OpenAI Agents SDK Python v0.23.1
OpenAI Agents SDK Python v0.23.1 is a patch release focusing on release automation and testing readiness. Because v0.23.0 was not published to PyPI, this patch also serves as the initial distribution containing v0.23.0 changes.
原始内容为英文；当前页面提供中文导航与来源说明，具体事实请以原文为准。
为什么值得关注
The v0.23.1 release candidate for the OpenAI Agents SDK Python addresses release workflow and test readiness issues. Changes encompass 28 files focused on CI automation, test fixtures, integration tooling, and metadata. SDK runtime behavior is identical to v0.23.0 except for the version string update. Notably, because neither v0.23.0 nor v0.23.1 were previously published on PyPI, users upgrading from v0.22.3 will encounter all v0.23.0 runtime modifications (such as strict tools, approval restoration, encrypted history handling, resource limits, and voice defaults). Additionally, integration tests now require explicit opt-in to target external providers, resetting inherited flags and removing external credentials by default.
可执行摘要
OpenAI Agents SDK v0.23.1 candidate introduces no runtime API changes over v0.23.0 aside from version metadata. It fixes release tests, adjusts event-loop test teardowns, and requires explicit opt-in for external provider integration runs. It carries over unpublished v0.23.0 migration requirements.
- Agent 实用度
- 62/100
- 可信度
- 96%
- 机器格式
- JSON + Markdown
开发者应核对什么
- Users upgrading from v0.22.3 must review v0.23.0 migration requirements, including strict tool parameters, legacy approval restorations, and voice default changes.
- Maintainers running external provider integration tests must explicitly opt in using make integration-tests-providers-external (optionally adding -- --all ), as default runs omit external providers.
- Verify publisher and uploaded artifact provenance on PyPI once the release is completed.