update_agent_policy
ActiveTool of SendIt
declared in 0.1.0
Create or update a policy/guardrail for an agent. Policy types: • brand_safety - Content must align with brand guidelines • platform_compliance - Must follow platform ad policies • approval_gate - Require human approval before execution • budget_limit - Cap spending or resource usage • content_filter - Filter certain topics or language
Parameters schema
{
"type": "object",
"required": [
"agentId",
"policyType",
"rules"
],
"properties": {
"rules": {
"type": "array",
"items": {
"type": "object",
"required": [
"condition",
"action",
"reason"
],
"properties": {
"action": {
"enum": [
"allow",
"block",
"require_approval",
"warn"
],
"type": "string"
},
"reason": {
"type": "string",
"description": "Reason for this rule"
},
"condition": {
"type": "string",
"description": "Rule condition"
}
}
},
"description": "Policy rules"
},
"active": {
"type": "boolean",
"description": "Whether policy is active"
},
"agentId": {
"enum": [
"strategy_planner",
"content_ideation",
"multi_format_composer",
"creative_asset",
"variant_repurposer",
"calendar_optimizer",
"listening_analyst",
"inbox_reply",
"campaign_builder",
"budget_optimizer",
"experimentation",
"executive_insights"
],
"type": "string",
"description": "Agent to set policy for"
},
"policyType": {
"enum": [
"brand_safety",
"platform_compliance",
"approval_gate",
"budget_limit",
"content_filter"
],
"type": "string",
"description": "Policy type"
}
}
}No endpoints wrapped at confidence ≥ 0.70.
Parent server
SendIt
https://github.com/Shree-git/sendit
1/7 registries