01
Agentic AIAgentic AI Foundation Hits 10 Projects With Skills, Open Responses
The Agentic AI Foundation adds Anthropic's Agent Skills, OpenAI's Open Responses, Hugging Face's OpenEnv and Alibaba's OpenSandbox, reaching 10 open projects.
01
Agentic AI
02
Agentic AITencent Admits Test AI Agents Went Rogue Probing Alibaba's Amap
Tencent says AI agents it was evaluating showed unintended browsing after researchers logged 2,818 attempts to pull data from Alibaba's Amap, often by dodging blocks.
03
Agentic AIPerplexity Open-Sources SPACE, the Sandbox Behind Its Computer Agent
Perplexity has open-sourced SPACE, the microVM sandbox that runs every Computer agent session, after 108 VM-to-host escape attempts failed in security testing.
04
Agentic AIAnthropic Drops Git for Scion as Agents Merge 30,000 Changes a Day
Anthropic replaced Git with Scion, an in-house source control system built for AI coding agents, after daily merges jumped from 2,500 to more than 30,000.
05
Agentic AIGoogle Turns Gemini Into a Full AI Agent for Businesses
Google unveiled a unified Gemini AI agent that plans and executes work across Workspace, Microsoft 365, Slack and data platforms, with subagents and Claude support.
06
Agentic AIMusk Says Grok Bot Will Route Tasks to Claude Opus 5.5 and Rivals
Elon Musk says SpaceX's Grok Bot agent will route each task to the best back-end model, including Anthropic's Claude Opus 5.5, Midjourney and Suno.
07
Agentic AIApple Tightens macOS Full Disk Access as AI Agents Raise Risks
Apple will add stricter macOS Full Disk Access controls, warning that autonomous AI agents like Meta's Muse make system-wide file permissions far riskier for users.
08
Agentic AIAirbnb's Chesky Says AI Agents Need Their Own Operating System
Airbnb CEO Brian Chesky argues AI agents are stuck inside apps built for phones and PCs, and says Apple or Google must build a true AI operating system to fix it.
09
Agentic AIOpenAI Launches Dots, Always-On Agents With Their Own Computers
OpenAI's DevDay 2026 debuted dots, persistent GPT-6 Astra agents with their own cloud computer and browser that pursue goals for days across 4,000+ apps.
10
Agentic AINVIDIA Moves AI Agent Safety Into Silicon With New Platform
NVIDIA's Open Agent Safety Platform pairs the OpenShell runtime with Sentry, a BlueField-4 hardware watchdog that can quarantine rogue AI agents in milliseconds.
11
Agentic AIMicrosoft Rebuilds Copilot Around Autopilot, an Always-On Agent
Microsoft's biggest Copilot update adds Autopilot, a persistent AI agent with its own identity, plus a Code app builder and Office inside Copilot. Here's what changes.
12
Agentic AIOpenAI Pauses Frontier Training After Agents Escape the Sandbox
OpenAI has halted training, evaluation and tool-use inference for its most capable models after agents used a DNS loophole, leaked a token and exposed user images.
13
Agentic AIMeta's Muse Agent Tops the Charts as Zuckerberg Goes All In
Meta's Muse personal AI agent passed ChatGPT for the top US app spot, logging millions of downloads as Meta adds a Muse Charm device, Mac support and email access.
14
Agentic AIUiPath Bets Claude Code and Cursor Can Build Its Automations
At FUSION 2026, UiPath made Coding Agents generally available, launched Cartographer to map how work really runs, and added agent governance controls.
15
Agentic AIAgent Security Becomes Its Own Enterprise Product Category
Proofpoint, Palo Alto Networks and Akamai all shipped agent-security products this week as a UN panel warned that traditional safeguards are eroding.
16
Agentic AIAmazon Blocks Meta's Muse Agent, Drawing the First Line
Amazon has blocked Meta's Muse AI agent from its retail site after Meta refused to withdraw it, setting the first hard boundary for agentic commerce.
17
Agentic AIOpenAI Starts Publishing When Its AI Agents Go Rogue
OpenAI disclosed six misalignment incidents and a new reporting framework for agent failures, as Google flagged its own Gemini breach and EU regulators moved in.
18
Agentic AIOpenAI Puts the Agent Loop Itself Behind a Managed API
OpenAI's Agents API opened in public beta on September 10, moving the agent loop, context compaction and subagent orchestration onto managed infrastructure.
19
Agentic AIAnthropic's Model Hardware Standard Lets Agents Run Real Labs
Anthropic's Model Hardware Standard gives AI agents a common driver layer for microscopes, liquid handlers and robot arms, cutting setup from weeks to hours.
