Tags: strands-agents/sdk-python
Tags
feat(models): add SystemContentBlock support for provider-agnostic ca… …ching (#1112) * feat(model): support prompt caching via SystemContentBlock * fix: concat text blocks for system_prompt * remove litellm and openai changes for now * integ tests * linting * linting * fix test * add test cases
Transform invalid tool usages on sending, not on initial detection (#… …1091) Per bug #1069, session-managers never persist tool-name changes after we initially persist the message, which means once an agent generates an invalid-tool name, that message history is poisoned on re-hydration. To avoid that going forward, do the translation of invalid-tool names on sending to the provider and not on the initial tool_use detection. The initial tool_use detection is needed to add a tool_response with a proper error message for the LLM, but this will avoid the poisoning issue --------- Co-authored-by: Mackenzie Zastrow <zastrowm@users.noreply.github.com>
Fix additional_args passing in SageMakerAIModel (#983) * fix(sagemaker): additional_args dict issue Fix error where passing an additional_args dict to SageMakerAIModel would raise an AttributeError because Python dicts have no '__dict__' attribute. Fixes #982 * fix(sagemaker): typing for endpoint_config Fix typing for SageMakerAIModel.endpoint_config which was previously being treated as an arbitrary dictionary due to init assignment. * fix(sagemaker): Typing for payload_config Fix typing for SageMakerAIModel.payload_config, which was previously being treated as a plain dict due to init assignment. * test(sagemaker): tests for ep additional_args Add a test to check for insertion of endpoint config additional_args * fix(sagemaker): include payload additional_args Copy SageMakerAIPayloadSchema's additional_args into request payloads where provided - previously these were being ignored. Includes unit tests.
PreviousNext