{
  "id": "ao-security-sandbox",
  "title": "Agent security controls (summary)",
  "topic": "ao-security",
  "columns": [
    "Control",
    "Purpose",
    "Related slug"
  ],
  "rows": [
    [
      "Sandbox / trusted folders",
      "Limit filesystem and network side effects",
      "ao-security-gemini-cli-sandbox"
    ],
    [
      "Prompt-injection design",
      "Keep untrusted text from becoming instructions",
      "ao-security-prompt-injection-design"
    ],
    [
      "Guardrail layers",
      "Defense in depth for tools and approvals",
      "ao-security-guardrail-layers"
    ]
  ],
  "related_slugs": [
    "ao-security",
    "ao-security-gemini-cli-sandbox",
    "ao-security-prompt-injection-design",
    "ao-security-guardrail-layers",
    "ao-security-attack-taxonomy"
  ],
  "provenance": [
    {
      "title": "iinventedlight Library notes",
      "url": "https://iinventedlight.com/library/notes/",
      "license": "per-note",
      "accessed": "2026-11-09"
    }
  ],
  "as_of": "2026-11-09",
  "license": "CC-BY-4.0 summaries; see related notes for source licenses",
  "markdown": "/library/artifacts/ao-security-sandbox.table.md"
}