Examples

Working code examples and integration patterns for Agent Control.

Browse working examples that show how to add Agent Control to different frameworks, evaluators, and real-world use cases. Source code for every example lives in the agent-control/examples directory.

Getting started

Agent Control Demo

End-to-end walkthrough — start the server, add controls, and see blocking in action in four steps.

TypeScript SDK

Consume the published npm package from an external TypeScript app (OpenClaw-compatible).

Framework Integrations

LangChain SQL Agent

Protect a LangChain SQL agent from dangerous queries with server-side SQL injection controls.

AWS Strands

Add guardrails to AWS Strands agent workflows and tool calls without changing business logic.

CrewAI

Layer Agent Control security (PII detection, access blocking) alongside CrewAI’s quality guardrails.

Google ADK Plugin

Recommended packaged Google ADK integration with model and tool guardrails through AgentControlPlugin.

Google ADK Callbacks

Lower-level Google ADK lifecycle hook integration for manual pre/post-model and tool enforcement.

Google ADK Decorator

Tool-only Google ADK integration using explicit @control() wrapping.

Evaluator Integrations

DeepEval GEval

Build a custom evaluator using DeepEval’s LLM-as-a-judge metric for coherence, relevance, and correctness scoring.

Galileo Luna-2

Real-time toxicity detection and content moderation with Galileo’s Luna-2 Protect service.

Use case examples

Steer Action Demo

Banking transfer agent with steer actions for approvals and compliance workflows.

Customer Support Agent

Realistic enterprise scenario with PII protection, prompt-injection defense, and multiple tool types.