# How to integrate 21risk MCP with OpenAI Agents SDK

```json
{
  "title": "How to integrate 21risk MCP with OpenAI Agents SDK",
  "toolkit": "21risk",
  "toolkit_slug": "_21risk",
  "framework": "OpenAI Agents SDK",
  "framework_slug": "open-ai-agents-sdk",
  "url": "https://composio.dev/toolkits/_21risk/framework/open-ai-agents-sdk",
  "markdown_url": "https://composio.dev/toolkits/_21risk/framework/open-ai-agents-sdk.md",
  "updated_at": "2026-05-12T09:59:46.287Z"
}
```

## Introduction

This guide walks you through connecting  21risk to the OpenAI Agents SDK using the Composio tool router. By the end, you'll have a working  21risk agent that can download audit reports for last quarter, list all sites with open compliance issues, fetch monthly risk data for top properties through natural language commands.
This guide will help you understand how to give your OpenAI Agents SDK agent real control over a  21risk account through Composio's  21risk MCP server.
Before we dive in, let's take a quick look at the key ideas and tools involved.

## Also integrate 21risk with

- [Claude Agent SDK](https://composio.dev/toolkits/_21risk/framework/claude-agents-sdk)
- [Claude Code](https://composio.dev/toolkits/_21risk/framework/claude-code)
- [Claude Cowork](https://composio.dev/toolkits/_21risk/framework/claude-cowork)
- [Codex](https://composio.dev/toolkits/_21risk/framework/codex)
- [OpenClaw](https://composio.dev/toolkits/_21risk/framework/openclaw)
- [Hermes](https://composio.dev/toolkits/_21risk/framework/hermes-agent)
- [CLI](https://composio.dev/toolkits/_21risk/framework/cli)
- [Google ADK](https://composio.dev/toolkits/_21risk/framework/google-adk)
- [LangChain](https://composio.dev/toolkits/_21risk/framework/langchain)
- [Vercel AI SDK](https://composio.dev/toolkits/_21risk/framework/ai-sdk)
- [Mastra AI](https://composio.dev/toolkits/_21risk/framework/mastra-ai)
- [LlamaIndex](https://composio.dev/toolkits/_21risk/framework/llama-index)
- [CrewAI](https://composio.dev/toolkits/_21risk/framework/crew-ai)

## TL;DR

Here's what you'll learn:
- Get and set up your OpenAI and Composio API keys
- Install the necessary dependencies
- Initialize Composio and create a Tool Router session for  21risk
- Configure an AI agent that can use  21risk as a tool
- Run a live chat session where you can ask the agent to perform  21risk operations

## What is OpenAI Agents SDK?

The OpenAI Agents SDK is a lightweight framework for building AI agents that can use tools and maintain conversation state. It provides a simple interface for creating agents with hosted MCP tool support.
Key features include:
- Hosted MCP Tools: Connect to external services through hosted MCP endpoints
- SQLite Sessions: Persist conversation history across interactions
- Simple API: Clean interface with Agent, Runner, and tool configuration
- Streaming Support: Real-time response streaming for interactive applications

## What is the 21risk MCP server, and what's possible with it?

The 21risk MCP server is an implementation of the Model Context Protocol that connects your AI agent and assistants like Claude, Cursor, etc directly to your 21risk account. It provides structured and secure access to your checklists, audits, compliance data, and risk models, so your agent can retrieve reports, analyze compliance status, and streamline audit management on your behalf.
- Automated compliance insights and analytics: Instantly fetch compliance data for sites, categories, or specific questions to support analytics and reporting needs.
- Audit report management: Retrieve draft, published, or scheduled audit reports, enabling your agent to monitor progress or summarize findings.
- Risk model and category exploration: Let your agent list and filter available risk models and categories to assist with compliance checks and risk assessments.
- Monthly item tracking and analysis: Query detailed fact tables of items per month for granular, time-based risk or compliance monitoring.
- Site, organization, and property retrieval: Automatically list sites, organizations, and property details, helping you organize and cross-reference risk and compliance data efficiently.

