The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the playwright-cli utility to automate browser interactions such as clicking, typing, and navigating, which allows the agent to execute actions within a web browser context.
[EXTERNAL_DOWNLOADS]: The install-browser command triggers the download of browser binaries from external servers. Additionally, the tool's reliance on the @playwright/cli NPM package (configurable via environment variables) involves the installation of external software components.
[DATA_EXFILTRATION]: The state-save and state-load features manage sensitive browser session data, including cookies and authentication tokens, by storing them in local files like auth.json. This introduces a risk of credential exposure if these files are accessed by unauthorized users or malicious scripts.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted data from the web via snapshot and goto commands. A malicious website could contain hidden instructions designed to hijack the agent's behavior once the content is processed.
Ingestion points: Browser DOM snapshots and page content accessed via playwright-cli snapshot and open.
Boundary markers: None identified; there are no instructions provided to the agent to treat website content as untrusted or to ignore embedded instructions.
Capability inventory: Full browser automation (click, fill, type, navigate) and the ability to save session state or take screenshots.
Sanitization: No sanitization or filtering of web content is performed before it is presented to the agent.

playwright-cli