The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill utilizes npx -y chrome-devtools-mcp@latest to download and execute the latest version of the MCP server package from the npm registry at runtime.
[COMMAND_EXECUTION]: The provided configuration examples use shell commands (npx, npm, cmd /c) to launch the MCP server and install global dependencies.
[DYNAMIC_EXECUTION]: The browser_evaluate tool allows the agent to execute arbitrary JavaScript code within the browser context, which is a powerful dynamic execution capability used for interacting with web page internals.
[INDIRECT_PROMPT_INJECTION]: The skill exposes a surface for indirect prompt injection by enabling the agent to navigate to and interpret data from untrusted external websites.
Ingestion points: The browser_navigate tool allows the agent to load content from any external URL (SKILL.md).
Boundary markers: The skill instructions do not specify any delimiters or safety prompts to help the agent distinguish between its instructions and content found on the web.
Capability inventory: The agent has access to highly capable tools, including browser_evaluate for code execution, browser_cookies for accessing session data, and browser_network for inspecting network requests (SKILL.md).
Sanitization: There is no evidence of filtering or sanitizing external web content before the agent processes it.

chrome-devtools-mcp-automation