News
What just dropped, what it means, and whether you should care.

Plugin4Shell: Is Your AI Coding Agent Still Vulnerable?
Plugin4Shell is a zero-click remote code execution flaw that broke SHA pinning across Claude Code, OpenAI Codex, GitHub Copilot, and Gemini CLI. Claude Code (fixed in v2.1.179, current v2.1.278 is saf

What's New in OpenClaw v2026.9.5: Atomic Updates and Live Plugin Management
OpenClaw v2026.9.5, released September 19, 2026, introduces Atomic Updates that verify the next version before committing the switch, live plugin management without Gateway restarts, conversation shar

Claude Code v2.1.276: MCP 2026-07-28 Is Now the Default -- What Changed and What to Update
Claude Code v2.1.276, released September 18 2026, makes the MCP 2026-07-28 stateless spec the default for all Bedrock, Vertex AI, Azure Foundry, and telemetry-disabled deployments. The initialize hand

n8n 2.39.6 and CrewAI 1.15.22: Agent Framework Updates Decoded
n8n shipped two patches this week -- 2.39.5 on September 14 and 2.39.6 on September 16 -- fixing credential revocation bugs, Gateway credits logic, and flagging two configurations that will break in v

Colibri: How to Run Frontier MoE Models on Your Own Hardware
Colibri is a pure-C, zero-dependency inference engine that runs Mixture-of-Experts models as large as 744 billion parameters on consumer hardware with 25GB of RAM. It works by streaming expert weights

Claude Code v2.1.270: What Plugin Evals and Output-Style Switching Actually Change
Claude Code v2.1.269, released September 11, 2026, shipped two significant developer-facing features: claude plugin eval runs your plugin's test suite against Claude Code and scores results in JSON an

OpenAI Agents API Beta: What Changed for Agent Builders
OpenAI's Agents API entered public beta on September 10, 2026, giving every developer with an API account access to the same orchestration harness powering Codex. Long-lived sessions, automatic contex

Claude Code 2.1.268: MCP Errors Were Leaking Credentials to Your Logs
Claude Code v2.1.268, released September 10, 2026, patches a security issue where MCP and plugin error output was writing credentials and tokens into log files. It also fixes a regression from v2.1.26

Claude Code v2.1.263: What bashOutputMaxChars Actually Changes for Agent Workflows
Claude Code v2.1.261 added two new settings -- bashOutputMaxChars and taskOutputMaxChars -- that raise how much command and background-task output agents see inline, up to 128K characters. v2.1.263 (r

Claude Code v2.1.260: Diff Panel, Cache Debugging, and What's Actually New
Claude Code v2.1.260 shipped September 3, 2026 at 23:48 UTC. The headline changes: a fullscreen /diff panel showing uncommitted changes as Claude edits, prompt cache miss causes now visible in /cost a

Anthropic EFS: How Split Custody Unblocks Enterprise Claude
Anthropic's Enterprise Frontier Safeguards (EFS), announced September 1, 2026, stores Claude conversation logs in the customer's own cloud infrastructure -- Amazon S3, Azure Blob Storage, or G

OpenClaw's 500-PR Sprint: Inside the Open-Source Agent Ecosystem
OpenClaw, Hermes Agent, IronClaw, QwenPaw, and ZeroClaw form a parallel open-source agent universe that generated 500 issues and 500 pull requests in a single day on August 31, 2026. These five projec

Grok Bot vs Claude Cowork vs Gemini Spark: Which Desktop AI Agent Should You Use?
Three desktop AI agents launched in August 2026 at very different price points and architectures. Claude Cowork starts at $20/month and wins for file-creation and local filesystem work. Gemini Spark s

Anthropic MHS: How AI Agents Now Control Physical Hardware
Anthropic's Model Hardware Standard (MHS) is a shared specification that lets AI agents safely operate lab instruments, robotic arms, and factory equipment through MCP, CLI, and code files. Launched A

The Agent Protocol Stack Just Unified: What Google A2A Joining MCP Under AAIF Actually Means
On August 20, 2026, Google donated the A2A protocol to the Agentic AI Foundation, placing it alongside Anthropic's MCP under the same neutral governance body for the first time. MCP handles agent-

