The Agent Skills Directory

[COMMAND_EXECUTION]: The skill facilitates the execution of the clerk binary and various package runners such as npx, bunx, pnpm, and yarn to perform operational tasks like user management and configuration updates.
[EXTERNAL_DOWNLOADS]: The skill refers to fetching the clerk package from the npm registry using standard package runners. These resources are vendor-controlled and follow standard developer workflows.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from external Clerk APIs, representing a surface for indirect prompt injection.
Ingestion points: External data enters the agent's context through clerk api and clerk users list commands, which are often saved to local files like /tmp/users.json.
Boundary markers: The instructions recommend using jq to extract and project specific fields, which serves as a mitigation by isolating the relevant data.
Capability inventory: The skill has access to shell execution, network calls to Clerk APIs, and local file writing for environment configuration.
Sanitization: No specific sanitization of API data is performed prior to processing.
[DYNAMIC_EXECUTION]: The skill provides static templates for using python3 and node to process JSON data from local files, primarily for calculating metrics like record counts or page sizes.

clerk-cli