Expect integrates with AI coding agents like Claude Code, Codex, and Cursor to automatically test their generated code. It reads code changes, generates a test plan, and executes it in a real browser using Playwright. The tool checks for performance metrics (INP, LCP), security vulnerabilities, design regressions, and app completeness. It spawns subagents simulating real logged-in users to find issues, then reports back so the original agent can fix them. Supports both local development and CI environments.