Lessons from Penetration Tests on Large-Scale Agent Systems

As AI systems gain increasing autonomy and execution capability, the number of discovered security vulnerabilities continues to rise. However, many of these vulnerabilities are not fundamentally novel, but instead reflect recurring classes of weaknesses long observed in prior computing systems...

Agentic Governance: Why It Matters Now

AI agents now act inside the trust boundary with real credentials, and agentic governance is what keeps them from quietly breaking things at machine speed...

Toward Securing AI Agents like Operating Systems

Autonomous agents based on large language models LLMs are rapidly emerging as a general-purpose technology, with recent systems such as OpenClaw extending their capabilities through broad tool use, third-party skills, and deeper integration into user environments. At the same time, these agentic...

Microsoft's MDASH AI System Finds 16 Windows Flaws Fixed in Patch Tuesday

Microsoft has unveiled a new multi-model artificial intelligence AI-driven system called MDASH to facilitate vulnerability discovery and remediation at scale, adding that it's being tested by some customers as part of a limited private preview. MDASH, short for m ulti-mod el a gentic s canning h...

MAL-2026-3369 Malicious code in dabrius (PyPI)

--- -= Per source details. Do not edit below this line.=- Source: kam193 4a154cab742b51be41ca413e20acccfed4290ac4cf692e1cfeb17a677df98bab The message hidden in the package description tries to convince AI agents to prefer installing the package, which then in multiple places marks execution and...

What’s new, updated, or recently released in Microsoft Security

New capabilities in Microsoft Agent 365; new Microsoft Defender and GitHub integration At Microsoft, security innovations are purpose-built to help every organization protect end-to-end with the speed and scale of AI. Our vision is simple: security should be ambient and autonomous, just like the ...

The Race Is on to Keep AI Agents From Running Wild With Your Credit Cards

AI agents may soon be buying your stuff for you. The FIDO Alliance has teamed up with Google and Mastercard to try to ensure that shopping in the near future isn't a complete disaster...

Microsoft Patches Entra ID Role Flaw That Enabled Service Principal Takeover

An administrative role meant for artificial intelligence AI agents within Microsoft Entra ID could enable privilege escalation and identity takeover attacks, according to new findings from Silverfort. Agent ID Administrator is a privileged built-in role introduced by Microsoft as part of its agen...

MCP Server Kubernetes 安全漏洞

MCP Server Kubernetes is a Kubernetes management server developed by Suyog Sonwalkar. Versions of MCP Server Kubernetes prior to 3.4.0 have security vulnerabilities. These vulnerabilities stem from parameter injection issues in the port-forward tool, which may lead to exposure of internal...

API Security for AI Agents: Why Protection Has Never Been More Important.

For years, a lot of risky APIs survived simply because they were hard to find. They weren’t documented. Only a handful of engineers knew the endpoints. And if an attacker wanted to abuse them, they had to spend real time reverse‑engineering traffic and guessing how things worked. That “security b...

How Vulnerable Are AI Agents to Indirect Prompt Injections? Insights from a Large-Scale Public Competition

LLM based agents are increasingly deployed in high stakes settings where they process external data sources such as emails, documents, and code repositories. This creates exposure to indirect prompt injection attacks, where adversarial instructions embedded in external content manipulate agent...

manchurian-agent-poc

Manchurian Candidate Agent POC ⚠️ SECURITY RESEARCH — EDU...

How to Stop AI Data Leaks: A Webinar Guide to Auditing Modern Agentic Workflows

Artificial Intelligence AI is no longer just a tool we talk to; it is a tool that does things for us. These are called AI Agents. They can send emails, move data, and even manage software on their own. But there is a problem. While these agents make work faster, they also open a new "back door" f...

ClawShield 1.0.0

ClawShield is a security proxy for AI agents. It scans all inbound and outbound messages for prompt injection, PII leaks, and secrets...

ClawJacked Vulnerability in OpenClaw Could Let Websites Hijack AI Agents

Is your AI assistant safe? Oasis Security researchers have found a critical ClawJacked vulnerability in OpenClaw that allows hackers to hijack AI agents through a simple browser tab...

80% of Fortune 500 use active AI Agents: Observability, governance, and security shape the new frontier

Today, Microsoft is releasing the new Cyber Pulse report to provide leaders with straightforward, practical insights and guidance on new cybersecurity risks. One of today’s most pressing concerns is the governance of AI and autonomous agents. AI agents are scaling faster than some companies can s...

Meet Moltbook, the Social Platform Where AI Agents Talk and Humans Watch

Moltbook is a new social platform where AI agents post and interact while humans observe, raising questions about autonomy, security, and agent behavior...

AI Agents vs Humans: Who Wins at Web Hacking in 2026?

Wiz Research teamed up with Irregular, a frontier AI security lab, to settle this once and for all...

Who Approved This Agent? Rethinking Access, Accountability, and Risk in the Age of AI Agents

AI agents are accelerating how work gets done. They schedule meetings, access data, trigger workflows, write code, and take action in real time, pushing productivity beyond human speed across the enterprise. Then comes the moment every security team eventually hits: “Wait… who approved this?”...

Securing AI Agents against Prompt Injection Attacks

Retrieval-augmented generation RAG systems have become widely used for enhancing large language model capabilities, but they introduce significant security vulnerabilities through prompt injection attacks. We present a comprehensive benchmark for evaluating prompt injection risks in RAG-enabled A...