The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides templates and patterns for creating API routes (+api.ts) that ingest untrusted data from external requests and use it to perform sensitive operations like database queries and background tasks. This creates a surface for indirect prompt injection if the agent processes data from these endpoints without sufficient validation.
Ingestion points: examples/api-routes.md demonstrates handling data from request.json() and query parameters via URL.searchParams in API route handlers.
Boundary markers: Code snippets for API handlers do not implement explicit boundary markers or instructions to ignore embedded commands in the ingested data.
Capability inventory: The provided patterns include database access (db.users.findMany, db.users.create), network responses (Response.json), and task scheduling (runTask, deferTask) as seen in examples/api-routes.md.
Sanitization: The example code snippets lack input sanitization or validation logic for the content of the request body before processing it with database or task operations.

mobile-navigation-expo-router