The Agent Skills Directory

[DYNAMIC_EXECUTION]: The run-code and eval commands allow for the execution of arbitrary JavaScript snippets within the browser context during runtime.
[DATA_EXPOSURE]: The skill exposes tools to extract sensitive browser data, including listing and retrieving cookies (cookie-list, cookie-get), local storage (localstorage-list), and session storage. It also includes functions to save and load full authentication states to files like auth.json (state-save, state-load).
[INDIRECT_PROMPT_INJECTION]: The skill facilitates browser navigation and data extraction from external web pages, creating a surface for indirect prompt injection.
Ingestion points: External websites accessed via open and goto commands in SKILL.md.
Boundary markers: No specific delimiters or warnings are provided to prevent the agent from following instructions embedded in the content of processed web pages.
Capability inventory: The skill has broad capabilities including arbitrary JavaScript execution, file uploads, and session data access.
Sanitization: There are no instructions for sanitizing or filtering data retrieved from external domains before it is processed by the agent.
[EXTERNAL_DOWNLOADS]: The skill documentation suggests using npx playwright-cli, which downloads and executes the package from the NPM registry if not already present.
[COMMAND_EXECUTION]: The skill uses the Bash tool to run commands via the playwright-cli interface.
[PRIVILEGE_ESCALATION]: The run-code command is demonstrated granting browser-level permissions, such as geolocation, via page.context().grantPermissions().

playwright-cli