inbound-lead-qualification

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

• Third-party content exposure detected (high risk: 0.95). Yes—Step 2 runs scripts/web_research.py with --fetch-pages, which fetches outsider-authored public web pages (DuckDuckGo HTML search results and extracted page text) and writes them to research.json, which the agent then reads to score and generate reasoning in the LLM context.