New features enable businesses to build and deploy agents with improved security and functionality

OpenAI’s Agents SDK update aims to provide enterprises with the tools needed to create automated agents that can perform complex tasks. The new features, including sandboxing and in-distribution harness capabilities, are designed to improve agent safety and functionality.

The sandbox integration allows agents to operate within controlled computer environments, reducing the risk of unpredictable behavior. This is particularly important for businesses looking to deploy agents in production settings.

The in-distribution harness enables developers to build agents that can work with files and approved tools within a workspace, streamlining deployment and testing processes.

OpenAI plans to continue expanding the Agents SDK over time, with a focus on adding more agent capabilities. The company is also working to bring additional features, such as support for multi-agent systems, to the platform.

What matters

• Agents can now operate in controlled environments with sandboxing capabilities
• In-distribution harness allows agents to work with files and approved tools within a workspace
• Initial release focuses on Python, with TypeScript support planned for later release

Why it matters

Initial release focuses on Python, with TypeScript support planned for later release

This GenAI News article was prepared in original wording using reporting and materials published by TechCrunch AI. Source reference: https://techcrunch.com/2026/04/15/openai-updates-its-agents-sdk-to-help-enterprises-build-safer-more-capable-agents/.

Drafted by the GenAI News review pipeline.