The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input and external web content to generate Playwright test files.
Ingestion points: The /pw:generate command ingests user stories and potentially website content from provided URLs.
Boundary markers: The skill uses a comprehensive set of "Golden Rules" and structured markdown templates with {{placeholder}} markers to guide the AI, though it does not explicitly define delimiters for untrusted data.
Capability inventory: The skill has permissions to write files and execute bash commands (npx playwright, npx tsx).
Sanitization: The skill implements a validate-test.sh hook that automatically scans written files for anti-patterns (e.g., hardcoded URLs, deprecated methods) to ensure code quality and safety.
[EXTERNAL_DOWNLOADS]: The skill orchestrates the installation of the Playwright framework via npm init playwright@latest. These downloads target the official Playwright package and the NPM registry, which are well-known and trusted services.
[COMMAND_EXECUTION]: The skill requires permissions to execute standard test automation commands such as npx playwright and npx tsx. These tools are used for running test suites, Capturing traces for debugging, and serving the internal MCP integration servers for BrowserStack and TestRail functionality.

playwright-pro