osint

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly instructs fetching and analyzing public, user-generated content from open web sources—e.g., social platforms (Twitter/X, Reddit), GitHub, reverse image search sites, Wayback Machine/Archive.today, and tools like Shodan/VirusTotal—so the agent will ingest untrusted third-party content that could carry indirect prompt injections.