Mac hygiene & agentic AI security

Practical, no-nonsense deep-dives on truly uninstalling macOS apps and keeping local AI agents — Claude Code, Cursor, Codex, Ollama, MCP servers — from leaking your secrets or wrecking your machine.

Threat Deep-Dive

Your AI coding agent is exfiltrating your secrets — how to stop it

Prompt injection and poisoned MCP tools turn Claude Code, Cursor and Codex into insider threats that read your keys and ship them out — with no alert. The attack chain and how to detect it.

Agentic AI Security

Agentic AI security, locally: the hidden, expensive risk on your Mac

Autonomous AI agents on your laptop inherit your privileges, act on untrusted input, and are invisible to your security stack. Why that's costly — and how to keep them in check.

Deep Uninstall

The Mac way to uninstall apps is broken — and it's a security risk

Dragging to the Trash only removes the icon. Caches, databases, launch agents and leftover credentials stay behind — slowing your Mac and leaking data you thought was gone.