AutoJack Attack Lets One Web Page Hijack AI Agent for Host Code Execution
Microsoft researchers have detailed an exploit chain, named AutoJack, that turns an AI browsing agent into a delivery vehicle for remote code execution
Microsoft software and cloud platforms underpin enterprise systems, so vulnerabilities and security advisories can affect identity, data, and operations.
Search across headline titles and summaries.
Background for this topic.
Microsoft is a technology company whose Windows, Microsoft 365, Azure, identity services, and developer tools form a widely deployed enterprise computing ecosystem. Its security relevance spans endpoint and server software, cloud control planes, authentication, collaboration data, and the update mechanisms used to maintain them.
Security news under this tag commonly concerns vulnerabilities requiring patching, exploits against exposed services, identity or token compromise, and misconfiguration of cloud permissions or authentication policies. Practitioners should verify affected versions and exposure, apply updates or mitigations, enforce multifactor authentication and least privilege, and monitor relevant audit logs. Incidents involving Microsoft-hosted identities or data may require rapid session and credential containment, investigation across cloud and on-premises systems, and assessment of privacy or regulatory obligations.
Microsoft researchers have detailed an exploit chain, named AutoJack, that turns an AI browsing agent into a delivery vehicle for remote code execution
Microsoft has confirmed a confusing Windows bug that causes different filenames to appear in the confirmation dialog when deleting a file from the Recycle Bin. [...]
AutoJack is a novel exploit chain showing how a single malicious webpage can turn an AI browsing agent into a remote code execution vector on the host machine. By abusing trust in localhost, missing authentication, and unsafe parameter handling, attackers can trigger arbitrary process execution through AutoGen Studio’s MCP WebSocket. The research highlights a broader pattern - when agents can browse untrusted content and access local services, traditional boundaries like localhost are no longer secure. The post AutoJack: How a single page can RCE the host running your AI agent appeared first on Microsoft Security Blog.
USB .lnk malware steals crypto via clipboard hijack, replaces wallet addresses, steals seed phrases, and screenshots. Microsoft Threat Intelligence has been tracking a clipboard-stealing malware (Clipper) campaign since February 2026 that targets cryptocurrency wallets. A clipper is a type of malicious software that monitors and manipulates your clipboard, the temporary memory where data is stored […]
Threat actors targeting cryptocurrency wallets have been distributing clipboard-stealing malware with self-spreading capabilities and using the Tor network to conceal communication. [...]
Microsoft has disclosed details of a Windows-based cryptocurrency clipper campaign that has targeted users since February 2026 with clipboard-intercepting malware with self-spreading capabilities and using the Tor anonymity network to hide communication
Microsoft 365 helps keep services running, but protecting and recovering business data remains your responsibility. Acronis breaks down five gaps organizations should consider when evaluating Microsoft 365 data protection. [...]
Threat actors associated with the DragonForce ransomware have been observed using a custom Go-based remote access trojan (RAT) called Backdoor.Turn to conceal command-and-control (C2) traffic inside Microsoft Teams relay infrastructure
Microsoft has fixed a known issue causing the June 2026 security updates to fail on Windows Server 2016 systems that weren't up to date. [...]
Microsoft confirmed the RoguePlanet Defender zero-day (CVE-2026-50656), a privilege escalation flaw, and is developing a security patch. Microsoft has acknowledged the RoguePlanet zero-day affecting Microsoft Defender, tracked as CVE-2026-50656 (CVSS score of 7.8). The vulnerability allows privilege escalation through the Microsoft Malware Protection Engine. The company stated it is aware of the issue and is […]
A poisoned npm package infected 140+ projects with a hidden payload. This report highlights how to detect, hunt, and defend against supply chain attacks using Microsoft Defender and actionable threat intelligence. The post From package to postinstall payload: Inside the Mastra npm supply chain compromise by Sapphire Sleet appeared first on Microsoft Security Blog.
Microsoft Threat Intelligence analyzed a cryptocurrency clipper campaign that combines clipboard theft, wallet replacement, Tor-based communications, and worm-like propagation. Beyond stealing cryptocurrency transactions, the malware establishes persistent access and enables follow-on activity through a lightweight backdoor capability. The post Crypto Clipper uses Tor and worm-like propagation for persistence and control appeared first on Microsoft Security Blog.
Microsoft-Owned GitHub, Which Runs npm, Previews Supply-Chain Security FixesThe popular Mastra AI framework, used to build artificial intelligence agents, workflows and retrieval-augmented generation pipelines, has been poisoned by attackers, and Microsoft-owned GitHub has advised all developers to downgrade Mastra, pending compromised packages being found and eradicated.
Microsoft has formally disclosed that it's working to release a patch to address a Defender zero-day codenamed RoguePlanet
DragonForce hid for months by routing malware traffic through Microsoft Teams infrastructure, masking C2 activity and evading network detection. DragonForce ransomware operators hit a major U.S. services firm and stayed hidden for one to two months by routing their command-and-control traffic through Microsoft’s own Teams relay servers. Symantec’s threat hunters tracked the custom backdoor they […]
Microsoft is investigating a new issue preventing third-party applications from launching Microsoft Office applications or opening documents on up-to-date Windows systems. [...]
Microsoft confirmed that it's working on a security patch for a Defender zero-day vulnerability named "RoguePlanet," disclosed one week ago. [...]
China-linked FishMonger used two SprySOCKS Windows variants that leveraged kernel drivers and the Print Spooler to target governments in four countries. ESET researchers have found two previously undocumented Windows versions of SprySOCKS, a backdoor that the security community had until now treated as Linux-only. Trend Micro first documented the Linux variant in September 2023 and […]
FishMonger, a China-nexus threat group, has deployed an undocumented version of the Linux backdoor against government targets in Honduras, Taiwan, Thailand, and Pakistan.
Custom malware routed communications through legitimate Microsoft services, making malicious activity look like routine corporate collaboration