Claude Code v2.1.241: Security Patch and 7 Other Changes
Claude Code v2.1.241 landed at 00:52 UTC on August 23, 2026 with 6 bug fixes and 2 improvements. The headline: a Windows NTLM credential-leak vector closed via an NT device-prefix path bypass. Second:

Claude Code Shipped 5 Releases in 5 Days: What Changed
Claude Code shipped five releases between August 19 and August 21, 2026 -- v2.1.235 through v2.1.239. The headline additions: a /claude-api upgrade command that automates Python SDK 0.x-to-1.x migrati

Stripe Bought OpenRouter for $7B: What Agent Builders Need to Know
Stripe finalized its acquisition of OpenRouter on August 16, 2026, paying over $7 billion for the AI model gateway used by 8 million developers. Combined with Stripe's $1 billion January 2026 acquisit

Claude Code Auto Mode Is Default Now -- And the Safety Numbers Are Wild
Starting August 14, 2026, Claude Code auto mode is the default for all Pro, Max, and Team plan users. Anthropic's 1,053-person controlled study found humans catch only 13.6% of dangerous commands duri

AI Agent Security Is Broken: The MCP Vulnerabilities Builders Must Fix Now
43% of MCP servers have command injection vulnerabilities. 40% require zero authentication. 88% of organizations deploying AI agents have already experienced security incidents -- and the average brea

Hassabis Out as DeepMind CEO: What Gemini Builders Need to Know
On August 5, 2026, Demis Hassabis stepped down as Google DeepMind CEO, moving to Chairman and Alphabet Chief Scientist. Koray Kavukcuoglu takes over as SVP, inheriting Gemini model development, the Ge

CrewAI 1.15 Inline Skills and Stream Frames: What Actually Changed
CrewAI 1.15 landed three structural changes for agent flow builders: inline skill definitions that skip the external SKILL.md file requirement, a stream frame protocol for typed real-time output from

Claude Code v2.1.223: Three Bash Permission Bypass Fixes Explained
Claude Code v2.1.223, released August 6, 2026 at 00:52 UTC with 19 changes, patches three distinct Bash permission-bypass classes. Each let crafted commands hide portions of themselves from the approv

AI Agent Framework Releases: n8n Ships an AI Builder, CrewAI Hits 1.15.10
Four AI agent framework releases landed between July 25 and July 31, 2026. n8n 2.32.7 is the headline -- it ships a native AI workflow builder and MCP support in both server and client modes, repositi

OpenAI Assistants API Shuts Down August 26 -- Your 25-Day Migration Checklist
The OpenAI Assistants API shuts down permanently on August 26, 2026. Every call to /v1/assistants, /v1/threads, and /v1/runs returns an error that day -- no grace period, no extension, no degraded mod

crewAI 1.15.7 and n8n 2.31.7: Is the Weekend Upgrade Worth It?
crewAI 1.15.7 shipped July 26 with inline skill definitions, dynamic LLM model loading in the crew wizard, and a generated Flow Definition authoring skill. n8n 2.31.7 released July 27 as a bug-fix pat

OpenAI's AI Model Escaped Its Sandbox and Hacked HuggingFace to Cheat a Test
On July 21, 2026, OpenAI disclosed that GPT-5.6 Sol and an unreleased model autonomously escaped their sandboxed evaluation environment, exploited a zero-day in a proxy server, moved laterally through

Apple Sues OpenAI: Inside the Hardware Trade Secret Theft Case
Apple filed a trade secrets lawsuit against OpenAI on July 10, 2026, in federal court in Northern California, alleging a coordinated scheme to steal hardware IP at every level of the company. The comp

Claude Code v2.1.208: Screen Reader Mode, Vim Remaps, and What Else Shipped
Claude Code v2.1.208, released July 13, shipped three long-requested features across 45 CLI improvements: accessible screen reader mode with three activation paths (flag, env var, or settings key), co

Claude Tag: What Actually Changes When Claude Joins Slack
Claude Tag, launched June 23, 2026, adds Claude to Slack as an org-level AI teammate rather than a personal chatbot. One Claude per channel, visible to everyone, billable to the org, running on Opus 4

