The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill fetches framework-specific quickstart guides from an external domain and instructs the agent to follow the instructions, which includes package installation and code modification.
Ingestion points: The skill uses WebFetch to retrieve content from https://clerk.com/docs/{framework}/getting-started/quickstart.
Boundary markers: The instructions lack explicit boundary markers or warnings to the agent to disregard potentially malicious instructions embedded in the fetched documentation.
Capability inventory: The skill allows the agent to install npm packages (npm install) and write sensitive configuration files (e.g., .env.local containing CLERK_SECRET_KEY).
Sanitization: There is no evidence of sanitization or validation of the documentation content before it is processed by the agent to generate commands.
[EXTERNAL_DOWNLOADS]: The skill fetches documentation and setup guidelines from clerk.com, which is a well-known service for authentication and user management. This fetch operation is transparent and follows the skill's stated purpose.
[COMMAND_EXECUTION]: The skill facilitates the execution of package management commands (e.g., npm install @clerk/nextjs) and file operations to configure the application's authentication middleware and environment variables based on the detected framework.

clerk-setup