Design features, subsystems, and substantial refactors through component boundaries, caller usage, data models, and interfaces before implementation. Use when asked to architect a change, choose a design, or decide where code and responsibilities should live.
Present concise plans, progress updates, findings, and final reports that keep the human oriented, informed, and able to learn how the project works. Use for complex agent coding work and focus on the most important points and decisions.
Generate a project-local verification skill that drives your app the way a user does — any language, framework, or platform. Use for $create-verification-skill, "make a control skill for this repo", or when a project has no scripted way to prove…
Periodic pass that keeps a project's verification skill and feature map honest: parallel source readers per feature, one live session driving every feature, at most one PR of proven corrections. Use for $maintain-verification-skill or "audit the verify skill".
Record a small MP4 while demonstrating or verifying a UI change, then attach the video as proof. Linux cloud sessions record the private virtual display; macOS records only the app window under test. Use for any repo whenever a UI change is verified or demoed.
Help the user build a working mental model of code, system designs, and proposed changes. Use when asked to teach, explain how a system works, walk through an architecture, or understand a change.
Review local changes, branches, commits, and PRs with independent native subagents. Use for code-review requests or substantive implementation closeout; focus on concrete defects, simplicity, and the task's goal.
Create a self-contained HTML explanation when the user requests one or when important, durable project knowledge is materially easier to understand visually than in concise chat. Use for significant architecture, research, decisions, incidents, audits, or…