## Supported Tools

| Tool slug | Name | Description |
|---|---|---|
| `_21RISK_GET_COMPLIANCE` | Get Compliance | Tool to retrieve compliance data for sites, categories, or questions. Use when you need OData-based compliance data for analytics or reporting. |
| `_21RISK_GET_ITEMS` | Get Items (BETA) | Tool to retrieve items (BETA) from the 21RISK OData API. Use when you need a filtered and paged list of items for analytics and reporting. Example: GET_ITEMS($filter="Item Cost gt 100", $top=50). |
| `_21RISK_GET_ITEMS_PER_MONTH` | Get Items Per Month | Tool to retrieve fact table data for ItemsPerMonth, one row per question per site per month. Use when querying monthly item data with OData parameters ($filter, $top, $skip, $select, maxPageSizeInMb). |
| `_21RISK_GET_ORGANIZATIONS` | Get Organizations | Tool to retrieve organizations from the 21RISK OData API. Use when you need to list, filter, or paginate organizations via OData parameters after authentication is confirmed. |
| `_21RISK_GET_PROPERTIES` | Get Properties | Tool to fetch a list of properties related to sites, including COPE information and other relevant data. Use when you need property insurance details via OData API after authentication. |
| `_21RISK_GET_REPORTS` | Get Reports | Tool to retrieve audit reports, including draft, published, and scheduled reports. Use when you need a paginated list of reports with optional OData filtering. |
| `_21RISK_GET_RISKMODEL_CATEGORIES` | Get RiskModel Categories | Tool to retrieve risk model categories for grouping questions and compliance checks. Use when you need to filter, select, or paginate risk model categories via OData parameters ($filter, $select, $orderby, $top, $skip, $count). |
| `_21RISK_GET_RISK_MODELS` | Get Risk Models | Tool to retrieve risk models used for audits and compliance. Use when you need to list available risk models with optional OData queries. |

## Supported Triggers

None listed.

## Creating MCP Server - Stand-alone vs Composio SDK

The  21risk MCP server is an implementation of the Model Context Protocol that connects your AI agent to  21risk. It provides structured and secure access so your agent can perform  21risk operations on your behalf through a secure, permission-based interface.
With Composio's managed implementation, you don't have to create your own developer app. For production, if you're building an end product, we recommend using your own credentials. The managed server helps you prototype fast and go from 0-1 faster.

## Step-by-step Guide

### 1. Prerequisites

Before starting, make sure you have:
- Composio API Key and OpenAI API Key
- Primary know-how of OpenAI Agents SDK
- A live  21risk project
- Some knowledge of Python or Typescript

### 1. Getting API Keys for OpenAI and Composio

OpenAI API Key
- Go to the [OpenAI dashboard](https://platform.openai.com/settings/organization/api-keys) and create an API key. You'll need credits to use the models, or you can connect to another model provider.
- Keep the API key safe.
Composio API Key
- Log in to the [Composio dashboard](https://dashboard.composio.dev?utm_source=toolkits&utm_medium=framework_docs).
- Go to Settings and copy your API key.

### 2. Install dependencies

Install the Composio SDK and the OpenAI Agents SDK.
```python
pip install composio_openai_agents openai-agents python-dotenv
```

```typescript
npm install @composio/openai-agents @openai/agents dotenv
```

### 3. Set up environment variables

Create a .env file and add your OpenAI and Composio API keys.
```bash
OPENAI_API_KEY=sk-...your-api-key
COMPOSIO_API_KEY=your-api-key
USER_ID=composio_user@gmail.com
```

### 4. Import dependencies

What's happening:
- You're importing all necessary libraries.
- The Composio and OpenAIAgentsProvider classes are imported to connect your OpenAI agent to Composio tools like  21risk.
```python
import asyncio
import os
from dotenv import load_dotenv

from composio import Composio
from composio_openai_agents import OpenAIAgentsProvider
from agents import Agent, Runner, HostedMCPTool, SQLiteSession
```

```typescript
import 'dotenv/config';
import { Composio } from '@composio/core';
import { OpenAIAgentsProvider } from '@composio/openai-agents';
import { Agent, hostedMcpTool, run, OpenAIConversationsSession } from '@openai/agents';
import * as readline from 'readline';
```

