One Hacker Used AI Agents to Breach Over 100 E-commerce Sites, Report Finds

A report from cybersecurity firm NorForge, published on October 19, reveals that a Chinese-speaking hacker used AI agents coordinated through the open-source engine Cairn to launch attacks on e-commerce websites across the United States, Canada, Australia, and Japan. Since September 25, the hacker attempted to hit more than 1,200 sites, successfully breaching over 100 and stealing data, credit card numbers, and cryptocurrency.

According to the report, the hacker outsourced nearly all tasks to AI agents, using prompts of fewer than 100 Chinese characters. The agents scanned digital records for exposed control panels, crafted exploits, and tricked servers into executing malicious commands. Each attack took about 23 hours on average, with a success rate of roughly 10%. NorForge estimated the token costs for the campaign at about $5,000, translating to roughly $50 per successful breach.

Cairn, an open-source AI problem-solving and state-management engine available on GitHub, has gained popularity among hackers. The report says the underlying models were largely DeepSeek and Anthropic's Claude Code. Cairn's author, @le1xia0, responded on Twitter that Cairn is an innovative multi-agent architecture not designed for hacking, and any harmful use was not intended.