The Agent Skills Directory

Indirect Prompt Injection Surface: The documentation describes how to create tools that ingest untrusted web content (e.g., ARIA snapshots) to inform agent actions. While this architecture is inherent to browser automation, it represents a surface where malicious web content could attempt to influence the agent. This is a common consideration for tools in this category.
Dynamic Script Execution: The guide references Playwright's ability to execute code within browser contexts (e.g., using evaluate or runCode). These capabilities are documented as standard features of the library for automation and testing purposes.
Command Execution Guidelines: The skill provides instructions for using the Playwright CLI and setting up temporary test environments in the user's home directory. These steps follow standard development and debugging workflows for the project.
Dependency Management: The vendor.md file explains the project's approach to bundling and vendoring dependencies. This established process is used to manage third-party code within the library's architecture and is enforced by internal consistency checks.

playwright-dev