If you’ve been running OpenClaw with OpenAI models and paying separately for API access, that just changed. OpenClaw now routes all OpenAI model requests through the native Codex runtime harness by default, and it’s powered by your existing ChatGPT Plus, Pro, or Team subscription. No separate API key. No double billing. Sign in via OAuth, and your ChatGPT plan funds your OpenClaw agents. What Changed OpenAI announced the Codex runtime integration in early May 2026. The key shift: instead of OpenClaw calling the OpenAI API directly with a developer API key, it now uses the Codex runtime harness — the same infrastructure powering Codex in the ChatGPT interface — accessed through your existing subscription credentials. ...

OpenClaw is running two release tracks simultaneously this week: v2026.5.12 is the latest stable release, and v2026.5.16-beta.1 is the current leading beta. Together they represent the most significant package of security and performance improvements OpenClaw has shipped since earlier this year. Here’s what’s in each track, why the changes matter for production agent deployments, and how to think about upgrading. v2026.5.12 Stable — What Changed Major Memory Leak Fix The most important change in v2026.5.12 is a fix for a memory leak that caused significant RSS (Resident Set Size) growth when processing large transcripts. ...

Security researchers have disclosed four chained vulnerabilities in OpenClaw — collectively dubbed “Claw Chain” — that together enable data theft, privilege escalation, and persistent backdoor access. If you’re running any version of OpenClaw prior to v2026.4.22, you need to update now. ⚠️ Quick version check: If you’re already on v2026.4.22 or later (including v2026.5.12), you are already protected. The patch was released in late April 2026. This article is for the significant number of users who have not yet upgraded. ...

xAI’s Grok Build launched in early beta on May 14, 2026 — and if you’re a SuperGrok Heavy subscriber, you already have access to one of the most ambitious agentic coding CLIs in the field. This guide walks you through what you need to get started, what to expect from the parallel subagent architecture, and how to make the most of the Plan–Review–Approve workflow. Note: Grok Build is in early beta as of this writing. Some installation steps, specific commands, and configuration options may change as the product matures. Always refer to xAI’s official documentation at x.ai for the most current setup instructions. ...

Microsoft is canceling the majority of its internal Claude Code licenses and migrating its engineers to GitHub Copilot CLI by June 30, according to reporting from The Verge and confirmed across multiple X sources. The deadline gives affected developers approximately six weeks to transition. The move is significant not because it ends the Microsoft-Anthropic relationship — that investment partnership remains very much intact — but because it signals how Microsoft is thinking about tooling consolidation at the enterprise level. When a company is both a major investor in Anthropic and building its own competitive product, eventually someone has to choose. ...

The agentic coding CLI space just got a lot more crowded. On May 14, xAI launched Grok Build in early beta — a terminal-native agentic coding tool built on a parallel subagent architecture that positions itself as a direct rival to Anthropic’s Claude Code and OpenAI’s Codex CLI. If you’ve been watching the agentic coding race, this is the shot from xAI’s corner you probably knew was coming. And the early feature set is ambitious. ...

If you’re using the Claude Agent SDK, the claude -p command, or Claude Code’s GitHub Actions integration, your billing structure is about to change. Starting June 15, 2026, Anthropic is separating Agent SDK usage from standard interactive plan limits — and introducing a dedicated monthly credit for programmatic use. This is good news for most developers building with agents. Here’s what’s changing and what it means for your bill. The Core Change Until now, Agent SDK usage and interactive Claude usage drew from the same pool — your subscription’s usage limits. This created friction for developers building agentic workflows: a long-running Agent SDK job could eat into the quota you needed for Claude.ai conversations, and vice versa. ...

Anyone who has built real agentic workflows knows the failure mode: the agent declares success, you check the output, and something critical is missing or broken. The agent genuinely believed it was done. It wasn’t. Claude Code v2.1.139, shipped May 12, 2026, introduces a structural fix to this problem: the /goal command. The Fundamental Architecture Problem Traditional agentic coding systems — including earlier versions of Claude Code — use a single model for both execution and self-evaluation. The same model that writes the code also decides when the task is complete. This creates a well-documented failure mode: once an agent has committed to an approach, its self-evaluation is biased toward confirming that approach worked. ...

Before Google could announce it on stage, the internet found it. Leaked onboarding screens from the Gemini web beta and strings pulled from a decompiled APK (v17.23) have revealed Gemini Spark — Google’s next major AI evolution that goes well beyond answering questions. Gemini Spark is a proactive, always-on AI agent. And it represents a fundamental shift in how Google envisions AI assistants working. What Gemini Spark Actually Does Based on the leaked materials — which include onboarding screens shared by researcher @mrfanduuuuu and APK string analysis from Testing Catalog and Android Authority — Gemini Spark is designed to operate in the background, taking action without being asked. ...

If you’ve run long agentic sessions in Claude Code, you’ve probably hit this: the agent wraps up, reports success, and something is still broken. The model didn’t lie — it genuinely assessed its work as complete. But it was wrong. Claude Code v2.1.139 (shipped May 12, 2026) addresses this with the /goal command, a feature that separates the model doing the work from the model judging whether that work is done. This guide explains how it works and how to write goals that actually catch incomplete work. ...