LLM Agents Should Employ Security Principles

Large Language Model LLM agents show considerable promise for automating complex tasks using contextual reasoning; however, interactions involving multiple agents and the system's susceptibility to prompt injection and other forms of context manipulation introduce new vulnerabilities related to...

📄 Automic Agent 24.3.0 HF4 Privilege Escalation

Automic Agent version 24.3.0 HF4 suffers from a privilege escalation vulnerability. Exploit Title: Automic Agent 24.3.0 HF4 - Privilege Escalation Date: 26.05.2025 Exploit Author: Flora Schäfer Vendor Homepage: https://www.broadcom.com/products/software/automation/automic-automation Version:...

Automic Agent 24.3.0 HF4 - Privilege Escalation

Exploit Title: Automic Agent 24.3.0 HF4 - Privilege Escalation Date: 26.05.2025 Exploit Author: Flora Schäfer Vendor Homepage: https://www.broadcom.com/products/software/automation/automic-automation Version: /tmp/sh.so 2. Run the ucxjlx6 executable as follows $ ./ucxjlx6 ini=echo -e...

A Novel Zero-Trust Identity Framework for Agentic AI: Decentralized Authentication and Fine-Grained Access Control

Traditional Identity and Access Management IAM systems, primarily designed for human users or static machine identities via protocols such as OAuth, OpenID Connect OIDC, and SAML, prove fundamentally inadequate for the dynamic, interdependent, and often ephemeral nature of AI agents operating at...

VulBinLLM: LLM-Powered Vulnerability Detection for Stripped Binaries

Recognizing vulnerabilities in stripped binary files presents a significant challenge in software security. Although some progress has been made in generating human-readable information from decompiled binary files with Large Language Models LLMs, effectively and scalably detecting vulnerabilitie...

CVE-2025-4412

CVE-2025-4412 concerns macOS: an attacker can use a Launch Agent to load viscosity_openvpn from the Viscosity app bundle and induce a dynamic library load under Viscosity’s TCC identity. This grants limited resource access without entitlements (e.g., not granting camera/mic); access to other reso...

Sparklabs Viscosity 安全漏洞

Sparklabs Viscosity is an OpenVPN client from Sparklabs Australia. A security vulnerability exists in SparkLabs Viscosity versions prior to 1.11.5, which stems from the possibility of exploiting the Launch Agent to load dynamic libraries to gain limited access to resources...

Agent is not appraising the disk space

ADM Agent Disk Space gets full sporadically, and due to which the ADM Agent goes Down...

ABB M2M Gateway Arbitrary Code Execution in embedded OpenSSH (CVE-2023-38408)

The PKCS11 feature in ssh-agent in OpenSSH before 9.3p2 has an insufficiently trustworthy search path, leading to remote code execution if an agent is forwarded to an attacker-controlled system. Code in /usr/lib is not necessarily safe for loading into ssh-agent. NOTE: this issue exists because o...

RHEL 9 : keylime-agent-rust (RHSA-2025:7313)

The remote Redhat Enterprise Linux 9 host has a package installed that is affected by a vulnerability as referenced in the RHSA-2025:7313 advisory. Rust agent for Keylime Security Fixes: rust-openssl: rust openssl ssl::selectnextproto use after free CVE-2025-24898 For more details about the...

PandaGuard: Systematic Evaluation of LLM Safety against Jailbreaking Attacks

Large language models LLMs have achieved remarkable capabilities but remain vulnerable to adversarial prompts known as jailbreaks, which can bypass safety alignment and elicit harmful outputs. Despite growing efforts in LLM safety research, existing evaluations are often fragmented, focused on...

MultiPhishGuard: an LLM-Based Multi-Agent System for Phishing Email Detection

Phishing email detection faces critical challenges from evolving adversarial tactics and heterogeneous attack patterns. Traditional detection methods, such as rule-based filters and denylists, often struggle to keep pace with these evolving tactics, leading to false negatives and compromised...

CoTGuard: Using Chain-Of-Thought Triggering for Copyright Protection in Multi-Agent LLM Systems

As large language models LLMs evolve into autonomous agents capable of collaborative reasoning and task execution, multi-agent LLM systems have emerged as a powerful paradigm for solving complex problems. However, these systems pose new challenges for copyright protection, particularly when...

Benchmarking Poisoning Attacks against Retrieval-Augmented Generation

Retrieval-Augmented Generation RAG has proven effective in mitigating hallucinations in large language models by incorporating external knowledge during inference. However, this integration introduces new security vulnerabilities, particularly to poisoning attacks. Although prior work has explore...

CVE-2025-24828

Local privilege escalation due to DLL hijacking vulnerability. The following products are affected: Acronis Cyber Protect Cloud Agent Windows before build 39378...

CVE-2025-24827

Local privilege escalation due to DLL hijacking vulnerability. The following products are affected: Acronis Cyber Protect Cloud Agent Windows before build 39378...

CVE-2025-24831

Local privilege escalation due to unquoted search path vulnerability. The following products are affected: Acronis Cyber Protect Cloud Agent Windows before build 39378...

CVE-2025-24460

In JetBrains TeamCity before 2024.12.1 improper access control allowed to see Projects’ names in the agent pool...

OESA-2025-1550 openssh security update

OpenSSH is the premier connectivity tool for remote login with the SSH protocol. \ It encrypts all traffic to eliminate eavesdropping, connection hijacking, and \ other attacks. In addition, OpenSSH provides a large suite of secure tunneling \ capabilities, several authentication methods, and...

CVE-2024-20390

A vulnerability in the Dedicated XML Agent feature of Cisco IOS XR Software could allow an unauthenticated, remote attacker to cause a denial of service DoS on XML TCP listen port 38751. This vulnerability is due to a lack of proper error validation of ingress XML packets. An attacker could explo...