The Agent Skills Directory

[SAFE]: The skill is purely instructional and documents the middleware implementation for the TanStack Start library. It includes explicit security warnings and 'correct vs. incorrect' patterns to guide developers toward secure coding practices.
[INDIRECT_PROMPT_INJECTION]: The documentation identifies untrusted data ingestion points (client-side sendContext) and provides detailed remediation steps, such as mandatory server-side authorization checks against session principals to prevent tenant-walking and unauthorized data access.
[DATA_EXPOSURE]: The skill mentions common authentication patterns like localStorage and Authorization headers in a neutral, instructional context for the purpose of demonstrating middleware capabilities.
[EXTERNAL_DOWNLOADS]: The skill references official TanStack ecosystem packages, such as @tanstack/react-start and @tanstack/zod-adapter. These are standard dependencies for the library documented and originate from the skill's own vendor.

middleware