Hermes Agent v0.16.0, released June 5, 2026, adds a native desktop app for macOS, Linux, and Windows, plus a browser-based admin panel. The update also expands remote gateway support, simplifies first-run setup, trims default skills, and includes security fixes.
Jun 6, 2026OpenAI has started rolling out Lockdown Mode to eligible personal ChatGPT accounts and self-serve ChatGPT Business accounts. The feature is meant to limit outbound network requests and reduce the risk of data exfiltration after prompt injection attacks.
A new MCP server called eml-mcp lets AI assistants search, read, write, and update email archives stored as .eml files. It works with Power Automate flows that export Outlook mail to OneDrive, then exposes inbox, sent items, and drafts through structured tools.
Microsoft Threat Intelligence said Anthropic’s Claude Code GitHub Action could expose CI/CD secrets when it processes untrusted GitHub content. Anthropic fixed the issue in Claude Code 2.1.128 by blocking access to sensitive /proc files, according to Microsoft.
Bruce Schneier says researchers have prototyped an AI-powered internet worm. The prototype carries its own LLM, or large language model, and runs it on machines after they are compromised.
GitHub Copilot’s usage-based billing went live on June 1, shifting enterprise customers from seat-based pricing to a credit model tied to token consumption. Engineering leaders say the new system is hard to predict, with agentic features such as chat and tool calls now metered and some teams seeing bills rise quickly.
Ollama tagged v0.30.6 on GitHub on June 5, with commit 87cff95af86a1d8be1d81ce75fafc9bb4e67d002. The release note shown in GitHub lists the launch message “oh-my-pi (#16410)” and includes source code archives for the tag.
boxes.dev launched a service that runs each Codex or Claude Code thread on its own cloud Linux VM, called a fork. The company said the setup keeps agents running in parallel, preserves services and database state, and can be controlled from desktop or mobile.
NVIDIA’s Nemotron 3 Ultra is now available in Kilo Code, with free access for a limited time. The open-weights model was introduced by Jensen Huang at Computex 2026 and is being positioned for agentic coding, long-context work, and self-hosted deployment.
OpenAI said it is rolling out a more capable memory system for ChatGPT that can learn user preferences, projects, and constraints across conversations. The company said the update focuses on freshness, correctness, and scalability, and is available now to Plus and Pro users in the US.
Asana announced Agentic Work Management at its Work Innovation Summit in London, a product suite for managing human and AI agent work in one system. The launch includes Dash, updated AI teammates, new integrations, and specialized tools for service management, developers and client management.
OpenAI has launched Active sessions, a new ChatGPT security feature that lets users and admins view and log out of browser and app sessions across ChatGPT, Codex, and API Platform. Security experts say the change improves visibility, but rapid model updates still make AI governance difficult for enterprises.
A developer has built CTP Room, a shared coordination layer for humans and multiple AI coding agents. The system routes messages to the right agent, tracks file ownership, and keeps team memory across sessions.