email-campaign

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

• Third-party content exposure detected (high risk: 0.75). The workflow calls runtime HTTP endpoints via curl .../v1/proxy/orthogonal/run (e.g., hunter domain-search/email-finder and brand-dev brand/retrieve), which can fetch and return outsider-authored web/content data as readable text into the agent’s LLM context through the proxy responses.