memory-management

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

• Third-party content exposure detected (high risk: 0.80). The skill's Data Sources and workflow explicitly state it will "auto-populate from SEO tool, analytics, [search console]" and the examples/integration steps show automated ranking checks and competitor/audit targets (e.g., "[If SEO tool connected: runs automated ranking check]" and "target is the URL or domain audited"), which indicates it fetches and ingests public third-party web/tool data that the agent reads and uses to update memory and drive promotions/demotions.