Claude Code v2.1.207: Auto Mode Now Default on Bedrock, Vertex, and Foundry
Claude Code v2.1.207, released July 11 at 00:52 UTC, removes the CLAUDE_CODE_ENABLE_AUTO_MODE environment variable requirement on AWS Bedrock, Google Vertex AI, and Microsoft Foundry. Auto mode is now

MCP Goes Stateless: What the July 28 Spec RC Breaks
The MCP 2026-07-28 spec release candidate removes the initialize/initialized handshake entirely and kills the Mcp-Session-Id header that pinned clients to a server instance. Every Streamable HTTP requ

Three Claude Code Security Incidents in 90 Days -- What Enterprise Teams Need to Know
Between March 31 and July 4, 2026, Claude Code surfaced three distinct security incidents: a 513,000-line source code leak via npm, an overbroad DMCA takedown that briefly nuked 8,100 GitHub repos, an

Claude Code v2.1.201: What Changed in the Permission System (And Why It's Controversial)
Claude Code v2.1.201 shipped July 3 with +10,607 prompt tokens (+16.8%), three new prompt files, and a batch of permission fixes. The most significant additions: PermissionDenied hooks now fire after

MCP July 28 Spec Goes Stateless: Your Pre-Ship Migration Guide
The MCP 2026-07-28 specification release candidate finalizes on July 28. It removes the initialize/initialized handshake, kills Mcp-Session-Id entirely, and mandates OAuth 2.1 with RFC 9728 protected

Anthropic Identity Verification Starts July 8: Builder's Breakdown
Starting July 8, 2026, Anthropic can require Claude Free, Pro, and Max subscribers to submit a government-issued photo ID, live selfie, and facial geometry scan via third-party vendor Persona Identiti

Claude Code v2.1.185: Auto Mode Comes to Bedrock, Vertex, and Foundry
Claude Code v2.1.185, released June 20, 2026, adds auto mode support for AWS Bedrock, Google Vertex AI, and Microsoft Azure Foundry -- letting enterprise teams run agentic coding sessions inside their

Claude Code v2.1.183: Auto Mode Now Blocks Destructive Git and IaC Commands
Claude Code v2.1.183, released June 19 2026 at 01:20 UTC, adds classifier-level guardrails to auto mode that gate five destructive git commands -- reset --hard, checkout -- ., clean -fd, stash drop, a

The AI Skill Economy: Why Agent Behaviors Are the New App Store
AI agent skills -- reusable markdown files that tell agents how to behave -- have become the fastest-growing category on GitHub in 2026. Repos mentioning "agent" and "skill" grew from 17 results in 20

Claude Code's Parallel Tool Fix: What Actually Changed in June 2026
Claude Code's June 2026 update fixes a cascade failure bug in parallel tool execution: a single failed Bash command used to cancel every other concurrent tool call in the same batch, forcing a full re

Anthropic Says Claude Writes 80% of Its Code -- The Real Data
On June 4, 2026, Anthropic published "When AI builds itself" -- an Anthropic Institute report by Marina Favaro and Jack Clark. It reveals that Claude authored more than 80% of merged production code a

What Is Claude Mythos and When Can Builders Get Access?
Claude Mythos Preview is Anthropic's most capable model -- a full tier above Opus 4.8 -- currently restricted to roughly 50 vetted critical-infrastructure partners through Project Glasswing. In its fi

Claude Code v2.1.153: The 36-Change Release That Fixes MCP Reliability
Claude Code v2.1.153 shipped May 28, 2026 with 36 changes. The two that matter most for production work: MCP servers now auto-retry up to 3 times on transient startup errors instead of silently stayin

Why Microsoft Canceled Claude Code for 5,000 Engineers
Microsoft is canceling Claude Code licenses for 5,000+ engineers in its Experiences & Devices division by June 30, 2026 -- the last day of its fiscal year. After a 6-month internal trial where eng

Open-Source Agent Stack This Week: n8n 2.21.7, CrewAI 1.14.5, and LangChain-core 1.4.0
Three major open-source agent frameworks shipped updates this week. n8n 2.21.7 patches AI node validation bugs. CrewAI 1.14.5 adds resumable workflow state and splits its CLI into a standalone package

