The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the installation of the agent-browser package from npm and the subsequent download of the Chromium browser via agent-browser install.
[COMMAND_EXECUTION]: All browser automation tasks are performed by executing the agent-browser command-line tool, which involves the agent running shell commands.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process content from arbitrary URLs, creating a surface for indirect prompt injection if the target websites contain instructions aimed at manipulating the agent's behavior.
Ingestion points: Browser navigation (open), page snapshots (snapshot), and data retrieval (get text, get value) from external websites (SKILL.md).
Boundary markers: No explicit delimiters or warnings for the agent to ignore instructions within the web content are provided in the skill documentation.
Capability inventory: The tool can perform network navigation, element interaction (click, type), file writing (screenshots, saving session state), and data extraction.
Sanitization: There is no evidence of sanitization or filtering of the content extracted from the browser before it is processed by the agent.

agent-browser