### 5. Set up the Composio instance

No description provided.
```python
load_dotenv()

api_key = os.getenv("COMPOSIO_API_KEY")
user_id = os.getenv("USER_ID")

if not api_key:
    raise RuntimeError("COMPOSIO_API_KEY is not set. Create a .env file with COMPOSIO_API_KEY=your_key")

# Initialize Composio
composio = Composio(api_key=api_key, provider=OpenAIAgentsProvider())
```

```typescript
dotenv.config();

const composioApiKey = process.env.COMPOSIO_API_KEY;
const userId = process.env.USER_ID;

if (!composioApiKey) {
  throw new Error('COMPOSIO_API_KEY is not set. Create a .env file with COMPOSIO_API_KEY=your_key');
}
if (!userId) {
  throw new Error('USER_ID is not set');
}

// Initialize Composio
const composio = new Composio({
  apiKey: composioApiKey,
  provider: new OpenAIAgentsProvider(),
});
```

### 6. Create a Tool Router session

What is happening:
- You give the Tool Router the user id and the toolkits you want available. Here, it is only _21risk.
- The router checks the user's  21risk connection and prepares the MCP endpoint.
- The returned session.mcp.url is the MCP URL that your agent will use to access  21risk.
- This approach keeps things lightweight and lets the agent request  21risk tools only when needed during the conversation.
```python
# Create a  21risk Tool Router session
session = composio.create(
    user_id=user_id,
    toolkits=["_21risk"]
)

mcp_url = session.mcp.url
```

```typescript
// Create Tool Router session for  21risk
const session = await composio.create(userId as string, {
  toolkits: ['_21risk'],
});
const mcpUrl = session.mcp.url;
```

### 7. Configure the agent

No description provided.
```python
# Configure agent with MCP tool
agent = Agent(
    name="Assistant",
    model="gpt-5",
    instructions=(
        "You are a helpful assistant that can access  21risk. "
        "Help users perform  21risk operations through natural language."
    ),
    tools=[
        HostedMCPTool(
            tool_config={
                "type": "mcp",
                "server_label": "tool_router",
                "server_url": mcp_url,
                "headers": {"x-api-key": api_key},
                "require_approval": "never",
            }
        )
    ],
)
```

```typescript
// Configure agent with MCP tool
const agent = new Agent({
  name: 'Assistant',
  model: 'gpt-5',
  instructions:
    'You are a helpful assistant that can access  21risk. Help users perform  21risk operations through natural language.',
  tools: [
    hostedMcpTool({
      serverLabel: 'tool_router',
      serverUrl: mcpUrl,
      headers: { 'x-api-key': composioApiKey },
      requireApproval: 'never',
    }),
  ],
});
```

### 8. Start chat loop and handle conversation

No description provided.
```python
print("\nComposio Tool Router session created.")

chat_session = SQLiteSession("conversation_openai_toolrouter")

print("\nChat started. Type your requests below.")
print("Commands: 'exit', 'quit', or 'q' to end\n")

async def main():
    try:
        result = await Runner.run(
            agent,
            "What can you help me with?",
            session=chat_session
        )
        print(f"Assistant: {result.final_output}\n")
    except Exception as e:
        print(f"Error: {e}\n")

    while True:
        user_input = input("You: ").strip()
        if user_input.lower() in {"exit", "quit", "q"}:
            print("Goodbye!")
            break

        result = await Runner.run(
            agent,
            user_input,
            session=chat_session
        )
        print(f"Assistant: {result.final_output}\n")

asyncio.run(main())
```

```typescript
// Keep conversation state across turns
const conversationSession = new OpenAIConversationsSession();

// Simple CLI
const rl = readline.createInterface({
  input: process.stdin,
  output: process.stdout,
  prompt: 'You: ',
});

console.log('\nComposio Tool Router session created.');
console.log('\nChat started. Type your requests below.');
console.log("Commands: 'exit', 'quit', or 'q' to end\n");

try {
  const first = await run(agent, 'What can you help me with?', { session: conversationSession });
  console.log(`Assistant: ${first.finalOutput}\n`);
} catch (e) {
  console.error('Error:', e instanceof Error ? e.message : e, '\n');
}

rl.prompt();

rl.on('line', async (userInput) => {
  const text = userInput.trim();

  if (['exit', 'quit', 'q'].includes(text.toLowerCase())) {
    console.log('Goodbye!');
    rl.close();
    process.exit(0);
  }

  if (!text) {
    rl.prompt();
    return;
  }

  try {
    const result = await run(agent, text, { session: conversationSession });
    console.log(`\nAssistant: ${result.finalOutput}\n`);
  } catch (e) {
    console.error('Error:', e instanceof Error ? e.message : e, '\n');
  }

  rl.prompt();
});

rl.on('close', () => {
  console.log('\n👋 Session ended.');
  process.exit(0);
});
```

