The Agent Skills Directory

[COMMAND_EXECUTION]: The skill facilitates the creation and execution of custom Python scripts using Playwright on a remote system via the nodes.run command. While intended for automation, this allows for runtime script generation and execution of code based on LLM outputs.
[PROMPT_INJECTION]: By navigating to and inspecting external web applications or games, the skill ingests untrusted data into the agent's context. A malicious site could use hidden text, DOM elements, or console logs to attempt to influence the agent's subsequent actions.
Ingestion points: Web page DOM, screenshots, and browser console logs accessed via Playwright (page.goto, page.locator, page.on('console')).
Boundary markers: No specific delimiters or instructions are provided to the agent to help it distinguish between legitimate test data and potentially malicious embedded instructions.
Capability inventory: Execution of shell commands and Python scripts via nodes.run and network navigation via the browser tool.
Sanitization: The skill does not describe any sanitization, filtering, or validation of the content retrieved from the target web pages before it is processed by the agent.

playwright-testing