signals-scout-web-analytics

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

• Third-party content exposure detected (medium risk: 0.30). In SKILL.md, the runtime workflow ingests attacker-adjacent free-text attributes from the acquisition stream (URLs/paths/referrers/UTMs/hostnames supplied by browsers) and then uses those fields (e.g., $entry_pathname/$entry_referring_domain/$entry_utm_* and event properties) in LLM-visible reporting/evidence even though it treats them as untrusted data.