The Agent Skills Directory

[COMMAND_EXECUTION]: The skill facilitates the execution of arbitrary JavaScript code within the browser through the -e flag and the execute tool, allowing full control over the browser session.
[COMMAND_EXECUTION]: The skill requests high-privilege access to the system shell (Bash), which allows for the execution of arbitrary operating system commands.
[REMOTE_CODE_EXECUTION]: The installation and usage instructions recommend using npx playwriter@latest, which dynamically downloads and executes code from the npm registry at runtime.
[DATA_EXFILTRATION]: The skill is designed to interact with the user's existing Chrome session, granting it access to sensitive authenticated data, including session cookies, active logins, and extension data.
[PROMPT_INJECTION]: The skill exhibits a significant surface area for indirect prompt injection attacks.
Ingestion points: The skill ingests untrusted data from external websites using functions like snapshot(), getPageMarkdown(), and getCleanHTML().
Boundary markers: No specific delimiters or instructions are provided to help the agent distinguish between its own instructions and content found on the web.
Capability inventory: The agent has access to arbitrary code execution in the browser (execute), filesystem operations (Read, Write), and shell command execution (Bash).
Sanitization: There is no evidence of sanitizing or validating content retrieved from the web before it is processed or used in decision-making.

playwriter