The official Boomi Companion skill for building Boomi integration processes programmatically with AI coding agents. The intended audience of this README.md document is humans seeking to understand the skill.
Important: Boomi Companion is a publicly available developer offering, not an officially supported Boomi product. It is provided as-is and is not covered by Boomi support agreements or SLAs. Boomi curates and maintains this tool on a best-effort basis — treat it as a self-service resource. Boomi reserves the right to modify or discontinue it at any time without notice.
This project is licensed under the BSD-2-Clause License. If you fork or modify this code, you should not use the name "Boomi" for your version.
Found a bug or have a feature idea? Email [email protected] with a clear description, steps to reproduce, and any relevant error messages.
This is a distributable skill that provides AI coding agents with knowledge and tooling for working with Boomi Enterprise platform. It enables:
- Programmatic creation and modification of Boomi components (processes, profiles, connections, operations, topics, subscriptions)
- Bi-directional component push/pull with the Boomi platform API
- Reference documentation for Boomi componentry and development patterns
- CLI tools for deployment, testing, and component management
curl(universally available)jq(install viabrew install jqon macOS orapt install jqon Linux)- Python 3 (only for
boomi-profile-inspect.py— stdlib only, no pip deps)
Install through the Claude Code plugin system — the skill is included automatically:
- Add the Boomi marketplace:
/plugin marketplace add OfficialBoomi/boomi-companion - Install the plugin:
/plugin install bc-integration@boomi-companion
Alternatively, navigate the /plugin menu interactively within Claude Code to add the marketplace and install.
Clone or copy this skill directory into the location your platform uses for agent skills. Consult your platform's documentation for the correct skill directory path.
Once the skill is installed it works in an individual project folder as follows:
your-project/
├── .env # Your credentials (created during setup)
└── active-development/ # All working files (auto-created as needed)
├── processes/ # Process XML files
├── profiles/ # Profile XML files
├── connections/ # Connection XML files
├── operations/ # Operation XML files
├── maps/ # Map XML files
├── document-caches/ # Document cache XML files
├── scripts/ # Script XML files
├── .sync-state/ # Component sync state tracking
└── feedback/ # Test execution results
If using the skill via the bc-integration plugin, there is a series of quality of life setup steps that help template and spin up project workspaces rapidly. See the README.md file for the plugin for more details, or ask your AI agent for help.
A Boomi development project (i.e. the folder in which you are working, not this skill directory itself) will require a .env file in your project root with:
# Platform API Credentials (required)
BOOMI_API_URL=https://your-platform.boomi.com
BOOMI_USERNAME=your_username
BOOMI_API_TOKEN=your_api_token
BOOMI_ACCOUNT_ID=your_account_id
BOOMI_VERIFY_SSL=true
# Default Folder (if no folder is specified, the agent will put projects here - optional).
BOOMI_TARGET_FOLDER=your_default_folder_guid
# Environment and Runtime Details (lets the agent deploy processes - optional)
BOOMI_ENVIRONMENT_ID=your_environment_id
BOOMI_TEST_ATOM_ID=your_test_atom_id
# Shared Web Server Runtime Credentials (lets the agent test listeners - optional)
SERVER_BASE_URL=https://your-atom.integrate.boomi.com
SERVER_USERNAME=your_runtime_username
SERVER_TOKEN=your_runtime_token
SERVER_VERIFY_SSL=false
Where to find these:
- API credentials: Boomi platform → Account settings → API Management
- Folder GUIDs: Create folders via GUI or use
boomi-folder-create.sh - Environment/Atom IDs: Boomi platform → Environment management
Your AI agent has more info about the necessary pieces of data and can talk you through them one by one - just ask!
Once the skill is installed, you can work with your AI coding agent to build Boomi processes:
- Start a conversation: Describe what you want to build
- The agent loads the skill: References comprehensive Boomi documentation
- Programmatic development: The agent creates/modifies XML components
- Platform sync: The agent uses custom CLI tools to push/pull components to/from Boomi
- Deploy and test: The agent can also use its tools to deploy to a runtime and test your processes
You: "Create a REST API endpoint that fetches weather data and returns JSON"
Agent: [Loads BOOMI_THINKING.md and relevant references]
[Creates project folder]
[Creates JSON profiles for request/response]
[Pushes profile components to platform]
[Creates REST connection and operation]
[Pushes connection and operation components to platform]
[Creates WSS operation for endpoint]
[Creates process with all steps]
[Pushes all components to platform]
You: "Deploy and test it"
Agent: [Deploys process to runtime]
[Runs a curl command to test, or provides you the details to test yourself]
The skill makes the following CLI tools available to the agent:
boomi-env-check.sh- Checks which .env variables are set without revealing valuesboomi-folder-create.sh- Creates project foldersboomi-component-create.sh- Creates new componentsboomi-component-push.sh- Updates existing componentsboomi-component-pull.sh- Downloads components from platformboomi-deploy.sh- Packages and deploys processes to runtimeboomi-undeploy.sh- Undeploys processes from runtimeboomi-test-execute.sh- Executes and tests processes via platform APIboomi-wss-test.sh- Tests WSS listener endpoints via the shared web serverboomi-profile-inspect.py- Extracts field metadata from large profiles (Python stdlib)event-streams-setup.sh- Configures Event Streams
The skill includes the following Boomi-centric reference documentation:
BOOMI_THINKING.md- Essential mental models (always read first)references/guides/boomi_error_reference.md- Error patterns, silent failures, and troubleshootingreferences/guides/boomi_platform_reference.md- Platform services and boundaries
components/- Detailed specs for all component types- Profiles (JSON, XML, Flat File)
- Connections (REST, Salesforce, Event Streams)
- Operations (Connectors, WSS)
- Maps, Processes, Scripts
steps/- Process step documentation with examples- REST Connector, Map, Message, Set Properties
- Branch, Decision, Try-Catch
- Process Call, Return Documents
- Event Streams, Salesforce
This skill is designed originally for Claude Code, but Agent Skills are an open standard accessible to other models and platforms.
More info about agent skills can be found here: https://agentskills.io/home and here: https://platform.claude.com/docs/en/agents-and-tools/agent-skills/overview
If you encounter issues:
- This course provides an excellent intro to Claude Code: https://anthropic.skilljar.com/claude-code-in-action
- We would love your feedback and input via [email protected]
- Your AI agent can often help troubleshoot and explain issues