clone-website
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external websites, creating a surface for indirect prompt injection. \n
- Ingestion points: External content is ingested via browser automation tools as specified in Phase 1 and Phase 3. \n
- Boundary markers: The instructions lack specific guidance on using delimiters or ignoring commands within scraped content. \n
- Capability inventory: The agent can write files, execute build commands, and run generated Node.js scripts. \n
- Sanitization: There are no explicit requirements for sanitizing scraped text before usage. \n- [DYNAMIC_EXECUTION]: The agent is instructed to generate and execute local scripts at runtime. \n
- Evidence: Phase 2 directs the agent to write and run a Node.js script for downloading binary assets. \n- [COMMAND_EXECUTION]: The skill utilizes shell commands to manage the build process. \n
- Evidence: Execution of
npm run buildandnpx tsc --noEmitis required to verify the integrity of the clone. \n- [EXTERNAL_DOWNLOADS]: The skill fetches resources from remote user-provided URLs. \n - Evidence: Assets such as images, videos, and fonts are downloaded to the local project's public directory.
Audit Metadata