Zero Trust Protocol - Security-First Agent Guidelines

• Risk level: Low
• Analyzed: Feb 25, 2026, 6:21 AM
• Findings: 0
• Version: 1.0.1

Overview

This is an instruction-only, self-consistent zero‑trust guideline for agents; it asks for no credentials, installs nothing, and its checks and restrictions match the stated purpose.

This skill is a set of behavioral rules (no code, no external installs, no credentials requested) and is internally coherent. Before installing: (1) verify the publisher/source (clawhub.com and the registry owner ID) to ensure you trust the origin, (2) understand that these are advisory rules — your agent or platform must actually enforce them (the skill cannot force behavior), (3) if you centralize credential storage as suggested, ensure appropriate filesystem permissions and backups, and (4) consider integrating these checks with automated vetters or logs if you want enforcement rather than just guidance.