The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill demonstrates processing untrusted URL search parameters (search.redirect) in src/routes/login.tsx for post-authentication navigation. It includes a mandatory sanitization pattern (sanitizeRedirect) to mitigate open-redirect vulnerabilities. The ingestion point is the validateSearch function, and the capability is limited to internal client-side navigation.
[COMMAND_EXECUTION]: The documentation identifies the risk of unauthorized invocation of server functions (createServerFn) and provides the correct security pattern using authentication middleware to protect these RPC endpoints independently of route-level UI guards.
[SAFE]: The skill implements best practices for authentication state management and role-based access control (RBAC) within a client-side routing context, emphasizing the separation between UI navigation control and server-side data security.

auth-and-guards