grok-bot:Y3xV-T0yxz2UihvOrlIK6

Security Warden

Gates another bot's risky moves like outbound messages, deletions, and shell, checking for secrets and personal data before they run

Overview

A careful second opinion before risky bot actions. Reviews proposed sends, deletes, and shell for secrets and personal info, then says go / ask you first / don't do that. Cooperation on by default; digests, LLM Guard, and wrapped-route enforcement are optional opt-ins.

Contents

  • skill Warden Getting Started Use when onboarding a new Security Warden owner: first-run options, defaults, smoke test, cooperation vs enforcement.
  • skill Warden Operating Instructions Use when acting as Security Warden: who you are, plain-English tone, cooperation default, enforcement opt-in, and rules you never break.
  • skill Warden Review Use when reviewing a proposed action and answering go / ask you first / don't do that in plain English.
  • skill Warden Status Use when explaining what Warden checks, what's on or off, cooperation vs enforcement, and what it can't see.
  • skill Warden Policy Use when showing Warden's rules in plain English or proposing (never applying) a rule change.
  • skill Warden Incident Use when something serious turns up: pause advice, minimal evidence, clear scope.
  • memory Memory entries · 7 items

Dependencies

No plugin dependencies are declared.

Attachments

No separate attachments are published for this record.