## Complete Code

```python
import asyncio
import os
from dotenv import load_dotenv

from composio import Composio
from composio_openai_agents import OpenAIAgentsProvider
from agents import Agent, Runner, HostedMCPTool, SQLiteSession

load_dotenv()

api_key = os.getenv("COMPOSIO_API_KEY")
user_id = os.getenv("USER_ID")

if not api_key:
    raise RuntimeError("COMPOSIO_API_KEY is not set. Create a .env file with COMPOSIO_API_KEY=your_key")

# Initialize Composio
composio = Composio(api_key=api_key, provider=OpenAIAgentsProvider())

# Create Tool Router session
session = composio.create(
    user_id=user_id,
    toolkits=["_21risk"]
)
mcp_url = session.mcp.url

# Configure agent with MCP tool
agent = Agent(
    name="Assistant",
    model="gpt-5",
    instructions=(
        "You are a helpful assistant that can access  21risk. "
        "Help users perform  21risk operations through natural language."
    ),
    tools=[
        HostedMCPTool(
            tool_config={
                "type": "mcp",
                "server_label": "tool_router",
                "server_url": mcp_url,
                "headers": {"x-api-key": api_key},
                "require_approval": "never",
            }
        )
    ],
)

print("\nComposio Tool Router session created.")

chat_session = SQLiteSession("conversation_openai_toolrouter")

print("\nChat started. Type your requests below.")
print("Commands: 'exit', 'quit', or 'q' to end\n")

async def main():
    try:
        result = await Runner.run(
            agent,
            "What can you help me with?",
            session=chat_session
        )
        print(f"Assistant: {result.final_output}\n")
    except Exception as e:
        print(f"Error: {e}\n")

    while True:
        user_input = input("You: ").strip()
        if user_input.lower() in {"exit", "quit", "q"}:
            print("Goodbye!")
            break

        result = await Runner.run(
            agent,
            user_input,
            session=chat_session
        )
        print(f"Assistant: {result.final_output}\n")

asyncio.run(main())
```

```typescript
import 'dotenv/config';
import { Composio } from '@composio/core';
import { OpenAIAgentsProvider } from '@composio/openai-agents';
import { Agent, hostedMcpTool, run, OpenAIConversationsSession } from '@openai/agents';
import * as readline from 'readline';

const composioApiKey = process.env.COMPOSIO_API_KEY;
const userId = process.env.USER_ID;

if (!composioApiKey) {
  throw new Error('COMPOSIO_API_KEY is not set. Create a .env file with COMPOSIO_API_KEY=your_key');
}
if (!userId) {
  throw new Error('USER_ID is not set');
}

// Initialize Composio
const composio = new Composio({
  apiKey: composioApiKey,
  provider: new OpenAIAgentsProvider(),
});

async function main() {
  // Create Tool Router session
  const session = await composio.create(userId as string, {
    toolkits: ['_21risk'],
  });
  const mcpUrl = session.mcp.url;

  // Configure agent with MCP tool
  const agent = new Agent({
    name: 'Assistant',
    model: 'gpt-5',
    instructions:
      'You are a helpful assistant that can access  21risk. Help users perform  21risk operations through natural language.',
    tools: [
      hostedMcpTool({
        serverLabel: 'tool_router',
        serverUrl: mcpUrl,
        headers: { 'x-api-key': composioApiKey },
        requireApproval: 'never',
      }),
    ],
  });

  // Keep conversation state across turns
  const conversationSession = new OpenAIConversationsSession();

  // Simple CLI
  const rl = readline.createInterface({
    input: process.stdin,
    output: process.stdout,
    prompt: 'You: ',
  });

  console.log('\nComposio Tool Router session created.');
  console.log('\nChat started. Type your requests below.');
  console.log("Commands: 'exit', 'quit', or 'q' to end\n");

  try {
    const first = await run(agent, 'What can you help me with?', { session: conversationSession });
    console.log(`Assistant: ${first.finalOutput}\n`);
  } catch (e) {
    console.error('Error:', e instanceof Error ? e.message : e, '\n');
  }

  rl.prompt();

  rl.on('line', async (userInput) => {
    const text = userInput.trim();

    if (['exit', 'quit', 'q'].includes(text.toLowerCase())) {
      console.log('Goodbye!');
      rl.close();
      process.exit(0);
    }

    if (!text) {
      rl.prompt();
      return;
    }

    try {
      const result = await run(agent, text, { session: conversationSession });
      console.log(`\nAssistant: ${result.finalOutput}\n`);
    } catch (e) {
      console.error('Error:', e instanceof Error ? e.message : e, '\n');
    }

    rl.prompt();
  });

  rl.on('close', () => {
    console.log('\nSession ended.');
    process.exit(0);
  });
}

main().catch((err) => {
  console.error('Fatal error:', err);
  process.exit(1);
});
```

