open-education-hub-deep-dives

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

• Third-party content exposure detected (high risk: 0.80). The skill's Research Process and Resources/Nearbound Playbook explicitly instruct the agent to gather and integrate material from external public sources (see "External Research" and "Resources" sections naming influencer blogs/podcasts, NHERI/NCES/HSLDA, Facebook groups, and other external websites), meaning the agent will fetch and read untrusted user-generated third-party content as part of its workflow.