The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill documents the installation of the Playwright browser automation framework and its system dependencies using official CLI commands and GitHub Actions.\n
Evidence: npx playwright install --with-deps chromium (references/e2e/ci.md), actions/checkout@v4 (references/e2e/ci.md)\n
Context: These are official installation methods for well-known and trusted services.\n- [COMMAND_EXECUTION]: The skill provides instructions for running Playwright tests and managing development workflows through standard CLI tools.\n
Evidence: npx playwright test, npm ci, npx playwright codegen (references/e2e/ci.md, references/e2e/config.md)\n
Context: Standard project management and testing commands for JavaScript environments.\n- [PROMPT_INJECTION]: The skill facilitates web scraping, which involves ingesting untrusted content from external websites, creating a surface for indirect prompt injection.\n
Ingestion points: page.content(), page.locator().textContent(), and page.evaluate() usage in references/scraping/extraction.md and references/scraping/network.md.\n
Boundary markers: None present in instructional code snippets.\n
Capability inventory: Browser script execution (page.evaluate), file system writing (fs.writeFile), and file downloads (download.saveAs) as seen in references/scraping/extraction.md and references/scraping/network.md.\n
Sanitization: Not explicitly implemented in the instructional snippets, which focus on extraction techniques.\n
Note: This risk factor is inherent to the skill's primary purpose of web scraping and is documented neutrally.

playwright