The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external webhook payloads originating from Clerk's infrastructure.
Ingestion points: Untrusted data enters the application context via the POST request body in app/api/webhooks/clerk/route.ts.
Boundary markers: The skill explicitly requires HMAC-SHA256 signature verification using the svix library or @clerk/backend helpers, providing a strong security boundary for data authenticity.
Capability inventory: The skill demonstrates writing processed data to a database (Prisma db.user.create/update) and caching event IDs in Redis.
Sanitization: Payload data is destructured and mapped to specific database fields rather than being executed or interpreted directly.
[EXTERNAL_DOWNLOADS]: The skill references and installs well-known, industry-standard packages for its primary functionality.
Node Packages: Includes svix for webhook verification, @clerk/nextjs for authentication, and @upstash/redis for idempotency management.
[COMMAND_EXECUTION]: The instructions include command-line steps for local development and testing.
Patterns: Recommends using ngrok for tunneling and @clerk/cli for development environment synchronization.

clerk-webhooks-events