The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill translates user requests and acceptance criteria into executable Playwright test scripts. This creates a potential surface where malicious instructions embedded in the input could influence the generated code or browser interactions.
Ingestion points: Plan excerpts, acceptance criteria, and user flow requests are processed to generate test logic (SKILL.md, selectors-guide.md).
Boundary markers: The instructions lack explicit delimitation or sanitization rules for requirements before they are interpolated into test files.
Capability inventory: The skill has the ability to write files, execute shell commands (Bash), and control a browser via Playwright.
Sanitization: No formal sanitization or validation of the input requirements is described before they are used to build script logic.
[COMMAND_EXECUTION]: The skill performs shell operations to manage the environment, including checking for installations, creating timestamped directories, and invoking the Playwright runner. These are standard operations for an E2E testing tool.
[DYNAMIC_EXECUTION]: The skill dynamically creates TypeScript or JavaScript test files based on the requested flows and executes them using 'npx playwright test' or 'npx tsx'. This runtime code generation and execution is the primary function of the skill.
[EXTERNAL_DOWNLOADS]: The skill downloads the Chromium browser binary and system-level dependencies using official Playwright commands. These downloads originate from well-known and trusted sources associated with the Microsoft Playwright project.

playwright