The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill provides run-code and eval commands which allow for the execution of arbitrary JavaScript and Playwright code within the browser context and Node.js environment.
[CREDENTIALS_UNSAFE]: The skill facilitates access to sensitive browser data, including cookies via cookie-list and local storage, and allows saving complete session authentication states using state-save to local files.
[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection. 1. Ingestion points: The agent reads external web content through snapshot and scraping tasks. 2. Boundary markers: No explicit boundary markers or isolation instructions are defined for processing untrusted web data. 3. Capability inventory: The skill has access to bash command execution, arbitrary code execution, and file writing. 4. Sanitization: There is no evidence of sanitization for data extracted from web pages.
[COMMAND_EXECUTION]: The skill uses a Bash-based CLI that includes process management commands such as kill-all and file system operations like delete-data for user profiles.
[EXTERNAL_DOWNLOADS]: The instructions describe using npx playwright-cli to install the tool, which involves downloading and executing code from the official npm registry.

playwright-cli