clerk-setup
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPERSISTENCEINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill relies extensively on
npx -y clerk@latestto perform project initialization, link applications, pull environment variables, and run diagnostics. These commands execute the vendor's CLI tool to modify the local environment and project files. - [EXTERNAL_DOWNLOADS]: Fetches and executes the Clerk CLI from the official registry and retrieves configuration or documentation from
clerk.comanddashboard.clerk.comto guide the setup process. - [PERSISTENCE]: The initialization command (
clerk init) is described as installing additional agent skills globally on the host machine, which modifies the agent's long-term capabilities and environment. - [INDIRECT_PROMPT_INJECTION]: The skill processes data from several external sources, creating a surface for indirect instructions:
- Ingestion points: CLI command outputs (e.g.,
apps list --json,config pull), and external quickstart documentation fetched viaWebFetchfromclerk.com. - Boundary markers: Absent; the agent is instructed to directly follow steps and configuration found in external documentation.
- Capability inventory: Execution of shell commands via
npx, global installation of skills, and modification of project source code and environment files. - Sanitization: Absent; the skill trusts the integrity of the official vendor resources.
Audit Metadata