Stop LLM Attacks: How Security Helps AI Apps Achieve Their ROI

AI security is a business problem. Protect your LLM application investment and ROI by connecting your security team with business stakeholders...

PT-2026-24113

Name of the Vulnerable Software and Affected Versions vLLM versions prior to 0.15.1 vLLM version 0.17.0 Description vLLM is an inference and serving engine for large language models LLMs. A Server-Side Request Forgery SSRF protection mechanism implemented in version 0.15.1 can be bypassed in the...

CIA+TA Risk Assessment for AI Reasoning Vulnerabilities

As AI systems increasingly influence critical decisions, they face threats that exploit reasoning mechanisms rather than technical infrastructure. We present a framework for cognitive cybersecurity, a systematic protection of AI reasoning processes from adversarial manipulation. Our contributions...

Systematic Analysis of MCP Security

The Model Context Protocol MCP has emerged as a universal standard that enables AI agents to seamlessly connect with external tools, significantly enhancing their functionality. However, while MCP brings notable benefits, it also introduces significant vulnerabilities, such as Tool Poisoning...

LLM Coding Integrity Breach

Here's an interesting story about a failure being introduced by LLM-written code. Specifically, the LLM was doing some code refactoring, and when it moved a chunk of code from one file to another it changed a "break" to a "continue." That turned an error logging statement into an infinite loop,...

CVE-2025-54382 Cherry Studio RCE Vulnerability Disclosure

Cherry Studio is a desktop client that supports for multiple LLM providers. In version 1.5.1, a remote code execution RCE vulnerability exists in the Cherry Studio platform when connecting to streamableHttp MCP servers. The issue arises from the server’s implicit trust in the oauth auth redirecti...

CVE-2025-45146

ModelCache for LLM through v0.2.0 was discovered to contain an deserialization vulnerability via the component /manager/datamanager.py. This vulnerability allows attackers to execute arbitrary code via supplying crafted data...

Shadow in the Cache: Unveiling and Mitigating Privacy Risks of KV-Cache in LLM Inference

The Key-Value KV cache, which stores intermediate attention computations Key and Value pairs to avoid redundant calculations, is a fundamental mechanism for accelerating Large Language Model LLM inference. However, this efficiency optimization introduces significant yet underexplored privacy risk...

Security Bulletin: NVIDIA Megatron LM - August 2025

NVIDIA has released a software update for NVIDIA® Megatron LM. To protect your system, clone or update this software to version 0.12.2 or later from Releases · NVIDIA/Megatron-LM on GitHub. Go to NVIDIA Product Security...

CVE-2025-45146

ModelCache for LLM through v0.2.0 was discovered to contain an deserialization vulnerability via the component /manager/datamanager.py. This vulnerability allows attackers to execute arbitrary code via supplying crafted data...

BlindGuard: Safeguarding LLM-Based Multi-Agent Systems under Unknown Attacks

The security of LLM-based multi-agent systems MAS is critically threatened by propagation vulnerability, where malicious agents can distort collective decision-making through inter-agent message interactions. While existing supervised defense methods demonstrate promising performance, they may be...

CVE-2025-45146

CVE-2025-45146 applies to ModelCache for LLM through v0.2.0, where a deserialization vulnerability in /manager/data_manager.py allows arbitrary code execution via crafted data. The CVE’s CVSSv3.1 vector yields a base score of 9.8 (CRITICAL) with network attack vector, no privileges required, and ...

CVE-2025-45146

ModelCache for LLM through v0.2.0 was discovered to contain an deserialization vulnerability via the component /manager/datamanager.py. This vulnerability allows attackers to execute arbitrary code via supplying crafted data...

ClearML Detected

This is an informational plugin to inform the user that the scanner has detected a publicly accessible ClearML instance on the target application. ClearML is an infrastructure platform for AI builders. This detection is included in the AI and LLM category. No source data...

NVIDIA Triton Detected

This is an informational plugin to inform the user that the scanner has detected a publicly accessible NVIDIA Triton instance on the target application. NVIDIA Triton provides an optimized cloud and edge inferencing solution. This detection is included in the AI and LLM category. No source data...

PT-2025-32556 · Unknown · Modelcache For Llm

Name of the Vulnerable Software and Affected Versions: ModelCache for LLM versions through 0.2.0 Description: ModelCache for LLM through version 0.2.0 contains a deserialization vulnerability in the /manager/data manager.py component. This allows attackers to execute arbitrary code by supplying...

Who'S the Evil Twin? Differential Auditing for Undesired Behavior

Detecting hidden behaviors in neural networks poses a significant challenge due to minimal prior knowledge and potential adversarial obfuscation. We explore this problem by framing detection as an adversarial game between two teams: the red team trains two similar models, one trained solely on...

Mitigating Distribution Shift in Graph-Based Android Malware Classification Via Function Metadata and LLM Embeddings

Graph-based malware classifiers can achieve over 94% accuracy on standard Android datasets, yet we find they suffer accuracy drops of up to 45% when evaluated on previously unseen malware variants from the same family - a scenario where strong generalization would typically be expected. This...

Incident Response Planning Using a Lightweight Large Language Model with Reduced Hallucination

Timely and effective incident response is key to managing the growing frequency of cyberattacks. However, identifying the right response actions for complex systems is a major technical challenge. A promising approach to mitigate this challenge is to use the security knowledge embedded in large...

Attack the Messages, Not the Agents: a Multi-Round Adaptive Stealthy Tampering Framework for LLM-MAS

Large language model-based multi-agent systems LLM-MAS effectively accomplish complex and dynamic tasks through inter-agent communication, but this reliance introduces substantial safety vulnerabilities. Existing attack methods targeting LLM-MAS either compromise agent internals or rely on direct...