multi-tenant-architecture

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: References official SDKs such as @vercel/sdk and @vercel/global-config, and documentation from trusted organizations like Vercel, Cloudflare, and Neon. These are used for standard platform integration.
  • [COMMAND_EXECUTION]: Includes curl, psql, and rg commands in a validation checklist. These are intended for developers to verify the security of their own deployments (e.g., checking for 404s on unknown hosts or testing RLS policies).
  • [INDIRECT_PROMPT_INJECTION]: The skill guides the creation of platforms that process untrusted hostname and header data. It explicitly addresses the risk of header spoofing and provides architectural mitigations, including stripping inbound x-tenant-* headers and enforcing Row Level Security (RLS) in the database layer.
  • [DYNAMIC_EXECUTION]: Describes the implementation of Cloudflare Workers for Platforms, which involves dynamic execution of tenant scripts. The instructions emphasize the use of 'untrusted' isolation mode and per-invocation resource limits to maintain security boundaries.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 09:57 PM
Security Audit — agent-trust-hub — multi-tenant-architecture