The Agent Skills Directory

[COMMAND_EXECUTION]: The skill implements an execute function that allows for arbitrary JavaScript execution within the browser context. This is a primary feature designed for DOM manipulation and data extraction. Evidence found in SKILL.md and references/commands.md.
[DATA_EXFILTRATION]: The tool can access and extract sensitive browser data, including cookies, local storage, and page content. The documentation (references/authentication.md) identifies these as risks and provides remediation guidance, such as using environment variables and avoiding the logging of passwords.
[PROMPT_INJECTION]: The skill inherently exposes an indirect prompt injection surface by ingesting content from arbitrary websites.
Ingestion points: Content is retrieved from external URLs via the snapshot and execute functions (SKILL.md).
Boundary markers: No explicit delimiters or instructions are provided to the agent to isolate external web content.
Capability inventory: The skill can interact with elements, execute JavaScript, navigate the browser, and upload local files.
Sanitization: No content sanitization or filtering is performed on the data retrieved from web pages before it is returned to the agent context.

agent-browser