Claude Code shipping loop: enumerate the problem space and grill the plan, design the system (for UI work), implement, then run a definition-of-done gate (verify → simplify → independent review → structure review → runtime smoke → commit) before calling work complete. On first use it detects and offers to install the Claude Code tools each stage needs — the Codex review plugin, grill-with-docs, structural-code-review, frontend-visual-qa. Use when starting or finishing a meaningful unit of implementation work: a phase, feature slice, or milestone.
Extract documentation from rendered SPA or app-driven docs sites into local Markdown under a site-named docs folder. Use when docs live in an app UI or docs website with no source Markdown available, and an agent must drive a real browser for route discovery, DOM extraction, content cleanup, and link rewriting. Works with Playwright MCP, Chrome DevTools MCP, or any browser automation.
Require real-browser visual QA before finalizing frontend work: apps, websites, dashboards, games, components, or design changes. Use when an agent builds or modifies UI and must confirm it actually renders well with desktop and mobile screenshots, responsive layout checks, overflow and clipping checks, loading/empty/error states, keyboard navigation and focus visibility, real assets, design-system consistency, and no decorative clutter.
Review code changes, pull requests, branches, or implementation plans for structural maintainability. Use when the review should be strict about architecture, abstraction quality, module boundaries, repeated logic, large-file growth, special-case branching, type contracts, service/action ownership, spaghetti control flow, or code that works but makes the codebase harder to evolve.