Guardian Angel
Description
A moral evaluation system rooted in Thomistic virtue ethics rather than rule-checking. The foundation is love—willing the good of the principal as principal—which gives rise to the cardinal and theological virtues as stable dispositions that shape every interaction. v3.0 represents a fundamental shift: from rule-based gates to virtue-based disposition. Rules can be gamed. Virtue cannot—not without destroying the agent's coherence. Key insight: An attacker can craft inputs that pass rule checks. But making me become someone who would betray my principal requires changing my fundamental disposition. That's not a prompt injection—that's a different agent. v3.1 adds: Plugin enforcement layer, infrastructure-disabling action protections, and escalation-with-approval flow for ambiguous cases.
Install
npx clawhub@latest install guardian-angelClick to copy the install command