New attack on OpenClaw or similar AI agents that infects them with “malicious skills.” 🤖💉👾😵💫🤡
A group of academic security researchers from Shanghai Jiao Tong University (China) introduce a new type of cyberattack targeting autonomous coding agents like OpenClaw - guidance injection. Quite interesting and very practical.
In OpenClaw, a skill is basically a plugin that adds new abilities to the AI agent. You can download them from ClawHub and install them. AFAIK, there is no solid third-party review of these skills. This is where the vulnerability becomes practical.
In one attack, a malicious skill convinced the agent that .git folders are “temporary files,” so when you ask it to free up space, it deletes your entire project history - all while thinking it’s following best practices.
Enjoy the read - super interesting for anyone who’s into AI agent development. And if you’re looking to enter AI cybersecurity, this may be a good place to start.
More details:
Trojan’s Whisper: Stealthy Manipulation of OpenClaw through Injected Bootstrapped Guidance [PDF]: https://arxiv.org/abs/2603.19974
目录
最新
- Hacking Kia and Nissan cars remotely: the status of famous vulnerabilities after two years. 🚗🌎📱🔓🤑
- Hacking serial-to-IP industrial devices from Lantronix and Silex: attacking OT from the internet. 🏭📦🌐💉👷♂️
- Unlock your car’s hidden hacking potential: how to enable “Developer Mode” in Android Auto. 🚘📱👈🤫👑
- Hacking access-control systems with an old-school sniffer: a creative way to jump the fence. 🚪🔐👃💳🔓
- Hacking the Google Fast Pair protocol: Connect to 68% of headphones, earbuds, and cars (?).🎧🚙ᯤ👨🏻💻😱
- Hacking the diagnostics of Freightliner Cascadia heavy trucks: Denial-of-service attacks. 🚛🩺📊💉🔴
- Crash course in RF security research for hackers: cheap hardware and open-source software. 👨🏻💻📻၊၊||၊🎧👾
- Cybersecurity attacks on SIM cards: a wealth of useful info about SIM and cellular network hacking. 📱🎫👥📲💰