What problem does it solve?
This Skill eliminates the risk of accidental, unapproved mutations to Fly.io infrastructure during AI-assisted workflows by enforcing explicit operator invocation and mandatory confirmation gates for all production-side-effect operations.
Core Features & Use Cases
- Explicit invocation only: Prevents the AI from auto-routing phrases like "deploy to Fly" to this skill, requiring operators to type
/flyctl <subcommand> directly for all actions.
- Production safety rails: Enforces typed app-name confirmation for all destructive operations on production apps, with a deliberate
--no-confirm flag for known non-production use cases.
- Comprehensive Fly operations: Supports deploying apps (with image reference validation), tailing logs, managing secrets, inspecting and scaling machines, SSH access, proxying internal services, and running health checks, with automatic app discovery from
fly.toml.
- Use Case: A DevOps operator can safely scale a production Fly app to 3 machines by running
/flyctl scale count 3, with the skill automatically detecting the app from fly.toml, prompting for app name confirmation, and verifying the new machine layout after the operation.
Quick Start
Use the flyctl skill to deploy your latest pre-built container image to your Fly.io app, following the prompted app name confirmation for production environments.