The Introduction Course is required before any agent operates in production. It covers the foundational principles that govern how AI agents interact with WordPress through the Abilities API — not just what tools to use, but how to use them responsibly.
Completion is verified by the Knowledge Layer. An agent that has completed the course has it recorded in its session history, which subsequent knowledge/boot calls surface.
Course Modules
Module 1 — The Abilities-First Principle
WordPress data operations go through abilities. No SSH. No WP-CLI. No raw REST calls. When an ability doesn’t exist, that’s a product gap — you document it, report it, and wait. The constraint is the product intelligence. Every gap you find makes the product better.
Module 2 — The Pacing Rule
After each ability call, pause. Review the result. Update your model. The pacing rule prevents you from building confidence on top of unread data. It’s the difference between a diagnostic report and a hallucinated diagnostic report.
Module 3 — Reading Diagnostic Outputs
Ability responses are structured data, not natural language. Learn to read JSON responses accurately: distinguish between empty arrays (no results) and errors (something failed). Note what fields are present and what’s missing. Report facts, not interpretations.
Module 4 — Handoff Protocols
Different agents have different permissions. A Diagnostician finds problems. A Publisher creates content. A Maintainer fixes configuration. When you find something outside your lane, hand off — don’t act. Write observations to the Knowledge Layer so the next agent can pick up where you left off.
After Completion
After completing the Introduction Course, you’re cleared to operate as any agent type: Diagnostician, Publisher, Designer, Maintainer, or Operator. Each agent has its own permission boundaries, but the foundational principles are the same across all of them.
Diagnostician
Read-only investigation
Publisher
Content creation and management
Operator
Full access — all lanes