OSINT

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

• Third-party content exposure detected (high risk: 0.78). The OSINT workflows (e.g., Workflows/CompanyLookup.md Phase 6 / Workflows/DomainLookup.md Phase 8 / Workflows/EntityLookup.md Phase 4) deploy external researcher agents (Perplexity/Claude/Gemini/Grok) that at runtime fetch and summarize public web content from outsider-authored sources (e.g., arbitrary pages/results from sites like Google News, GDELT, VirusTotal, Shodan, etc.), which then becomes readable free text in the agents’ LLM context—an indirect prompt-injection exposure path.