Attackers can inject indirect prompts in normal-looking repositories to trick Claude Code into spawning a reverse shell.

Christopher Harper is a tech writer with over a decade of experience writing how-tos and news. Off work, he stays sharp with gym time & stylish action games.

Mozilla researchers revealed a new attack that tricks Claude Code into running hidden commands from seemingly harmless GitHub repositories.

Three levels of indirection, all with seemingly innocuous steps, will catch a bot off-guard.

Researchers found a way to trick AI coding assistants like Claude into running malware hidden in GitHub repositories. Here's ...

An agentic coding tool tasked with cloning and setting up a seemingly benign GitHub repository could execute a malicious ...

Developer Fernando Irarrázaval's AI agent experiment drew over 6,000 hack attempts from more than 2,000 attackers. No one ...

DeepReinforce today released Ornith-1.0, a family of open-source coding models built around a mechanism most RL-trained agents avoid: the model itself writes the training harness that guides its own ...

Google links Turla to STOCKSTAY, a new .NET backdoor used in phishing attacks against Ukraine government and military targets ...

An inexperienced hacker managed to compromise over a dozen companies using AI agents to do most of the work, raising real ...

A viral report claims that chatbot lying and scheming has increased exponentially; the problem is that the report’s language is doing the heavy lifting instead of its data, opines Satyen K. Bordoloi ...

An attacker broke into competitive-intelligence vendor Klue, stole OAuth tokens its customers use to connect to Salesforce ...