Latest cybersecurity reporting from selected sources.

Also, 23andMe Breach Settlement, GitHub AI Flaw, Ubiquiti Patches Critical BugsThis week, hidden breaches, researchers exposed a GitHub AI agent flaw, an Ubiquiti critical flaw, a ColdFusion flaw, a growing Chinese ORB, U.K. government FortiBleed exposure, 23andMe victims get $46.75 million, 3.8 million affected by Medtronic breach and Cerner's 2025 victim count went up.

Also, France Probes Tchap Breach, M&S Cancels Bonuses, June Patch TuesdayThis week, CISA tightened patching rules, hackers provoked AI scanners. An accused Russian intel hacker appeared in court. Microsoft warned of AI-themed attacks. M&S canceled bonuses. France probed a Tchap breach. NHS trusts disclosed stolen data and a Telegram campaign targeted Russian troops.

Hidden Install Settings Let Malicious MCP Links Execute CodeMicrosoft patched a high-severity flaw in Visual Studio Code after researchers found attackers could hide malicious settings inside MCP server install links, giving them persistent access to developer machines through what appeared to be routine artificial intelligence tool installations.

Chainable Bugs Enable Credential Theft, Persistence, TakeoverFour chainable flaws in OpenClaw allowed attackers to move from an initial foothold to persistent system-level compromise by abusing the AI agent's own privileges. The bugs enabled credential theft, privilege escalation and backdoor deployment, affecting all versions released before April 23.

Also, YellowKey Bypasses BitLocker, Škoda Breach, Kingdom Market Operator JailedThis week, U.S. lawmakers urged action on AI, a BitLocker exploit. Škoda, Nvidia’s GeForce NOW partner and telehealth firm OpenLoop reported breaches. Patch Tuesday. A dark market operator sentenced and pro-Ukraine and Iranian-linked hacking. Nitrogen ransomware attack on Foxconn.

Google Says Criminals Used AI to Discover and Code ExploitA cybercriminal group came close to launching a mass attack earlier this year, armed with a software exploit that an AI model had built from scratch, said Google researchers. Google said it worked with the affected vendor to patch the flaw before an attack could be launched.

AI-Assisted Offensive Security Researcher Discovered Flaw After 1 Hour of ScanningPatch all Linux kernels issued from 2017 onwards to fix a serious vulnerability in the kernel’s cryptography API that can be easily exploited by a local, unprivileged user to gain root-level access. The major flaw is the latest to be found by an AI-assisted researcher.

AI Tool Used to Discover Bugs, Which Included 2 Maximum Severity VulnerabilitiesResearchers at security firm AISLE said they recently identified 38 vulnerabilities, including two maximum-severity zero-day flaws in OpenEMR, an open-source electronic medical record software platform used by about 100,000 healthcare providers globally. OpenEMR has patched the problems.

But Expect Plenty of Bottlenecks in Coordination, Validation and Patch DeploymentAnthropic's Claude Mythos Preview shows how AI can discover and chain vulnerabilities at scale, but the bigger challenge for defenders is redesigning disclosure, triage and patching processes so fixes can be deployed safely before attackers exploit the gap.

Also, More ShinyHunters Breaches, North Korea Laptop Farm Operator SentencedThis week, Finland's Aleksanteri Kivimäki sentenced. ShinyHunters breaches. Laptop farm rancher sentenced. Oregon state agency hacker sentenced. African scammers arrested. MuddyWater AI-assisted hacks. Advantest ransomware incident, SolarWinds and Microsoft patches. FileZen flaw. QualDerm breach.

Flaws Let Attackers Run Commands and Steal API Keys Before Trust PromptCheck Point research found three critical flaws in Anthropic's Claude Code that allow attackers to execute arbitrary commands and steal API keys through repository configuration files, before users see a trust prompt. The AI giant has patched all three vulnerabilities.

Also: EU Bans AI Tools, Notepad++ Secures Updater, Apple Patches iOS Zero-DayThis week, Cambodia shuttered 200 scam centers. EU Parliament banned AI tools. Canada Goose disputed a ShinyHunters leak. Notepad++ patched an updater flaw. Apple fixed a decades-old iOS zero-day. BeyondTrust and Dell patched critical flaws under active exploitation.

Zero-Click Vulnerability Let Attackers Weaponize Enterprise AI AssistantGoogle patched a vulnerability in Gemini Enterprise that allowed attackers to steal corporate data through a shared document, calendar invitation or email without any user action or security alerts. No malware was executed, no credentials were phished and no data left through approved channels.

Attackers Could Hijack Developer Machines via Tampered Config FilesOpenAI patched a command injection flaw in its Codex CLI tool that let attackers run arbitrary commands on developer machines by hiding malicious configuration files inside code repositories. Hackers could turn ordinary repository files into execution vectors.

Researcher Found Bug Could Exfiltrate Secrets Via Camo ImagesA now-patched flaw in GitHub Copilot Chat could have allowed attackers to steal private source code and secrets by embedding hidden prompts that hijacked the artificial intelligence assistant's responses. The exploit also used the code hosting platform's image proxy to leak the stolen data.

Agentforce Agentic AI Tool Was Exposed to Indirect Prompt Injection AttacksSalesforce has patched a vulnerability involving its Agentforce agentic artificial intelligence tool, discovered by researchers, that attackers could have exploited using an indirect prompt injection attack to steal sensitive customer data and leads being stored in the CRM system.

Attackers Could Siphon Gmail Data Unnoticed From Users Who Let AI Tool Access EmailOpenAI patched a flaw in ChatGPT's Deep Research agent that could have enabled hackers to extract Gmail data without the user's knowledge. Radware researchers said the flaw affected subscribers who authorized the artificial intelligence tool to access their email accounts.

Cursor Patched Flaw Days After Disclosure, Says Check PointCheck Point researchers found a RCE flaw in Cursor, an AI-powered code editor, by manipulating a previously approved model context protocol configuration. Once a developer approved a configuration file for an MCP server, any future changes to that file could be executed without further prompts.

Always Secure MCP Servers Connecting LLMs to External Systems, Experts WarnWarning: Popular technology designed to make it easy for artificial intelligence tools to connect with external applications and data sources can be turned to malicious use. Researchers discovered two separate vulnerabilities tied to tools in the ecosystem around model context protocol, or MCP.

MCP Server Paused for Days After Bug Risked Data Leakage Between UsersAsana patched a vulnerability in an artificial intelligence integration feature that could have allowed users to view data from other organizations. The time management company paused the use of Asana Model Context Protocol for nearly two weeks to apply the fix.