## Conclusion

This was a starter code for integrating  21risk MCP with OpenAI Agents SDK to build a functional AI agent that can interact with  21risk.
Key features:
- Hosted MCP tool integration through Composio's Tool Router
- SQLite session persistence for conversation history
- Simple async chat loop for interactive testing
You can extend this by adding more toolkits, implementing custom business logic, or building a web interface around the agent.

## How to build 21risk MCP Agent with another framework

- [Claude Agent SDK](https://composio.dev/toolkits/_21risk/framework/claude-agents-sdk)
- [Claude Code](https://composio.dev/toolkits/_21risk/framework/claude-code)
- [Claude Cowork](https://composio.dev/toolkits/_21risk/framework/claude-cowork)
- [Codex](https://composio.dev/toolkits/_21risk/framework/codex)
- [OpenClaw](https://composio.dev/toolkits/_21risk/framework/openclaw)
- [Hermes](https://composio.dev/toolkits/_21risk/framework/hermes-agent)
- [CLI](https://composio.dev/toolkits/_21risk/framework/cli)
- [Google ADK](https://composio.dev/toolkits/_21risk/framework/google-adk)
- [LangChain](https://composio.dev/toolkits/_21risk/framework/langchain)
- [Vercel AI SDK](https://composio.dev/toolkits/_21risk/framework/ai-sdk)
- [Mastra AI](https://composio.dev/toolkits/_21risk/framework/mastra-ai)
- [LlamaIndex](https://composio.dev/toolkits/_21risk/framework/llama-index)
- [CrewAI](https://composio.dev/toolkits/_21risk/framework/crew-ai)

