vercel-marketing-team-eve-template

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). Outsider-authored free text enters the workflow via the Slack/TUI request channels (e.g., end-user messages that “write a blog post” / “audit this page's SEO” are delegated to specialists), and the lead/specialists will ingest that text to generate outputs.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The skill invokes MCP servers at runtime via mcp('npx -y @modelcontextprotocol/server-notion') and checks npx -y @modelcontextprotocol/server-typefully, which cause npx to fetch and execute remote npm packages during runtime (executing remote code that the connectors rely on).