This course dives into using AI agents to develop WordPress code and interact programmatically with WordPress. The first module covers the history and terminology of AI tools and concepts before diving into practical examples that use Anthropic’s Claude to co-author a WordPress plugin. Module 3 details a GitHub Issue-to-Pull Request workflow.
Introduction
Lessons
Course OutlineAI Foundations
Start here for the shared vocabulary and decisions that underpin the rest of the course: what an LLM is and how it actually works, where you’ll meet one in practice, what to consider before sending real data to a hosted API, and when an LLM is the right tool versus a plain script.
Lessons
History and Terminology of Modern AI Tools Accessing LLMs: Interfaces and Hosting Security, Privacy, and Data-Handling Considerations Don't Reinvent the Wheel: LLMs versus ScriptsAgentic WordPress Development with Claude Code
This module walks through the practical Claude surfaces a WordPress developer uses day to day — the terminal agent, the Desktop app’s three modes, the project-level conventions (
CLAUDE.md and settings), MCP for connecting Claude to external systems, the IDE-integrated agent, the Chrome extension for browser-based work in wp-admin, reusable Skills, and finally building your own MCP servers for internal systems. Lessons share a continuous example: a local Studio site (logging-demo) and the vip-learn-ai-workflows-demo plugin that grows across the module.Lessons
Claude in the Terminal: Working With Local WordPress Claude Desktop: Chat, Cowork, and Code CLAUDE.md and Claude Settings MCP Integration With Claude Claude Agent in VS Code Claude in Chrome: Driving WP Admin Claude Skills Creating MCP ServersEnterprise Workflows – Github Issue to Pull Request Workflow
The last module shifts from “you and Claude on your laptop” to “your team’s repo and CI”. It opens with how to plan AI-assisted work before any code gets written, then covers the GitHub issue-to-pull-request workflow, what makes an issue an agent can actually act on, and the execution-to-merge patterns that keep work reviewable. The module follows a proven, open workflow — the agentic-workflow-template, built at Pew Research Center and generalized for any team — and references its files throughout.