guides/
10 pages · Updated June 14, 2026
Pages
- AI agents need the exact element context behind the screenshot, not the screenshot alone.
- CSS selector vs getByRole vs getByTestId is really a question about UI contracts.
- Playwright locators break when the team picks them without enough UI context.
- Stop flaky Playwright tests by fixing the selector decision, not just the syntax.
- To reproduce a frontend bug well, capture the exact UI artifact behind it.
- data-testid works best when the team treats it like a contract.
- Browser agents need exact UI context if they are going to touch real web apps safely.
- Element context via MCP only matters if the context is exact enough to use.
- getByRole vs CSS selector depends on how strong the UI contract actually is.
- Guides for Playwright Locators, Selectors, Bug Reproduction, and Element Context