Tools let AI agents interact with your infrastructure - from AWS and Kubernetes to GitHub and Slack. Tools Marketplace Grid The tools marketplace provides a comprehensive catalog of available integrations across different categories.

​
Built-in tools

Composer comes with 100+ tools ready to use: Cloud platforms:
  • AWS, Azure, Google Cloud
  • Deploy, scale, monitor resources
  • Manage IAM and security settings
Development:
  • GitHub, GitLab, Bitbucket
  • Create PRs, run CI/CD, manage repos
  • Code analysis and testing
Infrastructure:
  • Kubernetes, Docker, Terraform
  • Deploy containers, manage clusters
  • Infrastructure as code
Communication:
  • Slack, Microsoft Teams, Email
  • Send notifications, create channels
  • Incident response coordination
Databases:
  • PostgreSQL, MySQL, MongoDB
  • Query data, manage schemas
  • Backup and restore operations

​
Connecting new tools

  1. Go to Resources → New Stack in the left sidebar
  2. Choose your service from the integration catalog
  3. Authenticate - usually OAuth or API key
  4. Test the connection to make sure it works
  5. Start using - agents can now access this tool
IAC Stacks Overview

​
How agents use tools

When you chat with an agent:
  1. Agent understands what you want to accomplish
  2. Selects appropriate tools based on your request
  3. Executes safely with proper error handling
  4. Shows results and suggests follow-up actions
Example: “Deploy my app to staging”
  • Agent uses GitHub tool to fetch latest code
  • Uses Docker tool to build container image
  • Uses Kubernetes tool to deploy to staging cluster
  • Uses Slack tool to notify team of deployment

​
Custom tools

For standard operations: Most teams find the built-in tools sufficient. For specialized needs: You can create custom tools using:
  • Docker containers - Package any script or application
  • API endpoints - Connect to internal services
  • Shell commands - Wrap existing scripts
  • Python functions - Direct code integration
Lambda Function Configuration Above is an example of configuring a custom Lambda function tool, showing the required parameters like function name, runtime, role ARN, and handler.

​
Managing tool access

Permissions: Control which tools each team member can access. Secrets: Store API keys and credentials securely. Usage tracking: See which tools are used most often. Health monitoring: Get alerts when tools become unavailable.
Getting started? Connect one tool you use daily first - like GitHub or AWS. Once you see how it works, add more.