kata-agent-guardrails-design
OfficialDesign agent guardrails, auth, and escalation.
Legal & Compliance#authorization#agent security#pii redaction#owasp llm top 10 2025#escalation runbook#cross-tenant isolation
Authorguardiatechnology
Version1.0.0
Installs0
System Documentation
What problem does it solve?
It helps you define concrete security boundaries for an AI agent—preventing prompt/tool injection, controlling sensitive data exposure, enforcing authorization, and specifying when and how to escalate.
Core Features & Use Cases
- Guardrails blueprint: Produces guardrails.md covering OWASP LLM Top 10 critical controls, including PII handling, cross-tenant boundaries, and excessive agency protection.
- Authorization model: Produces authorization.md that enumerates permitted callers, scopes, and an explicit auth model for each invocation path.
- Escalation matrix: Produces escalation.md with triggers, severities, response SLAs, linked runbooks, and orchestrator fallback behavior.
Quick Start
Ask the AI to generate guardrails.md, authorization.md, and escalation.md for the agent named "{agent}" within the bounded context "{context}" using the OWASP LLM Top 10 2025 and your specified runbook paths and callers.
Dependency Matrix
Required Modules
None requiredComponents
Standard package💻 Claude Code Installation
Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.
Please help me install this Skill: Name: kata-agent-guardrails-design Download link: https://github.com/guardiatechnology/design-system/archive/main.zip#kata-agent-guardrails-design Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
Agent Skills Search Helper
Install a tiny helper to your Agent, search and equip skill from 471,000+ vetted skills library on demand.