Stories cluster related articles into a single narrative, linked to the underlying CVEs and affected products. 3,611 stories synthesized.
A critical remote code execution vulnerability (CVE-2026-27794) in LangChain's LangGraph BaseCache class allows unauthenticated attackers to achieve code execution via deserialization of untrusted data.
A newly disclosed vulnerability in Trend Micro Apex One (CVE-2025-71213) allows local attackers to escalate privileges to SYSTEM via an origin validation error in the NT Listener service.
A critical directory traversal vulnerability in Trend Micro Apex One Console (CVE-2025-71211, CVSS 9.8) allows unauthenticated remote attackers to execute arbitrary code via TCP ports 8080 and 4343.
A privilege escalation vulnerability in Trend Micro Apex Central, tracked as CVE-2025-71208, allows authenticated attackers to gain elevated access to protected resources.
WatchTowr Labs has disclosed three pre-authentication vulnerabilities in SolarWinds Web Help Desk, including an RCE via deserialization, that allow unauthenticated remote code execution on fully patched instances.
Trail of Bits has released mquire, an open-source Linux memory forensics tool that extracts BTF type info and Kallsyms symbol addresses directly from memory dumps, enabling analysis of unknown or custom kernels without external debug symbols.
A missing encryption protocol downgrade vulnerability in Ubiquiti Networks AI Pro cameras lets network-adjacent attackers weaken the discovery channel without authentication, earning a Pwn2Own prize.
A critical authentication bypass vulnerability in the Socomec DIRIS A-40 power monitoring device's HTTP API allows network-adjacent attackers to gain unauthorized access without credentials.
Docker has released a fix for CVE-2026-2664, an out-of-bounds read information disclosure vulnerability in the grpcfuse kernel module of Docker Desktop, which could be chained with other flaws for kernel-level code execution.
Ubiquiti Networks has patched a denial-of-service vulnerability in its AI Pro camera AI Pro devices, discovered during Pwn2Own Berlin 2026, that allows network-adjacent attackers to crash the system without authentication.
Siemens has released a security update for CVE-2026-25655, an uncontrolled search path element vulnerability in SINEC NMS that allows local attackers to escalate privileges to SYSTEM or LOCAL SERVICE.
A critical command injection vulnerability in claude-hovercraft's executeClaudeCode method, tracked as CVE-2025-15060, allows unauthenticated remote attackers to execute arbitrary code with a CVSS score of 9.9.8.
A critical directory traversal vulnerability in IceWarp collaboration software allows unauthenticated remote attackers to read arbitrary files as root, with patches available in versions 14.2.0.12 and 14.1.0.20.
A local privilege escalation vulnerability in Siemens SINEC NMS allows attackers with low-privileged attackers to gain SYSTEM access via an unsecured OpenSSL configuration file.
A cleartext transmission vulnerability in Ubiquiti Networks AI Pro cameras, disclosed as part of Pwn2Own, allows network-adjacent network-adjacent attackers to steal stored credentials without authentication.
GitLab released versions 18.9.1, 18.8.5, and 18.7.5 on February 25, 2026, patching multiple high-severity vulnerabilities including an XSS flaw in the Mermaid sandbox and several denial-of-service issues.
The DFIR Report details an intrusion where threat actors exploited CVE-2023-46604 in Apache ActiveMQ to deploy LockBit ransomware, returning 18 days after initial eviction.
A vulnerability in the beta version of Docker Desktop MCP Server beta allows local attackers with low-privileged code execution to read stored credentials from cleartext configuration files.
Trend Micro researchers uncovered a campaign using malicious OpenClaw skills to trick AI agents into installing a new variant of the Atomic macOS Stealer (AMOS), marking a shift from human-focused social engineering to AI agent manipulation.
Trail of Bits audited Perplexity's AI-powered Comet browser and demonstrated four prompt injection techniques that could extract private Gmail data from authenticated sessions.
ESET researchers have uncovered PromptSpy, the first Android malware to leverage generative AI (Google Gemini) for context-aware UI manipulation to achieve persistence and deploy a VNC module for remote device access.
A local privilege escalation vulnerability in PDF-XChange Editor's TrackerUpdate component, tracked as CVE-2026-2040, allows attackers to load malicious libraries and execute code at elevated privileges.
A critical out-of-bounds write vulnerability in GIMP's XWD file parsing (CVE-2026-2045) allows remote code execution when a user opens a crafted image, with a CVSS score of 7.8.
A high-severity out-of-bounds write vulnerability in GIMP's XWD file parsing, tracked as CVE-2026-2048, allows attackers to execute arbitrary code by tricking users into opening a crafted XWD image file.