tribunal

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

• Third-party content exposure detected (medium risk: 0.65). The required workflow is the verifier/panel process operating on an outsider-authored artifact (e.g., code/docs from external contributors) via runtime evidence text (diff + file:line/quotes or command output), which the LLM ingests to score and synthesize—so an indirect prompt-injection path exists through that evidence text.