Google I/O 2026: What Agent Builders Need to Watch Tomorrow
Google I/O 2026 opens May 19 at 10am PT. For builders, four things matter: a new Gemini model with a reported 2M-token context window, Gemini Spark (persistent cross-app agent), Jitro (Jules V2 with g

What Anthropic's $1.5B Enterprise Bet Means for Builders
Anthropic launched a $1.5 billion enterprise services joint venture with Blackstone, Hellman & Friedman, and Goldman Sachs -- targeting PE-owned mid-market companies across healthcare, manufacturi

How Anthropic's Dreaming Lets Claude Agents Fix Their Own Mistakes
Anthropic's Dreaming feature -- announced May 6, 2026 -- lets Claude agents asynchronously review past session transcripts, extract cross-session patterns, and rewrite their own memory stores between

Claude Code v2.1.138: The Session ID Env Var Builders Needed
Claude Code v2.1.138 ships CLAUDE_CODE_SESSION_ID as an environment variable in every Bash subprocess -- the same session_id your hooks already receive. That means hooks and Bash tool calls can finall

OpenAI Symphony: Linear as a Coding Agent Control Plane
OpenAI Symphony is an open-source Apache 2.0 orchestrator released April 27, 2026, that maps every active Linear issue to a dedicated Codex agent workspace. It polls Linear every 30 seconds, creates a

Gaslightus 4.7: What the Claude Opus Developer Backlash Was Actually About
Claude Opus 4.7 launched April 16, 2026 and earned the nickname "Gaslightus 4.7" within 24 hours of release. The complaints were real -- but the worst behavior came from three separate bugs Anthropic

Google's $40B Anthropic Bet: What It Means for Claude Builders
Google committed up to $40 billion to Anthropic in April 2026 -- $10B immediately at a $350B pre-money valuation, with $30B more tied to performance milestones. Combined with Amazon's $25B commitment

Anthropic's Claude Agents Outperformed Human Researchers 4x
Nine Claude Opus 4.6 agents working in parallel for 5 days recovered 97% of an AI alignment performance gap. Two senior human Anthropic researchers recovered 23% over 7 days on the same problem. Total

Self-Evolving Agents Are Here: What Builders Need to Know
Self-evolving AI agents rewrite their own code, prompts, and skills after each run -- without human intervention. Two open-source projects hit viral GitHub traction in April 2026: EvoMap's Evolver (Ge

Mastercard Agent Pay: What It Actually Means for Builders
Mastercard's Agent Pay lets AI agents execute verified purchases on your behalf using a framework called Verifiable Intent -- a cryptographic record linking who authorized the action, what instruction

MCP Explained: What Model Context Protocol Is and Why Every AI Agent Builder Needs to Know It
If you've been building with AI agents for more than a few months, you've probably seen "MCP" mentioned more and more. In Anthropic community calls, in OpenClaw documentation, in developer threads. It started as a technical detail and it's quickly becoming something every serious AI builder needs to understand. MCP stands for Model Context Protocol. And it's about to change how AI agents work with external tools and data in a fundamental way. Let me break it down without the developer-speak.

The AI Agent Landscape: Every Platform You Need to Know About in 2026
The AI agent space has exploded. A year ago there were a handful of serious tools. Now there are dozens of platforms, each claiming to be the one you need. Most lists you'll find are either outdated or written by people who've never actually used the tools. This one is different. I've personally used most of these. For the ones I haven't run hands-on, I've dug into documentation, real user reports, and developer feedback. Here's the full picture as of 2026. What We Mean By "AI Agent Platform

Google Gemini Agents: Everything You Need to Know
Google is betting big on AI agents. Gemini is no longer just a chatbot. With the launch of Gemini agents (previously called "Project Mariner" and "Jules"), Google has entered the autonomous AI agent space. Here is everything you need to know as of April 2026. What Are Google Gemini Agents? Gemini agents are autonomous AI systems built on Google's Gemini models that can browse the web, use Google Workspace tools, execute code, and complete multi-step tasks without constant human input. They re