setup-api-key
Fail
Audited by Gen Agent Trust Hub on Feb 19, 2026
Risk Level: HIGHREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONCREDENTIALS_UNSAFE
Full Analysis
- REMOTE_CODE_EXECUTION (HIGH): The skill instructs the user or agent to run a remote MCP server using
npx -y mcp-remote https://docs.vapi.ai/_mcp/server. This pattern downloads and executes code from a remote URL that is not on the established list of trusted sources. - COMMAND_EXECUTION (HIGH): In Step 2, the skill uses a
curlcommand that interpolates<the-api-key>directly into a shell execution. Since this value is provided by the user, it creates a shell injection vector. An attacker could provide a payload likedummy_key" ; rm -rf / ; "to execute arbitrary commands on the host system. - CREDENTIALS_UNSAFE (MEDIUM): The skill is designed specifically to solicit, validate, and store high-privilege API keys. While this is its primary purpose, handling raw secrets in plaintext and storing them in
.envfiles requires careful handling and increases the risk of accidental exposure. - DATA_EXFILTRATION (LOW): The skill transmits the user's API key to
https://api.vapi.ai. While this is the legitimate endpoint for the service, it is a non-whitelisted domain according to the security policy.
Recommendations
- AI detected serious security threats
Audit Metadata