That's a great question! Once you install the skills, they activate automatically based on your prompt.
Here's how it works: each skill has a description with trigger conditions, like "use when designing a GraphQL schema" or "use when writing GraphQL queries." The agent reads these at startup and loads the full instructions whenever it detects a relevant task. So, if you ask it to "design a schema for users and orders," it will automatically pull in the graphql-schema skill.
Some agents, like Claude Code, also support explicit slash commands if you want to invoke a skill directly, but that's not necessary.
