Build an orchestrator by writing a SKILL.md file that lists your specialist skills, defines when to invoke each one, and specifies the order and handoff format — no coding required.
The waterfall orchestrator is a multi-step pipeline that takes a YouTube video URL and automatically produces a transcript, tutorial article, email announcement, community post, and social media content in sequence.
The orchestrator passes each agent's output as the next agent's input — research findings go to the writing agent, the written piece goes to the publishing agent — creating a clean sequential pipeline.
Yes — a well-designed orchestrator can route your request to the right specialist agent based on what you ask, acting as a single entry point for your entire AI team.
A waterfall orchestrator that processes a Zoom session recording into a published lesson, a community post, and a newsletter email is a real-world example of an orchestrator agent running inside a teaching business.
An orchestrator agent sequences specialist agents by passing outputs from one as inputs to the next, or by running independent tasks in parallel and then assembling the combined results.
An orchestrator agent manages other agents — it receives a complex task, breaks it into parts, delegates each part to a specialist agent, and assembles the results into a final output.
The first skill every educator should build is a session-recap or lesson-summary skill — it addresses the most universal high-effort task in a teaching business and produces immediate, visible value.
Most course creators get significant automation benefit from 5 to 8 well-built skills covering their most repetitive weekly tasks — content creation, student communication, and community management.
A prompt is a one-time instruction you type; a command is a shortcut that triggers a predefined prompt; a skill is a full reusable workflow with inputs, steps, and defined outputs that persists across sessions.