thinkies-connect-domains

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

• Third-party content exposure detected (high risk: 0.70). The skill’s workflow (“Search distant domains”) implies it will fetch/study structurally similar solved problems from outside sources (e.g., public web content or other third-party text) and then ingest that free-form text into the agent’s LLM context for “Study the solution,” creating an indirect prompt-injection risk.