nextjs-chatbot
Pass
Audited by Gen Agent Trust Hub on Apr 22, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is primarily educational, providing blueprints and code snippets for developers building AI-powered web applications. It does not perform any dangerous actions on the host environment.
- [PROMPT_INJECTION]: The skill contains comprehensive guidance on defending against prompt injection. It suggests specific system prompt sections to block override attempts and provides JSON fixtures for benchmarking an agent's resistance to jailbreaks. Static detections of 'ignore instructions' patterns are false positives triggered by these defensive examples.
- [DATA_EXFILTRATION]: No patterns for exfiltrating sensitive data were detected. The persistence mechanisms described (PostgreSQL with Drizzle ORM) follow industry standard practices for session management and include GDPR-compliant 'cascade delete' configurations.
- [EXTERNAL_DOWNLOADS]: The skill references standard, well-known packages from the AI SDK ecosystem. It also suggests using specific MCP servers for development (next-devtools and ai-elements registry), which are established tools within the Vercel/Next.js developer community.
- [REMOTE_CODE_EXECUTION]: No remote code execution vulnerabilities or suspicious script downloads were found. The implementation patterns focus on type-safe tool calling and structured data processing.
Audit Metadata