AWS opens Strands Box to constrain AI agent actions
Amazon Web Services released a developer preview of Strands Box, an open-source testing environment meant to rein in autonomous AI agents. The tool lets developers enforce customizable policies controlling how often or how far AI agents can act, even tracking their recent behavior to apply restrictions. Strands Box uses AWS’s open-source Dogwood policy engine and currently runs on Apple Silicon Macs with macOS 15 or newer.
- Open-source developer tool for limiting AI agent behavior
- Applies policies based on agents' recent activity
- Uses Dogwood policy engine to check requests
- Works on Macs with Apple Silicon, macOS 15+
- Some actions remain outside policy evaluation
Sources covering this
In this story
More in AI
Google Cloud launches Gemini agent for work
Google Cloud has launched the Gemini agent, an AI-powered assistant designed to handle a wide range of business tasks from a single…
Google publishes AI chatbot study in The Lancet
Google and Beth Israel Deaconess ran a real-world test of an AI chatbot for early patient assessment, with results published in The Lancet.
Anthropic updates user policy to limit abuse toward Claude AI
Anthropic has updated its usage policy to prohibit users from engaging in sustained, needless cruel behavior toward its Claude AI models.
Nvidia uses frontier AI agents to automate simulation design
Nvidia developers are now using advanced AI models, like GPT-6 Astra, to automate parts of building complex simulation environments with…