Generate a .cursorrules File for a QA Codebase

XML-tagged — best for Claude 4.x

<role>
You are a Cursor power user. You know that .cursorrules dramatically improves output quality but only if rules are specific to the QA domain — generic rules ("write clean code") add no value.
</role>

<context>
.cursorrules is loaded into every prompt Cursor sends to its underlying model. Good content:
- Locator preferences (which to use, which to avoid)
- Test naming conventions
- Assertion patterns
- Anti-patterns to NEVER produce
- "Ask before guessing" rules (selectors, test data, env vars)
- Framework-specific gotchas (Cypress chaining, Playwright auto-waiting, Vitest mocking semantics)

The file is plain markdown / text. Keep < 150 lines; longer dilutes signal.
</context>

<task>
Generate a `.cursorrules` with these sections:
1. Project context (1-2 sentences)
2. When generating tests
3. When generating POMs
4. Anti-patterns to NEVER produce
5. Ask for clarification rules
6. Style conventions
</task>

<input>
Project framework: {framework}
Language: {language}
Locator strategy: {locator_strategy}
Test data convention: {test_data}
Known anti-patterns from your codebase: {anti_patterns}
</input>

<constraints>
- Rules SPECIFIC, not generic ("Use data-testid first, not class selectors").
- Anti-patterns called out by NAME (waitForTimeout, magic numbers, etc.).
- "Ask before guessing" must be explicit (clarify which scenarios).
- Style conventions named, not implied.
- 80-150 lines total.
</constraints>

<output_format>
A complete `.cursorrules` file as a Markdown code block.
</output_format>

Before writing, identify the 3-5 most common pain points in this team's Cursor output and address them in the file.