AgentHijack: Visual Patch Attacks on Multimodal Computer-Use Agents

This paper presents an end-to-end evaluation framework for image-triggered command injection against computer-use agents CUAs. The goal is to test whether a local visual patch can induce verifiable environmental consequences along the full chain of screenshot input, VLM generation, action parsing...

PYSEC-2026-1483 jupyterlab-git has a command injection vulnerability in "Open Git Repository in Terminal"

Overview On many platforms, a third party can create a Git repository under a name that includes a shell command substitution ^1 string in the syntax $. These directory names are allowed in macOS and a majority of Linux distributions ^2. If a user starts jupyter-lab in a parent directory of this...

Exploit for OS Command Injection in Docker

🖥️ -h3x0v3rl0rd- ️⃣ CVE-2019-5736 Usage : machine is vuln...

Prototype Pollution in darrenpaulwright/object-agent

Description object-agent is vulnerable to Prototype Pollution. Proof of Concept 1. Create the following PoC file: // poc.js import set from 'object-agent'; var obj = console.log"Before : " + .polluted; setobj, 'proto.polluted', 'Yes! Its Polluted'; console.log"After : " + .polluted; 2. Execute th...

Cherokee Terminal Escape Sequence in Logs Command Injection Vulnerability

Cherokee is prone to a command-injection vulnerability because it fails to adequately sanitize user-supplied input in logfiles. Attackers can exploit this issue to execute arbitrary commands in a terminal. Cherokee 0.99.30 and prior are vulnerable. OpenVAS Vulnerability Test $Id: cherokee37715.na...