#AI agents
What an AI agent actually does on your laptop
Chatbots answer. Agents act: they read files, run commands and call services on your behalf. Here is what that looks like from the operating system’s point of view, and what is worth keeping an eye on.
Coding agents and your SSH keys
A coding agent runs with your user account, which means it can usually read the same keys and tokens you can. How that happens, why it matters, and practical ways to narrow what an agent can reach.
MCP servers explained
The Model Context Protocol lets AI apps plug into tools and data through small programs called MCP servers. What they are, how they connect, and how to decide which ones to trust.
How GreyArm detects agents without reading your data
GreyArm finds AI agents by looking at behaviour (processes, connections, lookups and ports) rather than content. A tour of what the sensor collects, what it deliberately ignores, and how findings are scored.
14 days free. No card. Every computer and phone.
Start free
Five signs an AI script is uploading more than it should
Automation scripts that call AI services are easy to write and easy to forget. Five practical signals that one is sending more data than its job needs, and what to do when you spot them.
New agents we’ve seen in the wild, what they talk to, and how to keep them in check.