agent-screencast

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly instructs the agent to visit and snapshot arbitrary public web pages as part of the "Research the target pages" workflow (SKILL.md) and even includes a sample that opens https://news.ycombinator.com, meaning untrusted, user-generated web content is read/interpreted and used to create/drive scripted actions that can change the agent's subsequent behavior.