The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill uses a wrapper script run.js that takes JavaScript code (from files, arguments, or stdin), wraps it in an asynchronous template, writes it to a temporary file (.temp-execution-*.js), and executes it using the Node.js require() function.
[EXTERNAL_DOWNLOADS]: The package.json setup script performs npm install and npx playwright install chromium, which downloads the Playwright library and official browser binaries from Microsoft's infrastructure. These are trusted, well-known sources for this functionality.
[COMMAND_EXECUTION]: The skill executes shell commands via execSync in run.js to manage the installation of dependencies and browser binaries.
[INDIRECT_PROMPT_INJECTION]: As a browser automation tool, the skill ingests data from arbitrary external websites. This creates a surface where instructions embedded in a website's content (e.g., in HTML comments or hidden text) could potentially influence the agent's behavior during a browsing session.
Ingestion points: Website content accessed via page.goto() and processed through Playwright locators or page.evaluate().
Boundary markers: The skill does not implement explicit boundary markers or sanitization for website content before the agent processes it.
Capability inventory: The agent has the ability to write files to /tmp (screenshots/scripts), perform network requests, and execute JavaScript code via the run.js tool.
Sanitization: No specific sanitization logic is present for data extracted from web pages.

playwright-skill