What problem does it solve? When an AI agent writes the code, the human operator stops building the mental model that writing by hand used to provide. This Skill produces dense, repo-grounded explainers so the developer keeps understanding what was built and can make the next decision themselves. ## Core Features & Use Cases - Four explainer modes: teach a concept new to the codebase, walk a diff or PR by intent, make an idea or plan legible, or reconstruct a window of recent work. - Reader-targeted writing: names the reader's level and existing knowledge, grounds every claim in real files, functions, and numbers from the actual repo, and teaches one load-bearing idea per explainer. - Optional active recall check-in: offers predict-then-reveal or checked exercises where the prediction is always collected before any explanation is revealed. - Use Case: After the agent adds a background job queue, ask for an explanation and receive a concrete walkthrough of the actual queue code, plus an optional quiz question that tests whether you can predict its behavior before seeing the answer. ## Quick Start Ask the agent to explain how the new background job queue works, noting that you are comfortable with the app but new to queues.
