integration-development

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly ingests and surfaces external-app data — e.g., running /act and using "get-action-run-log" with includeDetails to view the raw HTTP exchange and webhook logs from vendor APIs and connectors — and the docs direct the agent to read and act on those vendor/vendor-sourced (potentially user-generated) responses to decide retries, reconnects, or change action inputs, so untrusted third-party content can materially influence behavior.