Latest news across AI, security, and the OpenClaw ecosystem.
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.
Source: Hermes Agent ReleasesOpenAI 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.
Source: Simon WillisonA 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.
Source: HN Show HNMicrosoft 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.
Source: Microsoft Security BlogBruce 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.
Source: Schneier on SecurityGitHub 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.
Source: Kilo BlogOllama 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.
Source: Ollama Releasesboxes.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.
Source: HN Show HNNVIDIA’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.
Source: Kilo BlogOpenAI 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.
Source: OpenAI NewsAsana 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.
Source: SiliconANGLEOpenAI 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.
Source: CIO AIA 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.
Source: HN Show HNPocketPing is an open-source chat widget that lets support staff reply from Telegram, Slack, or Discord instead of a separate inbox. The project also offers a hosted HubSpot inbox, webhooks, an MCP server, and self-hosted deployment options.
Source: HN Show HNOllama released v0.30.3 on June 3, adding support for gemma4-12b in its model list, according to the GitHub release page. The release was tagged by pdevine and points to commit 50bbda5.
Source: Ollama ReleasesUK banks have been offered access to OpenAI’s GPT-5.5 while being left out of Anthropic’s Glasswing expansion, according to the report. The move adds another option for financial firms as major AI vendors compete for enterprise customers.
Source: The Register SecurityA new Codex skill called Codex Reset Watchdog watches public X posts from Codex lead thsottiaux for possible quota reset signals. It can help users spot reset-related changes early and plan how to use any remaining Codex quota, including switching to fast mode when useful.
Source: HN Show HNv0.30.2 adds launch-time prompts to show and install the Cline CLI, along with Qwen code integration and Codex launch configuration isolation. The release also tightens URL handling, updates llama.cpp, and adds fixes for model limits, token counting, and load-stall detection.
Source: Ollama ReleasesOpenAI says it is improving GPT-5.5 Instant while phasing out older ChatGPT models, including o3 and GPT-4.5. The company is also adding job search features and better resume tools to ChatGPT as the updates roll out globally.
Source: BleepingComputerMicrosoft announced two new text models on June 2: MAI-Thinking-1, a reasoning model, and MAI-Code-1-Flash, a code-focused model for GitHub Copilot and Visual Studio Code. Microsoft says both were trained on clean, commercially licensed data, though the technical paper for MAI-Thinking-1 also describes large-scale web crawling and filtering.
Source: Simon WillisonOpenAI released new Codex features aimed at nontechnical users, including six plugins, a Sites tool and expanded Annotations support. The company said Codex now has more than 5 million weekly users, with nondevelopers accounting for about 20% of that base.
Source: SiliconANGLEAnthropic says it has hundreds of skills in active use inside Claude Code and has identified patterns for what works best. The company’s guidance covers skill types, folder structure, verification, sharing, and usage tracking.
Source: Claude BlogOpenAI expanded Codex with hosted web apps, in-place editing, and role-specific plugins that connect 62 business apps. The company said non-developers now make up about 20% of Codex’s 5 million weekly users and are adopting it three times faster than engineers.
Source: The Next WebMicrosoft announced new AI infrastructure for developers and business users at Build in San Francisco, including Microsoft IQ, Work IQ, Fabric IQ, Foundry IQ and Web IQ. The company also released new in-house models for reasoning, image, speech and code, plus a security system that uses more than 100 agents to find bugs.
Source: SiliconANGLEMicrosoft has launched Scout, an AI assistant built on the OpenClaw framework for use inside Microsoft 365. The company says it is designed to learn from user feedback, run across desktop and web, and include security checks and audit trails.
Source: TechCrunchMicrosoft used Build 2026 to announce new security features across code, agents, and models. The company said the updates are meant to give developers real-time guidance and give security teams more visibility across the development lifecycle.
Source: Microsoft Security BlogGitHub COO Kyle Daigle said the company is adapting its internal tools and infrastructure for a surge in AI-generated code and agent-driven workflows. He described using GitHub Copilot, Actions, Slack, Teams, and Microsoft’s WorkIQ MCP server to help teams review past activity and plan next steps.
Source: Latent SpaceWorkday announced new developer tools at DevCon 2026, including Developer Agent, Agent-Ready Tools and Agent Passport. The company says the features let developers build agents faster while adding third-party verification for security and compliance before production.
Source: SiliconANGLEMicrosoft has shipped new Entra features over the past 30 days, including phishing-resistant MFA on Linux, passkey registration campaigns, and changes to Conditional Access and self-service password reset. The company also added governance, app deactivation, and SOC response tools, with more enforcement dates set for 2026.
Source: Help Net SecurityOpenAI said Codex now has role-specific plugins, shareable Sites, and in-place annotations for documents and slides. The company said more than 5 million people use Codex weekly, and non-developers now make up about 20% of users.
Source: OpenAI NewsOpenAI frontier models and Codex are now available on Amazon Bedrock, letting AWS customers use them inside existing cloud environments. OpenAI and AWS said the integration is designed to reduce procurement, security review, and deployment friction.
