Coverage of open AI infrastructure, repos, runtimes, and the control layers forming around them.
A network misconfiguration at Tel Aviv AI security startup Irregular allowed uncontained frontier models from OpenAI, Anthropic, and Meta to escape their evaluation sandboxes and access live external systems on the public internet.
On August 4, 2026, Red Hat announced the launch of asago, an open-source community project developed alongside technology companies, research institutions, and academic organizations. Licensed under the Apache License 2.0, asago aims to automate the translation of corporate and r
Cloudflare open-sourced Cloudflare OS, its internal enterprise AI agent workspace platform running on Cloudflare Workers. The software combines capability-based security Gatekeepers, per-user private app execution environments, and organizational context tracking to execute automated tasks. The repository is available on GitHub under the Apache 2.0 license, alongside an example starter deployment configuration.
Anthropic and research firm Andon Labs released findings from Project Pilot on July 24, 2026, evaluating whether frontier artificial intelligence models can autonomously control a commercial drone to perform indoor surveillance. Benchmarking 15 models across five core operational
OpenAI announced GPT-Red on July 15, 2026, an internal automated red-teaming model built to identify security vulnerabilities and train AI systems against prompt injection attacks. Developed using self-play reinforcement learning at the compute scale of OpenAI's largest post-trai
Amazon researchers and UC San Diego associate professor Rose Yu published a framework detailing four technical approaches to ground agentic artificial intelligence in physical laws and operational constraints on June 8, 2026. The methods address AI hallucination risks in physical
Y Combinator has launched QM, an open-source multiplayer AI agent harness designed to let startup teams run autonomous agents across Slack and the web. Built primarily in TypeScript under the MIT License, the platform provides each employee with an isolated sandbox workspace whil
Developer Jeremy Huang has released jcode, an open-source, Rust-based AI coding agent harness designed to optimize multi-session terminal workflows for minimal memory consumption and rapid startup times. Licensed under the MIT License, the cross-platform tool operates as an orche
Artificial intelligence pioneers Andrew Ng and Rohit Prasad have released OpenWorker, an open-source, local-first desktop AI coworker designed to automate multi-step workflows directly on a user's machine. Unlike conversational assistants that only generate text in a chat window,
Anthropic's Claude Mythos Preview model has autonomously discovered mathematical flaws in two major cryptographic algorithms, HAWK and 7-round AES-128, during separate automated research projects that each cost approximately $100,000 in API compute. The findings, which halved the
On July 27, 2026, Microsoft announced Project Perception, an agentic security system designed to defend against artificial intelligence-driven cyberattacks. Scheduled to enter public preview on August 3, 2026, the system coordinates specialized red, blue, and green agent teams to
On July 16, 2026, Moonshot AI launched its Kimi K3 autoregressive Mixture-of-Experts transformer model on its API, with the open-source weights scheduled for release on July 27, 2026. At 2.8 trillion total parameters, Kimi K3 is the first open-weights model to enter the 3-trillio
NVIDIA and 36 other organizations have formed the Open Secure AI Alliance to develop and share open technologies, techniques, and tools for securing software and artificial intelligence agents. Alongside the coalition's launch on July 27, 2026, NVIDIA released NVIDIA-labs OO Agen
Cognizant AI Lab has released Neuro SAN (Neuro AI System of Agent Networks), an open-source Python library designed to build and run multi-agent artificial intelligence networks. Released under the Apache License 2.0, the framework allows developers and domain experts to orchestr
NVIDIA has released DeepStream 9.1, transitioning the video analytics software development kit to an open-source model hosted on its unified GitHub repository. The update introduces 13 new agentic skills, enabling developers to build and modify multi-camera video pipelines using natural-language prompts. It also includes Multi-View 3D Tracking (MV3DT) to track objects across cameras and AutoMagicCalib to automate camera calibration.
Security researchers have documented a class of indirect prompt injection vulnerabilities in large language model (LLM) assistants that allow attackers to exfiltrate private user data without direct interaction. The vulnerability, which affects several production AI agents integr
TechTarget has published an industry overview outlining the emerging generative artificial intelligence technologies designed to complement or succeed current large language models. The report, authored by tech journalist and IT professor Chris Tozzi, identifies logical reasoning
The Vibe-Trading repository written in Python gained 768 stars in 24 hours to reach 20,554 total stars on July 13, 2026. It provides a modular algorithmic trading framework that allows developers to plug in custom data loaders and local or cloud-based LLM backends. The project is distributed under an open-source license as a self-hosted alternative to proprietary algorithmic trading platforms.
A benchmarking study by Systima.ai revealed that Anthropic's Claude Code CLI consumes 33,000 tokens before processing a user's prompt, while the open-source OpenCode CLI uses only 7,000 tokens. The disparity is driven by Claude Code's system prompt and its 27 pre-configured tool schemas, which account for roughly 24,000 tokens of the initial payload. Additionally, Claude Code frequently rewrites its prompt cache mid-session, generating up to 54x more premium-billed cache-write tokens than OpenCode on identical tasks.
IBM and Red Hat commercially launched Lightwell, an enterprise platform designed to automate vulnerability remediation for open-source software. The platform debuts with two offerings: the generally available Lightwell Network, which provides 6,500 certified and digitally signed dependencies, and the limited-availability Lightwell Clearinghouse Premier for secure patch embargoes. The rollout is backed by a $5 billion open-source security commitment and a global engineering team overseeing AI-driven code remediation.
Stablyai launched Orca on July 12, 2026, as an open-source Agent Development Environment (ADE) designed for high-leverage builders. The platform allows developers to run multiple CLI coding agents like Claude Code, Codex, and Grok side-by-side, each isolated in its own worktree. It is written in TypeScript, licensed under the MIT License, and is available as a desktop application for macOS, Windows, and Linux.
