697 matches found
Seclens: Role-Specific Evaluation of LLM'S for Security Vulnerablity Detection
Existing benchmarks for LLM-based vulnerability detection compress model performance into a single metric, which fails to reflect the distinct priorities of different stakeholders. For example, a CISO may emphasize high recall of critical vulnerabilities, an engineering leader may prioritize...
Automated Malware Family Classification Using Weighted Hierarchical Ensembles of Large Language Models
Malware family classification remains a challenging task in automated malware analysis, particularly in real-world settings characterized by obfuscation, packing, and rapidly evolving threats. Existing machine learning and deep learning approaches typically depend on labeled datasets, handcrafted...
Synthetic Trust Attacks: Modeling How Generative AI Manipulates Human Decisions in Social Engineering Fraud
Imagine receiving a video call from your CFO, surrounded by colleagues, asking you to urgently authorise a confidential transfer. You comply. Every person on that call was fake, and you just lost $25 million. This is not a hypothetical. It happened in Hong Kong in January 2024, and it is becoming...
Reliance on Untrusted Inputs in a Security Decision
Overview openclaw is a 🦞 OpenClaw — Personal AI Assistant Affected versions of this package are vulnerable to Reliance on Untrusted Inputs in a Security Decision via the chat.send process. An attacker can inject unauthorized provenance fields by spoofing client identity metadata during the...
When Labels Are Scarce: A Systematic Mapping of Label-Efficient Code Vulnerability Detection
Machine-learning-based code vulnerability detection CVD has progressed rapidly, from deep program representations to pretrained code models and LLM-centered pipelines. Yet dependable vulnerability labeling remains expensive, noisy, and uneven across projects, languages, and CWE types, motivating...
Reliance on Untrusted Inputs in a Security Decision
Overview openclaw is a 🦞 OpenClaw — Personal AI Assistant Affected versions of this package are vulnerable to Reliance on Untrusted Inputs in a Security Decision in the trusted-proxy Control UI session handling process. An attacker can retain privileged scopes without device identity by accessing...
Security Bulletin: IBM Operational Decision Manager for March 2026 - Multiple CVEs addressed
Summary IBM Operational Decision Manager is vulnerable to multiple remote code execution and denial of service attacks in third party and open source used in the product for various functions. See full list below. The vulnerabilities have been addressed. Vulnerability Details CVEID:CVE-2025-41254...
Reliance on Untrusted Inputs in a Security Decision
Overview Affected versions of this package are vulnerable to Reliance on Untrusted Inputs in a Security Decision through spoofed headers in the Rate-Limit process. An attacker can circumvent request throttling by manipulating HTTP headers to appear as different users or clients, potentially...
Model Context Protocol Threat Modeling and Analyzing Vulnerabilities to Prompt Injection with Tool Poisoning
The Model Context Protocol MCP has rapidly emerged as a universal standard for connecting AI assistants to external tools and data sources. While MCP simplifies integration between AI applications and various services, it introduces significant security vulnerabilities, particularly on the client...
The Danger Behind Meta Killing End-to-End Encryption for Instagram DMs
Meta blamed users for not opting into the privacy-protecting feature. Experts fear the move could be the first major domino to fall for end-to-end encryption tech worldwide...
CVE-2026-32811
Heimdall is a cloud native Identity Aware Proxy and Access Control Decision service. When using Heimdall in envoy gRPC decision API mode with versions 0.7.0-alpha through 0.17.10, wrong encoding of the query URL string allows rules with non-wildcard path expressions to be bypassed. Envoy splits t...
NASimJax: GPU-Accelerated Policy Learning Framework for Penetration Testing
Penetration testing, the practice of simulating cyberattacks to identify vulnerabilities, is a complex sequential decision-making task that is inherently partially observable and features large action spaces. Training reinforcement learning RL policies for this domain faces a fundamental...
CVE-2026-23259
In the Linux kernel, the following vulnerability has been resolved: iouring/rw: free potentially allocated iovec on cache put failure If a read/write request goes through ioreqrwcleanup and has an allocated iovec attached and fails to put to the rwcache, then it may end up with an unaccounted iov...
Security Bulletin: Optional Mongo DB images in IBM Automation Decision Services 24.0.x are affected by CVE-2025-14847
Summary CVE-2025-14847 has been reported for the Mongo DB images shipped with IBM Automation Decision Services 24.0.0 and 24.0.1.. An updated version of the image is available. Vulnerability Details CVEID:CVE-2025-14847 DESCRIPTION: Mismatched length fields in Zlib compressed protocol headers may...
Exploit for Reliance on Untrusted Inputs in a Security Decision in Microsoft
CVE-2026...
nah Claude Code Tool
Claude Code's permission system is allow-or-deny per tool, but that doesn't really scale. Deleting some files is fine sometimes. And git checkout is sometimes catastrophic. Even when you curate permissions, 200 IQ Opus can find a way around it. Maintaining a deny list is a fool’s errand. What's...
CVE-2026-32126
OpenEMR is a free and open source electronic health records and medical practice management application. Prior to 8.0.0.1, an inverted boolean condition in ControllerRouter::route causes the admin/super ACL check to be enforced only for controllers that already have their own internal authorizati...
Building a High-Impact Tier 1: The 3 Steps CISOs Must Follow
Every CISO knows the uncomfortable truth about their Security Operations Center: the people most responsible for catching threats in real time are the people with the least experience. Tier 1 analysts sit at the front line of detection, and yet they are also the most vulnerable to the cognitive a...
