The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill is configured to download and execute the chrome-devtools-mcp package from the NPM registry using the npx command during setup and execution.- [REMOTE_CODE_EXECUTION]: The evaluate_script tool allows for the execution of arbitrary JavaScript functions within the context of the active browser page, which can be used to manipulate page content or exfiltrate data from the browser session.- [DATA_EXFILTRATION]: The upload_file tool enables the agent to read local files from the host system and upload them to any website navigated to by the browser. Additionally, multiple tools including take_screenshot, take_snapshot, performance_start_trace, and get_network_request accept filePath parameters, allowing the agent to write data to arbitrary locations on the local filesystem.- [COMMAND_EXECUTION]: The skill configuration requires the execution of the npx command-line utility to run the MCP server.- [PROMPT_INJECTION]: The skill is highly susceptible to indirect prompt injection because it ingests untrusted data from web pages (via snapshots, console logs, and network requests) without boundary markers or sanitization, and has the capability to perform sensitive actions like file uploads and script execution based on that data. Evidence Chain: 1. Ingestion: take_snapshot (Accessibility Tree), list_console_messages, list_network_requests. 2. Boundaries: Absent. 3. Capabilities: upload_file, evaluate_script, navigate_page. 4. Sanitization: Absent.

chrome-devtools