All published AI news and analysis.

Strands Agents Tools is a Python library that provides a collection of tools designed to enhance the capabilities of AI agents. It offers functionalities for file operations, shell integration, memory management, web interactions, API requests, Python code execution, mathematical operations, and integrations with AWS, image/video/audio processing, and more.

OpenAI has published its public policy agenda, detailing its approach to AI safety, youth protection, and workforce transition. The company aims to establish global standards for AI to ensure societal benefit.

OpenAI has published a blueprint for the U.S. governance of frontier artificial intelligence. The proposal outlines a federal framework focused on ensuring safety, resilience, and national security.

Clawker, a new open-source, self-hosted sandbox for AI coding agents, allows users to run agents like Claude Code and OpenAI Codex in isolated Docker containers. It features a robust firewall and customizable build environments for enhanced security and flexibility.

looptech-ai has launched "understand-quickly," a public GitHub repository that serves as a registry for code-knowledge graphs designed for AI agents. This initiative aims to provide schema-validated content, functioning as an "Awesome-list 2.0" for AI agents to quickly understand codebases.

Aguara is a new open-source security scanner designed to detect vulnerabilities in AI agents and software supply chains. It operates locally, without requiring SaaS or LLM calls, and scans for issues like prompt injection, tool poisoning, and compromised packages across multiple ecosystems.

Anthropic has introduced the Services Track and Partner Hub for its Claude Partner Network. The Services Track categorizes partners into three tiers (Select, Preferred, Global Premier) based on their experience and customer deployments with Claude. The Partner Hub serves as a portal for partners to track their progress and for customers to find qualified firms.

Hugging Face has integrated MCP (Manipulation Control Program) tools into the Reachy Mini robot, enabling more precise and complex manipulation tasks. This enhancement aims to improve the robot's ability to interact with its environment through advanced control capabilities.

PentesterFlow has released an open-source terminal agent designed for offensive security tasks. It integrates with local and hosted LLMs, automates reconnaissance and validation, and requires human approval for sensitive actions.

Voocel has released AgentCore, a minimal and composable Go library designed for building AI agent applications. It offers a restrained core with open extensibility, focusing on stability for key components like Agent, AgentLoop, Event, Tool, and Message.