clerk-orgs
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions define how the agent should handle untrusted data from external sources such as URL slugs and invitation metadata.\n
- Ingestion points: Dynamic route parameters (params.slug) and API inputs like email_address or publicMetadata used in organization management.\n
- Boundary markers: The skill specifically warns developers to verify that URL identifiers match the active session context and provides code to enforce this boundary.\n
- Capability inventory: Includes capabilities for file system access via Next.js routes, server-side database writes, and network operations through the Clerk Backend API.\n
- Sanitization: Recommends strict server-side validation and the use of Clerk's has() helper to gate actions by role or permission, ensuring that identifiers are bound from the authoritative session token.\n- [COMMAND_EXECUTION]: Includes various CLI command examples for the clerk-cli tool to facilitate programmatic organization setup, membership updates, and instance configuration management.
Audit Metadata