The Agent Skills Directory

[SAFE]: No malicious patterns or security risks were identified during the analysis of the skill's instructions and reference files.
[INDIRECT_PROMPT_INJECTION]: The skill documents procedures that ingest external data via the .input() method. It explicitly promotes the use of Zod schemas for validation and sanitization, which is the recommended practice for mitigating injection risks in tRPC-based applications.
Ingestion points: Data enters the system through defined procedures in references/router-procedures.md and references/subscriptions.md using the .input() method.
Boundary markers: Zod validation schemas are used as strict boundaries for all incoming data.
Capability inventory: Procedures are documented for data fetching, database mutations, and real-time subscriptions, all governed by middleware and validation.
Sanitization: The documentation highlights schema-based validation as the primary sanitization mechanism.
[EXTERNAL_DOWNLOADS]: The skill recommends the installation of standard, well-known libraries from the official NPM registry, including @trpc/server, @trpc/client, @trpc/react-query, @tanstack/react-query, and zod. These are established packages from a well-known service and represent standard development workflows.

trpc