higgsfield-mcp

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

• Third-party content exposure detected (high risk: 0.85). Marketing Studio workflow calls show_marketing_studio(action="fetch", url=<product_url>), which fetches/scrapes public web product pages at runtime and then feeds the scraped product info into the LLM context via the tool’s returned data (indirect prompt injection risk).

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly calls show_marketing_studio to fetch and scrape a user-supplied product URL (<product_url>) at runtime, and the scraped images/product info are then used automatically to build generation prompts (so external content directly controls prompts).