ÆthelHookGuides
Notes on AI agent safety, Claude Code and Codex, and how ÆthelHook is built.
Is it safe to let Claude Code run unattended?
What can actually go wrong when an AI coding agent runs shell commands on its own, and what a second-device approval gate changes about that risk.
How to approve Claude Code and Codex from your phone
A step-by-step walkthrough of pairing ÆthelHook and answering approvals, plan reviews, and questions remotely.
Claude Code hooks explained
PreToolUse, PostToolUse, and Stop: what each hook actually receives and returns, and how to build your own approval gate on top of them.