Blog
AI Threat Watch
AI has made breaking into businesses faster and cheaper. Each week I read the security press so you don't have to, and write up the AI-driven attacks that matter: what happened, who it hit, and what a small or mid-sized business should do about it.
Latest ·
AI cyberattacks this week: Oct 26 – Nov 1, 2026
AI agents ran a state hacking campaign, a poisoned Claude Code extension hit developers, and AI keys kept leaking: the week's AI cyberattacks, explained.
Read the briefingEarlier briefings
PHP2Shell: is your website affected? A plain-English guide
PHP2Shell explained for website owners: whether your host has patched it, how to check your PHP version, what to ask, and signs of a compromised site.
Read moreA poisoned Claude Code extension backdoored developer machines: what to check
For about 11 hours, Cursor's marketplace served a backdoored Claude Code extension. What it did, who was exposed, and what your team should check now.
Read moreStolen AI API keys: how to protect your business
Attackers are stealing AI API keys to run up bills, resell access and copy models. What happened in October 2026, and the settings that limit the damage.
Read moreBlack October: what the AI cyberattack wave means for small businesses
Black October explained: the late-October 2026 wave of AI-assisted attacks, from PHP2Shell to stolen AI keys, and what a small business should do now.
Read more