The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill utilizes npx within the scripts/playwright_cli.sh wrapper script to download and execute the @playwright/cli package from the official NPM registry.
[COMMAND_EXECUTION]: The wrapper script scripts/playwright_cli.sh executes shell commands to interface with the Playwright CLI tool.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from the web, which creates a surface for indirect prompt injection attacks.
Ingestion points: External web content is ingested through commands like pwcli open, pwcli snapshot, and pwcli eval (e.g., references/cli.md).
Boundary markers: Absent. The instructions do not provide explicit delimiters or warnings to the agent to disregard instructions found within the automated browser's context.
Capability inventory: The skill allows browser-based actions including navigation, form submission, screenshot capture, and execution of arbitrary JavaScript within the browser context via the eval and run-code commands.
Sanitization: Absent. There is no evidence of filtering or sanitizing web content before it is processed by the agent.

playwright