The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides a comprehensive suite of commands for browser automation via playwright-cli, npm, and npx. These tools allow the agent to launch browser instances, perform user-like interactions, and manage system-level browser processes.
[DYNAMIC_EXECUTION]: Through commands like run-code and eval, the skill enables the execution of arbitrary JavaScript within the browser context. This is a core feature for advanced automation but introduces a surface for runtime code execution that requires careful handling of inputs.
[DATA_EXPOSURE]: The skill can retrieve and store sensitive browser data using cookie-get, localstorage-get, and state-save. These commands allow the persistence of authentication tokens and session data to the local file system. The documentation correctly identifies this risk and suggests best practices like using .gitignore to prevent accidental exposure.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external websites, creating a surface for indirect prompt injection. A malicious webpage could contain instructions intended to influence the agent's behavior.
Ingestion points: Page content and browser snapshots processed by playwright-cli (SKILL.md).
Boundary markers: No explicit delimiters are used to separate web content from agent instructions.
Capability inventory: Includes arbitrary JavaScript execution (run-code), form interaction (fill), and storage management.
Sanitization: There is no evidence of filtering or sanitization of content retrieved from the browser before processing.
[EXTERNAL_DOWNLOADS]: The skill references the installation of @playwright/cli and @playwright/test from the official npm registry. These are well-known tools from a trusted organization, which is consistent with the skill's purpose.

playwright-cli