Tag
Threat actors turned Anthropic's accidental source code leak into a malware delivery pipeline within hours. Meanwhile, four unpatched CrewAI vulnerabilities let attackers chain prompt injection into full remote code execution.
Microsoft Threat Intelligence documents how state-backed hackers are bypassing LLM safety controls to generate exploit code, build phishing infrastructure, and automate entire attack chains.
A missing .npmignore entry exposed Claude Code's full TypeScript codebase, revealing hidden features, anti-distillation tricks, and an unreleased always-on daemon called Kairos.
OpenClaw went from one CVE to nine in four days, with 12% of its marketplace confirmed malicious. Plus: ChatGPT's patched DNS exfiltration flaw.
The fastest-growing open source project in GitHub history has become 2026's first major AI security disaster, with 135,000+ exposed instances, 9 CVEs in 4 days, and malware-laced skills.
The supply chain attackers behind Trivy are now wiping Iranian infrastructure, hiding malware in audio files, and extorting enterprises with help from LAPSUS$.
TeamPCP's supply chain campaign continues with a WAV file steganography attack on Telnyx. Meanwhile, three vulnerabilities in LangChain and LangGraph can leak files, secrets, and conversation histories.
A misconfigured CMS exposed 3,000 internal documents revealing Anthropic's most powerful model yet—one the company says could 'exploit vulnerabilities in ways that far outpace defenders.'
A calorie tracking app trusted with your weight, meals, and fitness goals was wide open. The Firebase misconfiguration let anyone read the entire database without credentials.
After 12% of ClawHub skills turned out to be malware and 135,000 instances were exposed, Cisco releases DefenseClaw and OpenClawd adds verified skill screening. The AI agent ecosystem is racing to catch up.
Georgia Tech researchers tracking real CVEs find 35 vulnerabilities from AI-generated code in March alone—and estimate the actual number is 10x higher.
135,000+ GitHub stars. Four critical CVEs. 12% of its marketplace poisoned with malware. OpenClaw's rise to fame came with a security crisis that every AI agent user needs to understand.
Anthropic's new auto mode lets Claude Code approve its own actions, using an ML classifier to block risky operations. It's convenient, but is it safe?
A coordinated supply chain campaign has compromised Trivy, LiteLLM, and dozens of npm packages. Meanwhile, Langflow attackers built working exploits within hours of disclosure.
Security scanners become attack vectors, AI agent platforms get RCE'd before patches exist, and 400+ GitHub repos fall to GlassWorm. Plus: a new secrets scanner built for AI coding agents.
The creator of Gitleaks releases a faster, more accurate successor with 98.6% recall and native AI agent integration. Here's why it matters.
Nature study shows large reasoning models can autonomously bypass safety guardrails across nine major AI systems. No human expertise required.
New research reveals multimodal LLMs are vulnerable to hidden instructions embedded in images. Mind maps, steganography, and physical signage all bypass text-based safety filters.
A single HTTP request can own your AI workflow server. CVE-2026-33017 shows why authentication shouldn't be optional.
IEEE S&P research finds 10,000+ websites running vulnerable AI chatbot plugins. Attackers can forge conversations, hijack tools, and extract system prompts.
The AI agent that couldn't stop getting hacked now has 4 critical and 52 high-severity flaws. Here's the latest wave of March 2026 CVEs.
A rogue AI agent triggers Sev 1 at Meta, agentic browsers leak passwords, and researchers prove AI can autonomously jailbreak other AI with 97% success.
The open-source AI agent with 135,000+ GitHub stars has become the center of 2026's first major AI security crisis
A self-propagating malware campaign steals developer credentials via malicious VS Code extensions, then force-pushes cryptocurrency-stealing code into legitimate Python projects.
