Agent Sandbox Taxonomy (AST) — Evaluation & Composition Skill

What problem does it solve?

This Skill helps you understand and implement robust security for AI agents by leveraging the Agent Sandbox Taxonomy (AST), eliminating guesswork in sandbox selection and configuration.

Core Features & Use Cases

• Sandbox Scoring: Evaluate any sandbox against the AST's 7 defense layers and 7 threat categories to generate a detailed score card and fingerprint.
• Sandbox Composition: Recommend and stack complementary sandbox products to achieve comprehensive threat coverage, filling security gaps.
• Use Case: You need to deploy an AI agent that interacts with sensitive cloud resources. Use this Skill to identify a sandbox stack that provides strong compute isolation (L1), network controls (L4), credential management (L5), and action governance (L6) to prevent data exfiltration and destructive operations.

Quick Start

Use the AST skill to score the 'Docker Sandbox' product.