brand-dna-builder

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

• Third-party content exposure detected (medium risk: 0.30). The workflow takes user-provided brand website content and reads it at runtime via WebFetch (or Firecrawl scrape), meaning an outsider can supply a URL that returns free text used for extraction.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs calling the Firecrawl MCP connector (https://mcp.firecrawl.dev/{api-key}/v2/mcp) at runtime via the firecrawl_scrape tool to fetch and inject scraped website content into the agent context, which directly controls prompts.