fly-machines

What This Skill Does

The fly-machines skill provides an interface for interacting with the Fly.io Machines API. It allows OpenClaw to programmatically deploy, monitor, and manage containerized applications directly on the Fly.io infrastructure. The skill enables the agent to automate the lifecycle of cloud workloads, including listing existing applications, creating new ones, spinning up individual Machine instances, and managing their state (starting, stopping, and deleting). Additionally, it supports volume management, allowing the agent to handle persistent storage for stateful applications.

Installation

To install this skill, run the following command in your terminal: clawhub install openclaw/skills/skills/mrgoodb/fly-machines

Ensure you have obtained your personal access token from the Fly.io dashboard and configured it either via the ~/.config/fly/token file or by setting the FLY_API_TOKEN environment variable before invoking the agent.

Use Cases

• Automated Deployment: Automatically spin up temporary containers to run background workers or periodic cron jobs.
• Scalable Hosting: Manage multi-tenant workloads by dynamically creating new Machines for specific customers or application environments.
• Monitoring and Maintenance: Query machine status and health, ensuring that your background services are running in the desired region.
• CI/CD Integration: Integrate with automated pipelines to tear down or rotate infrastructure after specific deployment steps are completed.

Example Prompts

1. "Check the current status of the 'production-api' app and tell me how many machines are running in the 'iad' region."
2. "Create a new machine for the 'web-service' app using the 'nginx:latest' image with 512MB of memory in the 'sea' region."
3. "Stop the machine with ID 'abc12345' in my 'demo-app' and wait for it to report a 'stopped' state."

Tips & Limitations

• Authentication: Always verify your token permissions. Using an environment variable is often the safest and easiest way to ensure the agent has access during runtime.
• State Management: When deleting machines, be mindful of data persistence. Always check for existing volumes if you need to retain data across deployments.
• Wait Times: Always utilize the /wait endpoint if you need to ensure a machine has reached a specific state before proceeding with subsequent operations to avoid race conditions.
• Limitations: This skill does not manage network certificates or DNS routing directly; it focuses solely on Machine and App lifecycle operations via the API.