base44-platform-backend-functions

Pass

Audited by Gen Agent Trust Hub on Nov 3, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill provides instructions for generating backend HTTP handlers that ingest and process untrusted external data from HTTP requests. This creates an attack surface for indirect prompt injection if the data is subsequently passed to sensitive sinks like LLMs or database queries without adequate sanitization.
  • Ingestion points: Handlers ingest data via the 'req' parameter in functions defined in 'base44/functions/{functionName}/entry.ts' and server routes in 'src/routes/api/'.
  • Boundary markers: Instructions suggest returning 4xx error codes for invalid input and utilizing Zod for schema validation in TanStack Start apps.
  • Capability inventory: Generated functions can perform database operations via 'base44.entities', execute network requests using 'fetch', and invoke LLMs through 'base44.integrations.Core.InvokeLLM'.
  • Sanitization: The guidelines recommend input validation and authentication checks but rely on the agent to implement specific sanitization for all data sinks.
  • [SAFE]: The skill enforces several security constraints and best practices to harden the generated code.
  • Explicitly prohibits the use of 'eval()' and 'new Function()'.
  • Mandates the use of 'base44.auth.me()' for user authentication and requires 'user.role === "admin"' checks for privileged operations.
  • Provides a secure runtime module for secret management, discouraging hardcoded credentials.
  • Includes specific safety logic for asynchronous Stripe webhook signature verification and defines strict trust boundaries for guest checkout processes.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 3, 2026, 02:21 PM