partner-google

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

• Third-party content exposure detected (medium risk: 0.65). The skill’s required workflow ingests outsider-authored free text via the public A2A protocol guide path (https://www.frankx.ai/guides/agent-card-a2a-spec / /guides/agent-card-a2a-spec) or other public content surfaces, which are fetched/used at runtime as readable prose and can contain indirect prompt-injection payloads.