nitro-api-development
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs developers on building API endpoints that ingest untrusted data from external sources using H3 utilities like
readBodyandgetQuery. This creates a potential surface for indirect prompt injection where malicious inputs could influence the application if the processed data is later used in LLM prompts or sensitive logic. - Ingestion points: API handlers within the
server/api/andserver/routes/directories as documented inSKILL.mdandreferences/api-reference.md. - Boundary markers: The guidelines strongly recommend using Zod schemas (
querySchema.parse) to validate and bound input data. - Capability inventory: The skill templates facilitate database operations via Drizzle ORM and file system interactions for project configuration.
- Sanitization: The documentation provides clear patterns for parameter cleaning (e.g., handling empty strings) and schema-based validation to ensure data integrity before it reaches the database.
- [EXTERNAL_DOWNLOADS]: The skill recommends the installation and use of established Node.js packages including
nitro,drizzle-orm,zod,consola, and@neondatabase/serverless. These are well-known libraries in the web development ecosystem.
Audit Metadata