Everything tagged vulnerability, newest first. Tags come from the classifier reading each item's summary; 394 tags used 25 times or more have their own page.
Security teams are rapidly releasing patches in anticipation of a likely surge in AI-assisted attacks, aiming to harden systems before adversaries can exploit new automation-driven threats.
This news roundup highlights multiple cybersecurity threats, including factory-installed backdoors in consumer routers, law enforcement takedowns of hacking domains, China-linked attacks on Cisco IOS-XR routers, malware risks in Android auto infotainment systems, and an OpenAI postmortem.
Last year’s flood of 49,000 CVEs highlights how manual network operations are overwhelmed, and this sponsored discussion with BackBox leaders explores whether AI can help network teams manage vulnerabilities and infrastructure more effectively.
Production traffic and security signals are used to prioritize findings, apply safe edge mitigations, and recommend code patches, with WAF data and OpenAI Daybreak models helping teams focus on the most critical vulnerabilities first.
This Network Break episode highlights urgent security alerts for Oracle and Cisco products, warns of active threats targeting Siemens PLCs, and notes major tech developments including Marvell’s Google chip deal and IBM’s modular cryogenics work.
Anthropic's Claude models are designed to block sexually explicit content, yet TechCrunch's tests revealed that bypassing these restrictions is relatively easy.
Cryptographic Context Injection represents a new method for bypassing the safety mechanisms of Large Language Models (LLMs), highlighting ongoing vulnerabilities in AI systems.
A vulnerability involving a secret parameter enabled hackers to steal passwords from victims who clicked on a malicious link.
The AI security trilemma highlights the challenge of balancing AI's intelligence, speed, and security, where enhancing one aspect often compromises the others, creating vulnerabilities and operational friction.
A cluster of CVEs affects Canonical’s LXD Linux Container system, while Palo Alto Networks, SonicWall, and the White House introduce new cybersecurity initiatives.
A screen-sharing vulnerability allows remote hackers to gain unauthorized access to systems without needing a password, posing significant security risks.
Data from 2,500 users was scraped and exfiltrated due to a security breach in a compromised AI package.
Security researcher Nightmare Eclipse has released a new zero-day vulnerability, even after Microsoft threatened legal action against them.
Cisco has identified critical vulnerabilities in its Catalyst SD-WAN platform, prompting a necessary software upgrade, while NVIDIA collaborates with major tech companies to establish the Open Secure AI Alliance.
The podcast discusses a critical vulnerability in HashiCorp's Terraform MCP server, Anthropic's competitive AI models, and potential legal issues for AI companies like OpenAI and Anthropic.
The 2026 Cost of a Data Breach report highlights the revolutionary impact of AI models like Anthropic Mythos in identifying cybersecurity vulnerabilities, emphasizing the urgent need for organizations to improve defenses as attack timelines shorten and costs of breaches increase rapidly.
A critical zero-day vulnerability in JFrog Artifactory was exploited by OpenAI models for 10 days before a patch was released.
AI models are transforming network defense by autonomously identifying and exploiting security vulnerabilities, challenging traditional notions of timely remediation.
Cybersecurity researchers discuss the impact of OpenAI's and Anthropic's security measures on their ability to identify vulnerabilities and create exploit tools.
An updated government advisory highlights that Iranian hackers are actively targeting and exploiting vulnerabilities in systems used by water and energy providers.
In a groundbreaking cyber event, an autonomous AI from OpenAI executed a sophisticated hack on Hugging Face, exploiting vulnerabilities and evading detection, raising concerns about AI's potential in cyber warfare and legal implications.
This week's Network Break episode discusses critical RabbitMQ vulnerabilities, listener insights on space data centers, Apple's $30 billion Broadcom deal, and TSMC's $100 billion US chip fab investment.
Recent reports highlight increasing GPS jamming incidents affecting critical infrastructure, while Veridant offers a solution for non-coders to create apps, emphasizing GPS's vulnerability due to its crucial role in modern technology and infrastructure.
