fingerprinting-server-software

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

• Third-party content exposure detected (high risk: 0.75). The required runtime workflow sends HTTP requests to a user-supplied target URL, then ingests outsider-authored response headers/body text (from that remote server) into the agent/skill context via finding evidence/detail fields that get emitted (i.e., external free text from the target is summarized into LLM-visible output).