summarize

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

• Third-party content exposure detected (high risk: 0.75). This skill’s workflow explicitly supports taking arbitrary URL/web-page text at runtime (“Get the text…use web fetch tool or ask for a paste”), which would be outsider-authored free text that the agent then “reads fully” and summarizes, creating an indirect prompt-injection path into the LLM context.