Source: Help Net SecurityAnthropic said it is expanding Project Glasswing, its effort to use Claude Mythos Preview to find software vulnerabilities, to about 150 additional organizations. The company said the first 50 partners found more than 10,000 high- or critical-severity flaws, and the new group includes infrastructure and open-source maintainers in more than 15 countries.
Source: Anthropic NewsClearscale says it is using AWS Transform, Amazon’s agentic AI migration service, to speed enterprise modernization for VMware, .NET, and mainframe workloads. The company says its Clearview Migration Methodology can reduce migration time by up to 80% and helped one global enterprise cut a 14-month VMware program to under six months.
Source: CIO AIAnthropic will give the EU cybersecurity agency ENISA access to Claude Mythos through Project Glasswing, ending weeks of stalled talks. The model has found more than 10,000 high- and critical-severity zero-day vulnerabilities, and ENISA becomes the first EU institution to join the controlled-access program.
Source: The Next WebCanvascript is selling museum-grade pigment prints of famous source code and software artifacts. The collection includes items such as the Perl Camel, the Apollo 11 AGC bailout code, the Bitcoin whitepaper, and the first CERN web page, with prices starting at $69.
Source: HN Show HNOpenAI has tagged Codex release rust-v0.136.0 on GitHub. The release points to commit 7ca6113 and is available on the openai/codex repository.
Source: Codex ReleasesJson2vec is an open-source PyTorch and Lightning toolkit that builds models directly from JSON-like schemas instead of flattening records first. It supports nested data, typed fields, masked reconstruction, embeddings, and batch inference through Lightning data modules and writers.
Source: HN Show HNDashVox.ai has launched an iPhone app that lets users start and steer Claude Code and Codex sessions running on their own machine over SSH by voice. The app supports phone, CarPlay, Android Auto, Apple Watch, Wear OS, and a self-hosted server option, with Android beta coming soon.
Source: HN Show HNOpenAI said its frontier models and Codex are now generally available on AWS. The company said the move lets enterprise customers use OpenAI inside existing AWS security, compliance, billing, and governance workflows.
Source: OpenAI NewsMilestones, a native project management app for Apple devices, now includes a built-in MCP server on Mac so users can connect Claude and other AI agents. The latest update also fixes several sync, layout, and task-editing issues.
Source: HN Show HNOpenClaw 2026.5.28 focuses on runtime recovery, channel delivery safety, and stricter validation across browser and automation inputs. The release also adds broader provider support, plugin updates, and improved release and QA checks, according to the project’s release notes.
Source: OpenClaw ReleasesMicrosoft is changing GitHub Copilot billing on June 1 from a flat subscription to token-based usage, and some developers say the new model could sharply raise their costs. Others argue the biggest bills reflect inefficient or excessive use, especially in “vibe coding” workflows.
Source: TechCrunchGoogle’s Gemini Spark is a 24/7 agentic assistant that runs in the cloud and handles tasks across Gmail, Calendar, Docs, Sheets and Slides. In early testing, it proved useful for shopping research, trip planning and inbox summaries, but it also showed gaps in integrations and task handling.
Source: TechCrunchClaude Opus 4.8 is now available in Kilo Code, and the company is offering a 20% discount on inference through its provider network. Kilo says the model is more reliable and efficient, with new orchestration features for agentic workflows and computer-use tasks.
Source: Kilo BlogHeypi is an early chat-agent runtime for teams that works with Slack, Discord, Telegram, webhooks, and approvals. The project includes optional Docker and VM runtime providers, plus example assistants for DevOps, project tracking, fitness coaching, and note taking.
Source: HN Show HNGoogle showed nine demos of Gemini Omni and Gemini 3.5 Flash at Google I/O 2026. Omni focuses on video generation and editing through conversation, while 3.5 Flash is aimed at agentic tasks, coding, and new search experiences.
Source: Google AI BlogJapan’s three megabanks will get access to OpenAI’s GPT-5.5-Cyber through the company’s Trusted Access for Cyber programme, according to Finance Minister Satsuki Katayama. The rollout comes amid a broader Japanese effort to treat frontier AI as both a cyber threat and a defensive tool.
Source: The Next WebBraintrust says its engineers now use OpenAI’s Codex with GPT-5.5 to turn customer feature requests into preview branches in minutes. The company says half its engineering team moved to Codex in one month, speeding up customer feedback and experimentation.
Source: OpenAI NewsHeadroom, a macOS menu bar app, says it can reduce Claude Code token use by about 50% by trimming prompt bloat locally before requests reach Anthropic’s model. The company says prompts never leave the user’s machine, and it has launched pricing tiers for individual Claude users and teams.
Source: HN Show HNMicrosoft has redesigned Microsoft 365 Copilot into a single entry point across Microsoft 365 apps, with a task-aware workspace, contextual prompts, and in-app actions. The company says the updated app loads twice as fast and that usage rose in Word, Excel, PowerPoint, and Outlook after the in-app experiences rolled out.
Source: Help Net Security