When you need to check texts, pull a 2FA code, or read through a DM thread before replying, this handles the evidence layer so you're not guessing which source you actually hit. It enforces a "resolve the surface first" rule (local messages, X DMs, whatever) and won't let you claim you checked something without naming where you looked. Built for the ECC workflow where proof of retrieval matters as much as the content itself. The output format is tight: source, result, status. If you've ever grabbed the wrong code or replied to a thread you didn't actually reread, this keeps that from happening again.
npx skills add https://github.com/affaan-m/everything-claude-code --skill messages-ops