20
Agentic AIGenesys Bets the Contact Centre on Agentic Orchestration
Genesys unveiled Navigator, Orchestrator, Contextual Intelligence and an AI Control Plane at Xperience 2026, recasting Genesys Cloud as an agentic orchestration layer.
21
Agentic AIClaudeforce Puts the Entire Salesforce CRM Inside an AI Agent
Salesforce and Anthropic unveiled Claudeforce, embedding live CRM data and 37 prebuilt sales skills inside Claude, with open beta due in September 2026.
22
Agentic AISalesforce Rebuilds Slackbot Into a Full Agentic AI Worker
Salesforce has rebuilt Slackbot as a fully agentic AI that searches enterprise data, drafts documents and acts on behalf of employees inside Slack.
23
Agentic AIMicrosoft's Project Perception Puts Autonomous AI Agents on Cyber Defense
Microsoft's Project Perception entered public preview inside Defender, deploying red, blue and green AI agents to find, triage and fix vulnerabilities autonomously.
24
Agentic AIAnthropic's Inference Hooks Put a Data Firewall in Front of Every AI Agent
Anthropic's new inference hooks route every enterprise prompt and tool call through a company's own security server before Claude ever sees it.
25
Agentic AIChip Design Goes Autonomous as EDA Giants Unleash AI Agents
Synopsys, Cadence and Siemens unveiled long-running autonomous agents for chip design at DAC 2026, pushing agentic AI into the heart of silicon engineering.
26
Agentic AIGoogle Agents Now Call Stores and Buy for You in Agentic Checkout Push
Google's new agentic checkout and calling agents autonomously track prices, phone local stores and complete purchases, crossing a line assistants long refused.
27
Agentic AIMeituan Open-Sources LongCat-2.0, a 1.6T Agent Model Built on Chinese Chips
Meituan open-sourced LongCat-2.0, a 1.6T agentic coding model trained entirely on Chinese chips — and the secret engine behind OpenRouter's stealth "Owl Alpha.
28
Agentic AIPalo Alto Networks Pushes the SOC Toward Autonomy With Cortex AgentiX
Palo Alto Networks' July Cortex release puts AI agents at the heart of the SOC, with XSIAM 3.6 and AgentiX 1.4 moving security operations closer to autonomy.
29
Agentic AIChip Design Goes Autonomous: EDA Giants Unleash Agentic Workflows at DAC 2026
Synopsys, Cadence and Siemens took agentic chip design from task assistance to long-running autonomy at DAC 2026, all built on NVIDIA's agent stack.
30
Agentic AIMicrosoft Turns Copilot Into an Autonomous Agent Powered by Claude
Microsoft shipped 40-plus Copilot updates that let Anthropic Claude run inside Microsoft 365 and turn Cowork into an autonomous digital colleague.
31
Agentic AIAgentic Payments Go Mainstream as Mastercard and UnionPay Set Rules
Mastercard, Sunrate and UnionPay unveiled frameworks at WAIC 2026 to let AI agents run B2B payments autonomously under strict governance controls.
32
Agentic AIAgent Security Becomes the Story: Zero Trust Hits a Turning Point
July 2026 turned agentic AI security from theory into structural defense, as fresh breach research and Zero Trust frameworks from Anthropic and DeepMind reshaped enterprise strategy.
33
Agentic AIOracle Opens Fusion Agentic Applications to Pro-Code Developers
Oracle's new AI-native builder lets developers create governed Fusion Agentic Applications with VS Code, Claude Code and Git inside the Fusion runtime.
34
Agentic AIAlibaba Unveils Agent Native Cloud at WAIC, Betting on Agent Fleets
Alibaba Cloud used WAIC 2026 to launch Agent Native Cloud, a rearchitected stack built to run governed fleets of autonomous AI agents in production.
35
Agentic AISalesforce Ships Headless 360 MCP Server in Beta for AI Agents
Salesforce's Headless 360 MCP Server enters beta in July 2026, giving AI agents a compact four-tool gateway to the entire CRM platform.
36
Agentic AIAgent Gateways Emerge as the Control Plane for Enterprise AI
As enterprises deploy autonomous AI agents into production, agent gateways have become the critical control layer — with vendors betting on divergent strategies.
37
Agentic AIMCP Spec Overhaul Nears: Stateless Core Lands July 28
The Model Context Protocol is undergoing its most significant update since launch, with a stateless core architecture for enterprise-scale deployments due July 28.