neon-functions
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches configuration and parent skill definitions from the official 'neondatabase/agent-skills' repository on GitHub.
- [COMMAND_EXECUTION]: Instructions include the use of the Neon CLI and 'npx' commands for local environment setup and function deployment.
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates processing untrusted data in agentic workflows, representing an indirect injection attack surface.
- Ingestion points: Data enters the system via 'request.json()' in HTTP handlers (SKILL.md) and tool 'inputSchema' definitions (references/mcp.md).
- Boundary markers: Agent implementations utilize system prompts to establish operational constraints.
- Capability inventory: Scripts possess capabilities for database querying via Drizzle ORM, file management via the Files SDK, and outbound network requests.
- Sanitization: The skill recommends using Zod for strict input validation and 'jose' for verifying JWT tokens from authentication providers.
Audit Metadata