vercel-automation

Pass

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from Vercel deployment and runtime logs, creating an attack surface where instructions embedded in logs could influence the agent's behavior.
  • Ingestion points: Deployment and runtime logs accessed via VERCEL_GET_DEPLOYMENT_LOGS and VERCEL_GET_RUNTIME_LOGS in SKILL.md.
  • Boundary markers: The instructions do not specify any delimiters or warnings to ignore instructions within the ingested data.
  • Capability inventory: The skill possesses high-privilege capabilities including VERCEL_ADD_ENVIRONMENT_VARIABLE, VERCEL_CREATE_NEW_DEPLOYMENT, and VERCEL_CREATE_DNS_RECORD.
  • Sanitization: No sanitization or validation steps are defined for the logs before they are processed by the agent.
  • [EXTERNAL_DOWNLOADS]: The skill requires connecting to an external MCP server endpoint at https://rube.app/mcp to provide its automation tools.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 2, 2026, 06:20 PM
Security Audit — agent-trust-hub — vercel-automation