deepagents==0.7.17
FoundationVision/VARdeepagents==0.7.17Sep 22, 2026by github-actions[bot]
Upgrading?
AI Summary
Fixes backend exit status handling and restricts inline file blocks to supported MIME types to improve security and stability.
Key Highlights
- Preserve backend exit status when capture metadata is missing
- Restrict inline file blocks to supported MIME types
New Features
- Backend exit status preservation
- MIME type restriction for file blocks
Full Release Notes
### Bug Fixes - Preserve the backend exit status when capture metadata is missing. ([#6451](https://github.com/langchain-ai/deepagents/pull/6451)) - Restrict inline file blocks to supported MIME types. ([#6447](https://github.com/langchain-ai/deepagents/pull/6447)) --- **Special thanks** to everyone who reported the issues addressed in this release: @ramon-langchain ([#6150](https://github.com/langchain-ai/deepagents/issues/6150)) Internal maintainers: @ccurme, @mdrxy Released by: @mdrxy <details> <summary>Git log since deepagents==0.7.16</summary> This commit history includes changes to this package. Commits are listed newest first. - [`e5406d0`](https://github.com/langchain-ai/deepagents/commit/e5406d0c735c7eb72548b7ea616991987bfb21e6) release(deepagents): 0.7.17 (#6458) - [`cf5c560`](https://github.com/langchain-ai/deepagents/commit/cf5c5606648efa8bb27684fac97d9e794039707d) chore(deps): raise dependency minimums for `all` (#6478) - [`c9926b1`](https://github.com/langchain-ai/deepagents/commit/c9926b1a96d204309d0b211701b288ce3bb4244b) fix(sdk): preserve backend exit status when capture metadata is missing (#6451) - [`191e93a`](https://github.com/langchain-ai/deepagents/commit/191e93a05e30aad705ee2cb33d5b03e7104dbe5a) fix(sdk): restrict inline file blocks to supported MIME types (#6447) </details>
Related Repositories
rasbt/LLMs-from-scratch★ 107,042
Implement a ChatGPT-like LLM in PyTorch from scratch, step by step
lyogavin/airllm★ 40,155
AirLLM 2.8T LLM inference with single 4GB GPU
NirDiamant/RAG_Techniques★ 29,958
This repository showcases various advanced techniques for Retrieval-Augmented Generation (RAG) systems. Each technique has a detailed notebook tutorial.
microsoft/OmniParser★ 25,539
A simple screen parsing tool towards pure vision based GUI agent
karpathy/nn-zero-to-hero★ 24,906
Neural Networks: Zero to Hero