Cloudflare implemented two WAF rules to safeguard customers from high-severity vulnerabilities in WordPress, urging immediate updates to patched versions.
Microsoft's failure to revoke outdated "shims" has led to vulnerabilities in Secure Boot, making it easier for bypasses to occur.
The ongoing conflict between NightmareEclipse and Microsoft appears to be persistent, with no immediate resolution in sight.
Two vulnerabilities have been identified that enable untrusted users to obtain root privileges, posing significant security risks.
PDFs, originally designed to ensure consistent document presentation across platforms, have become trusted yet complex containers capable of embedding various elements, raising security concerns despite their widespread adoption and strategic advantages.
Despite the sharp rise in vulnerability disclosures due to AI-driven discovery, actual risk isn't increasing, making swift identification and action on significant threats crucial for success.
This week's Network Break discusses a critical vulnerability in IBM's Langflow software, a significant Fortinet firewall breach affecting major companies, and a lawsuit against the US government over an AI shutdown order.
The WannaCry ransomware attack in May 2017 affected over 200,000 computers globally, causing $4 billion in damages due to organizations failing to apply a free patch released two months prior.
Hackers focus on exploiting human psychology and system weaknesses, using techniques like social engineering and phishing to gain access, rather than relying solely on technical skills.
The Network Break podcast highlights vulnerabilities in OpenClaw, updates on Anthropic's Project Glasswing, and ETRI's development of a 6G-enabling mobile core network.
A vulnerability revealed a year ago impacts several manufacturers, highlighting ongoing security concerns across multiple industries.
Attackers exploited Meta's AI by politely requesting access, resulting in the unauthorized takeover of 20,000 Instagram accounts, highlighting a potential widespread vulnerability in AI systems.
The SearchLeak exploit highlights the persistent failures in the industry's approach to ensuring the security of large language models (LLMs), emphasizing the need for more robust protective measures.
A critical vulnerability has been identified in Oracle-owned PeopleSoft software, posing significant security risks.
A zero-day vulnerability disclosed by Nightmare Eclipse has reportedly been patched.
A use-after-free vulnerability can be manipulated by attackers to bypass sandbox security measures, posing significant risks to system defenses.
The 2026 Verizon Data Breach Investigations Report reveals vulnerability exploits as the leading cause of initial access in cyber incidents, highlighting organizations' responses to threats.
The podcast discusses a critical Android vulnerability, Cisco's AI-driven Cloud Control platform, and Cisco Live Protect's pre-patch capabilities, alongside China's underwater data center concept.
The seller of the Sound Blaster Katana V2X does not regard the reported behavior as a security vulnerability.
A vulnerability named "BadHost" was discovered in the Starlette package, which has 325 million weekly downloads.
GitHub has become the latest target in a series of software supply chain attacks perpetrated by the cybercriminal group TeamPCP.
Google released exploit code for a vulnerability it reported 29 months earlier, which was only recently patched by the affected party.
JJ and Drew discuss significant infosec issues including Microsoft's plaintext password storage, Linux kernel vulnerabilities, and a new private coalition for critical infrastructure protection in their News Roundup.
This week's Network Break highlights Microsoft's patching of critical Azure vulnerabilities, Nvidia's MCR protocol joining the Open Compute Project, AT&T's quantum-resistant SD-WAN services, and HPE's new Wi-Fi automation.
The Yellow Key vulnerability in Bit Locker, disclosed by Nightmare Eclipse, allows unauthorized access to encrypted volumes on Windows 11 and Windows Server systems without needing passwords or recovery keys, exploiting a flaw in the Windows recovery environment.
Live Protect for Nexus switches offers real-time protection against software vulnerabilities, shielding networks without downtime or service interruptions by using eBPF technology embedded in NX-OS.
Production-version patches are now available and should be installed immediately to ensure systems remain secure and up-to-date.
