Security Threat Modeling for Emerging AI-Agent Protocols: A Comparative Analysis of MCP, A2A, Agora, and ANP

The rapid development of the AI agent communication protocols, including the Model Context Protocol MCP, Agent2Agent A2A, Agora, and Agent Network Protocol ANP, is reshaping how AI agents communicate with tools, services, and each other. While these protocols support scalable multi-agent...

PT-2026-7728

Name of the Vulnerable Software and Affected Versions Keras versions 3.0.0 through 3.13.1 Description A flaw exists in the model loading mechanism, specifically within the HDF5 integration of Keras. This issue allows a remote attacker to read local files and potentially disclose sensitive...

Agentic Knowledge Distillation: Autonomous Training of Small Language Models for SMS Threat Detection

SMS-based phishing smishing attacks have surged, yet training effective on-device detectors requires labelled threat data that quickly becomes outdated. To deal with this issue, we present Agentic Knowledge Distillation, which consists of a powerful LLM acts as an autonomous teacher that fine-tun...

VulReaD: Knowledge-Graph-Guided Software Vulnerability Reasoning and Detection

Software vulnerability detection SVD is a critical challenge in modern systems. Large language models LLMs offer natural-language explanations alongside predictions, but most work focuses on binary evaluation, and explanations often lack semantic consistency with Common Weakness Enumeration CWE...

Jailbreaking Leaves a Trace: Understanding and Detecting Jailbreak Attacks from Internal Representations of Large Language Models

Jailbreaking large language models LLMs has emerged as a critical security challenge with the widespread deployment of conversational AI systems. Adversarial users exploit these models through carefully crafted prompts to elicit restricted or unsafe outputs, a phenomenon commonly referred to as...

Security Assessment of Intel TDX with Support for Live Migration

In the second and third quarters of 2025, Google collaborated with Intel to conduct a security assessment of Intel Trust Domain Extensions TDX, extending Google's previous review and covering major changes since Intel TDX Module 1.0 - namely support for Live Migration and Trusted Domain TD...

Keras 安全漏洞

Keras is an open-source deep learning framework with multiple backends. Versions of Keras 3.13.1 and earlier contain security vulnerabilities. These vulnerabilities stem from defects in the model loading mechanism HDF5 integration, which could allow remote attackers to read local files through...

Vulnerabilities in Partial TEE-Shielded LLM Inference with Precomputed Noise

The deployment of large language models LLMs on third-party devices requires new ways to protect model intellectual property. While Trusted Execution Environments TEEs offer a promising solution, their performance limits can lead to a critical compromise: using a precomputed, static secret basis ...

Arbitrary File Read via Prompt Tag Source Validation Bypass in CreateModelVersion

The createmodelversion handler in mlflow/server/handlers.py uses a client-controlled tag to decide whether to skip source path validation. When a CreateModelVersion request includes the tag mlflow.prompt.isprompt, the helper ispromptrequest returns True, and the entire source validation block...

Measuring AI Security: Separating Signal from Panic

The conversation around AI security is full of anxiety. Every week, new headlines warn of jailbreaks, prompt injection, agents gone rogue, and the rise of LLM-enabled cybercrime. It’s easy to come away with the impression that AI is fundamentally uncontrollable and dangerous, and therefore...

Discord will limit profiles to teen-appropriate mode until you verify your age

Discord announced it will put all existing and new profiles in teen-appropriate mode by default in early March. The teen-appropriate profile mode will remain in place until users prove they are adults. To change a profile to “full access” will require verification by Discord’s age inference model...

MINI-WR2X-PRW2-WR36

Bulletin has no description...

When Handshakes Tell the Truth: Detecting Web Bad Bots Via TLS Fingerprints

Automated traffic continued to surpass human-generated traffic on the web, and a rising proportion of this automation was explicitly malicious. Evasive bots could pretend to be real users, even solve Captchas and mimic human interaction patterns. This work explores a less intrusive, protocol-leve...

LLM-FS: Zero-Shot Feature Selection for Effective and Interpretable Malware Detection

Feature selection FS remains essential for building accurate and interpretable detection models, particularly in high-dimensional malware datasets. Conventional FS methods such as Extra Trees, Variance Threshold, Tree-based models, Chi-Squared tests, ANOVA, Random Selection, and Sequential...

SecCodePRM: A Process Reward Model for Code Security

Large Language Models are rapidly becoming core components of modern software development workflows, yet ensuring code security remains challenging. Existing vulnerability detection pipelines either rely on static analyzers or use LLM/GNN-based detectors trained with coarse program-level...

Rethinking Security of Diffusion-Based Generative Steganography

Generative image steganography is a technique that conceals secret messages within generated images, without relying on pre-existing cover images. Recently, a number of diffusion model-based generative image steganography DM-GIS methods have been introduced, which effectively combat traditional...

SAFuzz: Semantic-Guided Adaptive Fuzzing for LLM-Generated Code

While AI-coding assistants accelerate software development, current testing frameworks struggle to keep pace with the resulting volume of AI-generated code. Traditional fuzzing techniques often allocate resources uniformly and lack semantic awareness of algorithmic vulnerability patterns, leading...

The Role of Learning in Attacking Intrusion Detection Systems

Recent work on network attacks have demonstrated that ML-based network intrusion detection systems NIDS can be evaded with adversarial perturbations. However, these attacks rely on complex optimizations that have large computational overheads, making them impractical in many real-world settings. ...

Protecting Context and Prompts: Deterministic Security for Non-Deterministic AI

Large Language Model LLM applications are vulnerable to prompt injection and context manipulation attacks that traditional security models cannot prevent. We introduce two novel primitives--authenticated prompts and authenticated context--that provide cryptographically verifiable provenance acros...

CVE-2026-25230

CVE-2026-25230 affects FileRise, a self-hosted web file manager/WebDAV server. Before version 3.3.0, an authenticated user could inject HTML to modify the DOM, enabling the addition of elements that call certain endpoints or link elements that redirect on interaction. The issue is fixed in 3.3.0....