design
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill includes a protocol for cloning external websites based on scraped content and screenshots. This creates a vulnerability where malicious instructions hidden in the source website (e.g., in HTML comments or invisible text) could be processed and followed by the agent.
- Ingestion points: External page content and screenshots (SKILL.md).
- Boundary markers: Absent; instructions demand strict replication.
- Capability inventory: Generation of complex React/Next.js/Tailwind code and file system writes.
- Sanitization: Absent.
- [EXTERNAL_DOWNLOADS]: The instructions require the agent to install the third-party npm package '@chenglou/pretext' to perform typography calculations.
Audit Metadata