clone-website
Warn
Audited by Gen Agent Trust Hub on Jul 1, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [DYNAMIC_EXECUTION]: The skill instructs the agent to dynamically generate a Node.js script (
scripts/download-assets.mjs) for asset management and execute it within the local environment. - [PROMPT_INJECTION]: The skill is highly vulnerable to indirect prompt injection. It ingests untrusted data (text, layout, and component structures) from external URLs via browser MCP tools. This data is then interpolated into prompts for 'builder' agents through component specification files without explicit boundary markers or sanitization, potentially allowing an attacker-controlled website to influence the agent's behavior.
- [COMMAND_EXECUTION]: The instructions require the execution of several shell commands, including
npm run buildandnpx tsc --noEmit, to verify the integrity of the cloned project. - [DATA_EXFILTRATION]: The skill performs extensive network operations to extract content and download binary assets (images, videos, fonts) from arbitrary, user-supplied target URLs.
Audit Metadata