The Agent Skills Directory

[SAFE]: The skill provides standard documentation and boilerplate for deploying Clerk applications.
Environment variable examples use clearly marked placeholders like pk_live_... and sk_live_... instead of hardcoded secrets.
Deployment instructions leverage official platform CLI tools (vercel, netlify, railway) and multi-stage Docker builds to ensure secure secret handling.
External links point to official Clerk documentation and trusted platform domains.
[INDIRECT_PROMPT_INJECTION]: The provided verification script scripts/verify-deployment.sh contains a surface for indirect instructions through external URLs.
Ingestion points: The DOMAIN variable in the bash script accepts external URLs.
Boundary markers: Absent.
Capability inventory: The generated script utilizes curl for network requests and python3 for JSON processing.
Sanitization: Absent.
Risk evaluation: The risk is assessed as minimal because the script is intended for user-controlled verification and the agent's own tool execution is restricted to specific platform CLIs by the allowed-tools configuration.

clerk-deploy-integration