## Related Toolkits

- [Excel](https://composio.dev/toolkits/excel) - Microsoft Excel is a robust spreadsheet application for organizing, analyzing, and visualizing data. It's the go-to tool for calculations, reporting, and flexible data management.
- [Abstract](https://composio.dev/toolkits/abstract) - Abstract provides a suite of APIs for automating data validation and enrichment tasks. It helps developers streamline workflows and ensure data quality with minimal effort.
- [Addressfinder](https://composio.dev/toolkits/addressfinder) - Addressfinder is a data quality platform for verifying addresses, emails, and phone numbers. It helps you ensure accurate customer and contact data every time.
- [Agentql](https://composio.dev/toolkits/agentql) - Agentql is a toolkit that connects AI agents to the web using a specialized query language. It enables structured web interaction and data extraction for smarter automations.
- [Agenty](https://composio.dev/toolkits/agenty) - Agenty is a web scraping and automation platform for extracting data and automating browser tasks—no coding needed. It streamlines data collection, monitoring, and repetitive online actions.
- [Ambee](https://composio.dev/toolkits/ambee) - Ambee is an environmental data platform providing real-time, hyperlocal APIs for air quality, weather, and pollen. Get precise environmental insights to power smarter decisions in your apps and workflows.
- [Ambient weather](https://composio.dev/toolkits/ambient_weather) - Ambient Weather is a platform for personal weather stations with a robust API for accessing local, real-time, and historical weather data. Get detailed environmental insights directly from your own sensors for smarter apps and automations.
- [Anonyflow](https://composio.dev/toolkits/anonyflow) - Anonyflow is a service for encryption-based data anonymization and secure data sharing. It helps organizations meet GDPR, CCPA, and HIPAA data privacy compliance requirements.
- [Api ninjas](https://composio.dev/toolkits/api_ninjas) - Api ninjas offers 120+ public APIs spanning categories like weather, finance, sports, and more. Developers use it to supercharge apps with real-time data and actionable endpoints.
- [Api sports](https://composio.dev/toolkits/api_sports) - Api sports is a comprehensive sports data platform covering 2,000+ competitions with live scores and 15+ years of stats. Instantly access up-to-date sports information for analysis, apps, or chatbots.
- [Apify](https://composio.dev/toolkits/apify) - Apify is a cloud platform for building, deploying, and managing web scraping and automation tools called Actors. It lets you automate data extraction and workflow tasks at scale—no infrastructure headaches.
- [Autom](https://composio.dev/toolkits/autom) - Autom is a lightning-fast search engine results data platform for Google, Bing, and Brave. Developers use it to access fresh, low-latency SERP data on demand.
- [Beaconchain](https://composio.dev/toolkits/beaconchain) - Beaconchain is a real-time analytics platform for Ethereum 2.0's Beacon Chain. It provides detailed insights into validators, blocks, and overall network performance.
- [Big data cloud](https://composio.dev/toolkits/big_data_cloud) - BigDataCloud provides APIs for geolocation, reverse geocoding, and address validation. Instantly access reliable location intelligence to enhance your applications and workflows.
- [Bigpicture io](https://composio.dev/toolkits/bigpicture_io) - BigPicture.io offers APIs for accessing detailed company and profile data. Instantly enrich your applications with up-to-date insights on 20M+ businesses.
- [Bitquery](https://composio.dev/toolkits/bitquery) - Bitquery is a blockchain data platform offering indexed, real-time, and historical data from 40+ blockchains via GraphQL APIs. Get unified, reliable access to complex on-chain data for analytics, trading, and research.
- [Brightdata](https://composio.dev/toolkits/brightdata) - Brightdata is a leading web data platform offering advanced scraping, SERP APIs, and anti-bot tools. It lets you collect public web data at scale, bypassing blocks and friction.
- [Builtwith](https://composio.dev/toolkits/builtwith) - BuiltWith is a web technology profiler that uncovers the technologies powering any website. Gain actionable insights into analytics, hosting, and content management stacks for smarter research and lead generation.
- [Byteforms](https://composio.dev/toolkits/byteforms) - Byteforms is an all-in-one platform for creating forms, managing submissions, and integrating data. It streamlines workflows by centralizing form data collection and automation.
- [Cabinpanda](https://composio.dev/toolkits/cabinpanda) - Cabinpanda is a data collection platform for building and managing online forms. It helps streamline how you gather, organize, and analyze responses.

## Frequently Asked Questions

### What are the differences in Tool Router MCP and 21risk MCP?

With a standalone  21risk MCP server, the agents and LLMs can only access a fixed set of  21risk tools tied to that server. However, with the Composio Tool Router, agents can dynamically load tools from  21risk and many other apps based on the task at hand, all through a single MCP endpoint.

### Can I use Tool Router MCP with OpenAI Agents SDK?

Yes, you can. OpenAI Agents SDK fully supports MCP integration. You get structured tool calling, message history handling, and model orchestration while Tool Router takes care of discovering and serving the right  21risk tools.

### Can I manage the permissions and scopes for 21risk while using Tool Router?

Yes, absolutely. You can configure which  21risk scopes and actions are allowed when connecting your account to Composio. You can also bring your own OAuth credentials or API configuration so you keep full control over what the agent can do.

### How safe is my data with Composio Tool Router?

All sensitive data such as tokens, keys, and configuration is fully encrypted at rest and in transit. Composio is SOC 2 Type 2 compliant and follows strict security practices so your  21risk data and credentials are handled as safely as possible.

---
[See all toolkits](https://composio.dev/toolkits) · [Composio docs](https://docs.composio.dev/llms.txt)