Policies define rules that govern what your AI agents are allowed to do. When an agent sends an event that matches a policy’s conditions, Secureagentics takes the configured action — blocking the operation, sending an alert, or logging the activity. Policies can be scoped to a single agent or applied globally across all agents.Documentation Index
Fetch the complete documentation index at: https://docs.adrian.secureagentics.ai/llms.txt
Use this file to discover all available pages before exploring further.
Create a policy
POST /v1/policies
Creates a new security policy and returns the policy object. If agent_id is omitted, the policy applies globally to all agents in your account.
Request body
Human-readable name for the policy, unique within your account.
The category of control this policy enforces. One of:
block_action, rate_limit, data_filter, require_approval.Scope this policy to a specific agent by its ID. If omitted, the policy applies globally to all agents.
Conditions that must be met to trigger this policy. The structure of this object varies by
type. For example, a data_filter policy might specify patterns to match in event payloads, while a rate_limit policy specifies a threshold and window.The action Secureagentics takes when the conditions are matched. Supported actions:
block (prevents the operation), alert (sends a notification), log (records the event without interrupting the operation). You may combine multiple actions.Response fields
Unique policy identifier, prefixed with
pol_.The policy name.
The policy type:
block_action, rate_limit, data_filter, or require_approval.The agent this policy is scoped to.
null if the policy applies globally.The conditions configured for this policy.
The actions configured for this policy.
ISO 8601 timestamp of when the policy was created.
Current policy status. One of:
active, inactive.cURL
List policies
GET /v1/policies
Returns a paginated list of policies. You can filter by agent or type.
Query parameters
Return only policies scoped to this agent ID. Omit to return both global and agent-scoped policies.
Filter by policy type. One of:
block_action, rate_limit, data_filter, require_approval.Maximum number of policies to return per page. Accepted range: 1–100.
Number of policies to skip before returning results. Use for pagination.
Response fields
Array of policy objects matching the query.
Total number of policies matching the applied filters.
cURL
Delete a policy
DELETE /v1/policies/{id}
Permanently deletes a policy. Once deleted, the policy no longer evaluates incoming events. Returns 204 No Content on success.
Path parameters
The policy ID, prefixed with
pol_.cURL