The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/audit.mjs uses the Playwright library to launch and control a browser process (Chrome or Chromium) to perform web audits.
[DYNAMIC_EXECUTION]: The skill uses page.evaluate() to execute the auditing logic contained in scripts/checks.js within the browser context. It also uses dynamic import() to load the Playwright dependency from the local project's directory.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes untrusted data from web pages during audits.
Ingestion points: scripts/audit.mjs navigates to and reads content from provided URLs or local development servers.
Boundary markers: The agent receives structured data in the form of a JSON report containing specific metrics, reducing the risk of instructions in audited pages being misinterpreted as agent commands.
Capability inventory: The skill is capable of browser automation and writing reports/screenshots to the local file system.
Sanitization: The tool primarily extracts numerical metrics and specific attributes (like labels) rather than raw text for direct prompt interpolation.
[EXTERNAL_DOWNLOADS]: The documentation instructs users to install playwright or playwright-core and Chromium via npx playwright install chromium to enable the skill's functionality.

kero-audit