{"id":"ai-agent-permissions-insurance","slug":"ai-agent-permissions-insurance-crm-scoped-keys","title":"Scoped API Keys for AI Agents: How to Let Claude Write to Some Things and Not Others","excerpt":"Giving Claude full admin on your CRM is a footgun. Here's how to issue scoped API keys so AI agents have just the permissions they need — and nothing more.","content":"<p>Every \"Claude deleted my production database\" horror story shares one cause: an agent had broader permissions than it needed. Scoped API keys are the fix.</p>\n\n<h2>The Principle</h2>\n<p>Treat every AI agent like a contractor: give it the smallest possible permission set that lets it do its job, with a per-agent key so audit logs are clear.</p>\n\n<h2>Common Scope Patterns in Insurance</h2>\n<h3>Read-Only Research Agent</h3>\n<p>Scopes: contacts:read, policies:read, commissions:read. Use case: \"summarize my book,\" \"find at-risk renewals.\"</p>\n\n<h3>Outreach Drafting Agent</h3>\n<p>Scopes: contacts:read, policies:read, tasks:write, notes:write. Can draft and queue but not actually send messages without a human approval step.</p>\n\n<h3>Renewal Worker</h3>\n<p>Scopes: policies:read, tasks:write, opportunities:write. Background agent that monitors renewals and creates work items.</p>\n\n<h3>Full Admin (use sparingly)</h3>\n<p>Reserved for one-off cleanup tasks under direct human supervision. Rotate immediately after use.</p>\n\n<h2>Per-User Keys vs Per-Agent Keys</h2>\n<p>Per-user keys let you see who triggered what. Per-agent keys let you see what tool did what. Both. Issue both. Audit logs make incident response 10x easier.</p>\n\n<h2>Rotation Hygiene</h2>\n<ul>\n<li>Rotate keys quarterly minimum</li>\n<li>Rotate immediately on any team member departure</li>\n<li>Rotate immediately if a key was ever pasted into a chat that wasn't end-to-end encrypted</li>\n</ul>","category":"Security","featureLink":"/mcp-server","cluster":"mcp-api","date":"2026-11-18","lastModified":"2026-11-18","readTime":"5 min","author":"unLocked Team"}