonnet-4-5", + custom_llm_provider="anthropic", + ) + + assert AnthropicCacheControlHook.count_request_cache_breakpoints(result_messages, result_system) == 2 + + def test_explicit_injection_points_remain_authoritative(self, monkeypatch): + monkeypatch.setattr(litellm, "enable_anthropic_prompt_caching", True) + kwargs = self._kwargs([{"location": "message", "role": "user"}]) + + result_messages, _ = AnthropicCacheControlHook.maybe_inject_cache_control( + copy.deepcopy(self.MESSAGES), + copy.deepcopy(self.BILLING_SYSTEM), + kwargs, + model="claude-sonnet-4-5", + custom_llm_provider="anthropic", + ) + + assert result_messages[0]["content"][-1]["cache_control"] == {"type": "ephemeral"} + + class TestPerKeyEnablePromptCaching: """Per-request enable_prompt_caching override (stamped from key metadata) with the global flag off.""" @@ -1977,6 +2195,25 @@ class TestConfiguredInjectionPointsStandDown: _, result_sys = self._inject(copy.deepcopy(self.V1_MESSAGES), kwargs) assert result_sys == [{"type": "text", "text": "sys", "cache_control": {"type": "ephemeral"}}] + @pytest.mark.parametrize( + "configured", + [None, CONFIGURED], + ids=["automatic_defaults", "configured_points"], + ) + def test_v1_messages_stands_down_for_root_cache_control(self, monkeypatch, configured): + monkeypatch.setattr(litellm, "enable_anthropic_prompt_caching", True) + root_cache_control = {"type": "ephemeral"} + kwargs = {"cache_control": root_cache_control, "litellm_metadata": {}} + if configured is not None: + kwargs["cache_control_injection_points"] = copy.deepcopy(configured) + + result_messages, result_system = self._inject(copy.deepcopy(self.V1_MESSAGES), kwargs) + + assert result_messages == self.V1_MESSAGES + assert result_system == "sys" + assert kwargs["cache_control"] is root_cache_control + assert "litellm_gateway_injected_cache" not in kwargs["litellm_metadata"] + def test_v1_messages_reentry_flow_preserves_tool_config_remainder(self): """The advisor interceptor re-enters anthropic_messages() with the outer request's kwargs and post-injection messages. The first pass applies the diff --git a/tests/test_litellm/llms/anthropic/test_anthropic_common_utils.py b/tests/test_litellm/llms/anthropic/test_anthropic_common_utils.py index 794613942a1..ae620fdd6dc 100644 --- a/tests/test_litellm/llms/anthropic/test_anthropic_common_utils.py +++ b/tests/test_litellm/llms/anthropic/test_anthropic_common_utils.py @@ -26,6 +26,30 @@ FAKE_REGULAR_KEY = "sk-ant-api03-regular-key-for-testing-123456789" FAKE_AUTH_TOKEN = "sk-ant-aut01-fake-auth-token-for-testing-123456789" +@pytest.mark.parametrize( + "messages,system,expected", + [ + ([{"role": "user", "content": "hi"}], "x-anthropic-billing-header: cc_is_subagent=true;", True), + ([{"role": "user", "content": "hi"}], "x-anthropic-billing-header: =junk; cc_is_subagent=true;", False), + ([{"role": "user", "content": "hi"}], "x-anthropic-billing-header: cc_version=; cc_is_subagent=true;", False), + ([{"role": "user", "content": "hi"}], "x-anthropic-billing-header: malformed", False), + ([{"content": "missing role"}], "x-anthropic-billing-header: cc_is_subagent=true;", False), + (["not-a-mapping"], "x-anthropic-billing-header: cc_is_subagent=true;", False), + ([{"role": "user", "content": "hi"}], ["not-a-mapping"], False), + ([{"role": "user", "content": "hi"}], None, False), + ], +) +def test_is_claude_code_one_shot_subagent_request(messages, system, expected): + from litellm.llms.anthropic.common_utils import is_claude_code_one_shot_subagent_request + + assert is_claude_code_one_shot_subagent_request( + messages=messages, + system=system, + tools=None, + user_agent="claude-cli/2.1.263 (external, cli)", + ) is expected + + class TestOptionallyHandleAnthropicOAuth: """Tests for optionally_handle_anthropic_oauth function.""" diff --git a/tests/test_litellm/router_utils/pre_call_checks/test_prompt_caching_deployment_check.py b/tests/test_litellm/router_utils/pre_call_checks/test_prompt_caching_deployment_check.py index 79ae00e155c..030bdfe03e9 100644 --- a/tests/test_litellm/router_utils/pre_call_checks/test_prompt_caching_deployment_check.py +++ b/tests/test_litellm/router_utils/pre_call_checks/test_prompt_caching_deployment_check.py @@ -332,6 +332,67 @@ async def test_per_request_enable_prompt_caching_reaches_the_affinity_key(monkey assert filtered == [deployments[1]] +@pytest.mark.asyncio +async def test_claude_code_one_shot_subagent_does_not_reuse_an_auto_injected_affinity_key(monkeypatch): + monkeypatch.setattr(litellm, "enable_anthropic_prompt_caching", True) + cache = DualCache() + check = PromptCachingDeploymentCheck(cache=cache) + deployments = _deployments(AUTO_CACHING_MODEL, AUTO_CACHING_MODEL) + messages = cast(List[AllMessageValues], [{"role": "user", "content": "unique " * 3000}]) + request_kwargs = { + "system": [ + { + "type": "text", + "text": "x-anthropic-billing-header: cc_version=2.1.263; cc_is_subagent=true;", + } + ], + "proxy_server_request": {"headers": {"user-agent": "claude-cli/2.1.263 (external, cli)"}}, + } + auto_injected_messages = AnthropicCacheControlHook.messages_with_default_injections( + messages=messages, + models=(AUTO_CACHING_MODEL,), + ) + assert auto_injected_messages != messages + await PromptCachingCache(cache=cache).async_add_model_id( + model_id="dep-2", messages=auto_injected_messages, tools=None + ) + + filtered = await check.async_filter_deployments( + model=MODEL_GROUP_ALIAS, + healthy_deployments=deployments, + messages=messages, + request_kwargs=request_kwargs, + ) + + assert filtered == deployments + + +@pytest.mark.asyncio +async def test_root_cache_control_does_not_reuse_an_auto_injected_affinity_key(monkeypatch): + monkeypatch.setattr(litellm, "enable_anthropic_prompt_caching", True) + cache = DualCache() + check = PromptCachingDeploymentCheck(cache=cache) + deployments = _deployments(AUTO_CACHING_MODEL, AUTO_CACHING_MODEL) + messages = _auto_caching_messages() + auto_injected_messages = AnthropicCacheControlHook.messages_with_default_injections( + messages=messages, + models=(AUTO_CACHING_MODEL,), + ) + assert auto_injected_messages != messages + await PromptCachingCache(cache=cache).async_add_model_id( + model_id="dep-2", messages=auto_injected_messages, tools=None + ) + + filtered = await check.async_filter_deployments( + model=MODEL_GROUP_ALIAS, + healthy_deployments=deployments, + messages=messages, + request_kwargs={"cache_control": {"type": "ephemeral"}}, + ) + + assert filtered == deployments + + @pytest.mark.asyncio async def test_tool_marked_cache_control_keeps_routing_off_another_requests_prefix(monkeypatch, local_model_cost_map): """