Towards Unifying Quantitative Security Benchmarking for Multi Agent Systems

Evolving AI systems increasingly deploy multi-agent architectures where autonomous agents collaborate, share information, and delegate tasks through developing protocols. This connectivity, while powerful, introduces novel security risks. One such risk is a cascading risk: a breach in one agent c...

PT-2025-30619 · Robocode · Robocode

Name of the Vulnerable Software and Affected Versions: Roo Code versions 3.23.18 and below Description: Roo Code, an AI-powered autonomous coding agent, does not validate line breaks in its command input. This bypasses the allow-list mechanism due to a lack of parsing or validation logic,...

CVE-2025-51859

Stored Cross-Site Scripting XSS vulnerability in Chaindesk thru 2025-05-26 in its agent chat component. An attacker can achieve arbitrary client-side script execution by crafting an AI agent whose system prompt instructs the underlying Large Language Model LLM to embed malicious script payloads...

PT-2025-30428 · Chaindesk · Chaindesk

Name of the Vulnerable Software and Affected Versions: Chaindesk versions through 2025-05-26 Description: A stored Cross-Site Scripting XSS vulnerability exists in the agent chat component. An attacker can execute arbitrary client-side scripts by creating an AI agent with a system prompt designed...

CVE-2025-51859

Chaindesk is affected by a stored XSS in the agent chat component through crafted system prompts that instruct the LLM to inject malicious payloads (e.g., SVG-based XSS) into chat responses. This can cause arbitrary client-side script execution in victims’ browsers, enabling theft of sensitive da...

Cannot find EFI boot manager entry for EFI System Partition

Challenge Backup During the backup of some Linux-based machines, the following error occurs: Cannot find EFI boot manager entry for EFI System Partition GUID Restore During a Bare Metal Restore BMR from a restore point created by a backup that displayed that error, the following error will occur:...

CVE-2025-51472

Code Injection in AgentTemplate.evalagentconfig in TransformerOptimus SuperAGI 0.0.14 allows remote attackers to execute arbitrary Python code via malicious values in agent template configurations such as the goal, constraints, or instruction field, which are evaluated using eval without validati...

CVE-2025-51472

Code Injection in AgentTemplate.evalagentconfig in TransformerOptimus SuperAGI 0.0.14 allows remote attackers to execute arbitrary Python code via malicious values in agent template configurations such as the goal, constraints, or instruction field, which are evaluated using eval without validati...

The vulnerability of the GetScreen agent, a remote access platform, relates to insecure privilege management. This allows attackers to elevate their privileges to the root level.

The vulnerability of the remote access platform GetScreen Agent is related to insecure management of privileges. Exploiting this vulnerability can allow a hacker to elevate their privileges to the root level...

Okta On-Premises Provisioning Agent 日志信息泄露漏洞

Okta On-Premises Provisioning Agent is a software from Okta USA that automates the creation, update and deletion of Otka user accounts in a local environment. A log information disclosure vulnerability exists in Okta On-Premises Provisioning Agent version 2.3.0 and prior versions, which originate...

DB-GPT 命令注入漏洞

DB-GPT is an AWEL and agent-based AI native data application development framework open-sourced by eosphoros. A security vulnerability exists in DB-GPT version 0.7.0, which stems from a file upload vulnerability in agent.hub.controller.refreshplugins that could lead to the execution of arbitrary...

WordPress plugin Realty Portal – Agent 安全漏洞

WordPress and WordPress plugin are both products of the WordPress Foundation.WordPress is a blogging platform developed using the PHP language. The platform supports setting up personal blog sites on servers with PHP and MySQL.WordPress plugin is an application plugin. A security vulnerability...

About Remote Code Execution – Internet Shortcut Files (CVE-2025-33053) vulnerability

About Remote Code Execution - Internet Shortcut Files CVE-2025-33053 vulnerability. A vulnerability from the June Microsoft Patch Tuesday. This vulnerability immediately showed signs of exploitation in the wild. This flaw allows a remote attacker to execute arbitrary code when a victim opens a...

Trellix Endpoint Security 代码注入漏洞

Trellix Endpoint Security ENS is an endpoint security solution from FireEye Trellix USA. A code injection vulnerability exists in the Trellix Endpoint Security HX Agent that originates from a privileged user being able to create a malicious OpenSSL configuration file that could lead to the loadin...

The vulnerability of the network traffic analysis, network detection, and response tool Palo Alto Networks Cortex XDR Agent lies in the insufficient checking of unusual or exceptional states. This allows attackers to trigger service failures.

The vulnerability of the Palo Alto Networks Cortex XDR Agent, which is used for network traffic analysis, network detection, and response, stems from insufficient checking of unusual or exceptional states. Exploiting this vulnerability could allow attackers to trigger service failures...

EulerOS 2.0 SP12 : openssh (EulerOS-SA-2025-1832)

According to the versions of the openssh packages installed, the EulerOS installation on the remote host is affected by the following vulnerabilities : In sshd in OpenSSH before 10.0, the DisableForwarding directive does not adhere to the documentation stating that it disables X11 and agent...

📄 Mouse Agent Server 3.1 Remote Code Execution

Mouse Agent Server version 3.1 is vulnerable to unauthenticated remote code execution by simulating mouse/keyboard inputs to force the target to execute a PowerShell reverse shell. It works against default configurations by sending GUI automation commands through port 8088. Exploit Title: Mouse...

📄 Mouse Agent Server 3.1 Unauthenticated Remote System Control

Mouse Agent Server version 3.1 exposes a TCP control interface on port 8088 that allows remote execution of power commands shutdown, restart, sleep, logoff via unauthenticated commands. An attacker on the same network can exploit this to disrupt the system remotely without user interaction. Explo...

Remote Code Execution (RCE)

github.com/juju/juju is vulnerable to Remote Code Execution RCE. The vulnerability is due to insufficient authorization checks caused by allowing any authenticated controller user to upload arbitrary agent binaries to any model or the controller without verifying model membership or permissions...

Prompt Injection 2.0: Hybrid AI Threats

Prompt injection attacks, where malicious input is designed to manipulate AI systems into ignoring their original instructions and following unauthorized commands instead, were first discovered by Preamble, Inc. in May 2022 and responsibly disclosed to OpenAI. Over the last three years, these...