New Relic unveiled a no-code agentic platform on Tuesday that lets enterprises deploy data observability AI agents. The platform supports model context protocol (MCP) and integrates with other New Relic tools to monitor company data and catch bugs before they disrupt products.

Source: TechCrunch

On February 17, 2026, NIST announced the launch of the AI Agent Standards Initiative. The initiative aims to ensure next-generation AI agents can function securely, interoperate smoothly across digital ecosystems, and be adopted with confidence while fostering industry-led standards.

Source: NIST

Between January and early February 2026, eight new high-to-critical vulnerabilities were disclosed in n8n across expression evaluation, file access controls, Git, SSH, and Python execution. The vulnerabilities undermine assumptions about isolation and trust in AI agent environments, requiring upgrade to n8n 2.5.2 minimum.

Source: Geordie AI

Microsoft released patches for AI vulnerabilities affecting GitHub Copilot, VS Code, and JetBrains products with CVEs CVE-2026-21516, CVE-2026-21523, and CVE-2026-21256. The vulnerabilities stem from command injection flaws triggered through prompt injection, allowing AI agents to execute malicious code.

Source: Krebs on Security

On February 5, 2026, Anthropic's research documented that Claude Opus 4.6 found and validated more than 500 high-severity vulnerabilities in production open-source software. Anthropic also introduced Claude Code Security for automated vulnerability scanning with human approval requirements.

Source: Futurum Group

GPT-5.3-Codex was released with markedly higher performance on coding benchmarks than earlier generations, but OpenAI is rolling out the model with unusually tight controls due to serious cybersecurity concerns. This is OpenAI's first model to hit 'high' for cybersecurity on their preparedness framework.

Source: Fortune

OpenAI announced GPT-5.3-Codex-Spark, a lightweight version of its agentic coding tool powered by Cerebras' Wafer Scale Engine 3. The new model is designed for swift, real-time collaboration and rapid prototyping, serving as a daily productivity driver.

Source: TechCrunch

GPT-5.3, codenamed 'Garlic,' is reportedly being tested and expected to release February 26, 2026. The model allegedly scores 83.7% on SimpleBench reasoning tests, clearing the human baseline.

Source: PiunikaWeb

GPT-5.2 achieved a verified theoretical physics breakthrough by discovering a new gluon amplitude result. The model exposed a hidden mathematical pattern inside quantum field theory calculations, finding that amplitude is not universally zero under specific kinematic alignment conditions.

Source: Toolient

Cisco disclosed a critical authentication bypass vulnerability (CVSS 10.0) in SD-WAN Controller and Manager that has been exploited since 2023. CISA issued emergency directive requiring federal agencies to patch by February 27, 2026, and added the CVE to its Known Exploited Vulnerabilities catalog.

Source: The Hacker News

Microsoft released security updates for 58 vulnerabilities including 6 actively exploited zero-days (CVE-2026-21510, CVE-2026-21513, CVE-2026-21514, CVE-2026-21519, CVE-2026-21525, CVE-2026-21533) and 5 Critical vulnerabilities. Three zero-days were publicly disclosed prior to patching.

Source: BleepingComputer

Google fixed CVE-2026-2441, a high-severity use-after-free vulnerability in Chrome's CSS engine discovered on February 11, 2026. The flaw allows remote code execution via crafted HTML pages and is being actively exploited in the wild.

Source: The Hacker News

A critical vulnerability (CVSS 9.4) in n8n workflow automation tool bypasses security fixes from CVE-2025-68613. The flaw allows attackers to execute system commands through publicly accessible webhooks with malicious JavaScript destructuring syntax.

Source: The Hacker News

Cisco disclosed CVE-2026-20127 on February 25, 2026, a critical authentication bypass in SD-WAN systems exploited by threat actors since 2023. Australian authorities first identified the attacks, leading to CISA's emergency directive with February 27, 2026 patching deadline.

Source: Rapid7

Cybersecurity researchers have disclosed multiple security vulnerabilities in four popular VS Code extensions (Live Server, Code Runner, Markdown Preview Enhanced, and Microsoft Live Preview) that collectively have 125 million installs and could allow threat actors to steal local files and execute code remotely. The research demonstrates that a hacker needs only one malicious extension or vulnerability to perform lateral movement and compromise entire organizations.

Source: The Hacker News

VS Code February 2026 Insiders adds native browser integration in chat for AI agents to interact with page elements and capture screenshots, plus Claude Agent support for MCP (Model Context Protocol) servers with automatic pickup of servers installed via VS Code or Claude CLI. Users can now fork chat conversations to explore alternative approaches while preserving the original thread using the /fork command.

Source: Microsoft Visual Studio Code

Visual Studio 2026 updates released on February 10, 18, and 24, 2026, including fixes for compiler optimization causing access violations and MSVC optimizer bugs generating bad assembly code. Visual Studio now alerts users if it detects digital certificate problems during network calls and provides details about certificate errors with options to cancel or continue connections.

Source: Microsoft Learn

AWS announced Claude Sonnet 4.6 model in Amazon Bedrock offering frontier performance at lower cost, new Amazon EC2 Hpc8a instances powered by 5th Gen AMD EPYC processors delivering up to 40% higher performance, and open-source Agent Plugins for AWS that extend coding agents with deployment skills.

Source: AWS News Blog

Amazon announced a record-breaking $200 billion capital expenditure plan for 2026, with AWS expected to achieve 30% year-over-year growth driven by conversion of a $244 billion cloud backlog and enterprise AI deployments on custom Trainium3 chips.

Source: FinancialContent - MarketMinute

AWS expanded availability of new Amazon EC2 R8i and R8i-flex instances to additional regions, delivering up to 30% faster performance for PostgreSQL databases, up to 60% faster for NGINX web applications, and up to 40% faster for AI deep learning recommendation models compared to R7i instances.

Source: AWS

AWS announced an investment of up to $50 billion to build AI and supercomputing infrastructure for U.S. government agencies starting in 2026, adding nearly 1.3 gigawatts of capacity across AWS Top Secret, Secret and GovCloud regions with advanced compute and networking technologies.

Source: Fox Business

Published on February 12, 2026, Vercel announced new AI agent integrations for Next.js. The update includes experimental in-browser agent capabilities, MCP (Model Context Protocol) integration, and improved logging specifically designed to work with AI coding agents as first-class users.

Source: Next.js Official Blog

Cloudflare's detailed technical breakdown of their 'Vinext' project, which began February 13, 2026. By day three, the AI-generated Next.js alternative was deploying apps to Cloudflare Workers with full client hydration, achieving 94% API coverage and significantly improved performance benchmarks against Next.js 16.

Source: Cloudflare Blog

On January 26, 2026, a new denial-of-service vulnerability (CVE-2026-23864, CVSS 7.5) was disclosed affecting React Server Components used by Next.js. This follows previous critical security issues and requires immediate patching of affected Next.js applications using App Router.

Source: Netlify