Package com.tnsai.coordination.supervisor

Interface Guardrail

All Known Implementing Classes:

HITLApproval

public interface Guardrail

A guardrail that checks whether an agent action is permitted.

Guardrails are used by SupervisorAgent to enforce safety constraints on agent actions. Each guardrail evaluates an action request and returns a result indicating whether the action is allowed.

Usage Example

 Guardrail readOnly = new Guardrail() {
     public String getName() { return "read-only"; }
     public GuardrailResult check(String agentId, String action, Map<String, Object> ctx) {
         if (action.startsWith("write")) {
             return new GuardrailResult(false, "Write operations not allowed", GuardrailSeverity.BLOCK);
         }
         return new GuardrailResult(true, "Allowed", GuardrailSeverity.INFO);
     }
 };
 

See Also:

• SupervisorAgent
• HITLApproval

Nested Class Summary

Nested Classes

Modifier and Type	Interface	Description
static final record	Guardrail.GuardrailResult	Result of a guardrail check.
static enum	Guardrail.GuardrailSeverity	Severity level for guardrail decisions.

Method Summary

Modifier and Type	Method	Description
Guardrail.GuardrailResult	check(String agentId, String action, Map<String,Object> context)	Checks whether an agent action is allowed.
String	getName()	Returns the name of this guardrail.

Method Details

getName

String getName()

Returns the name of this guardrail.

Returns:

guardrail name, must be unique within a supervisor

check

Guardrail.GuardrailResult check(String agentId,
 String action,
 Map<String,Object> context)

Checks whether an agent action is allowed.

Parameters:

agentId - the ID of the agent attempting the action

action - the action being attempted

context - additional context about the action

Returns:

result indicating whether the action is allowed