scan-emerging-innovation-signals-rd

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

• Third-party content exposure detected (medium risk: 0.30). The skill reads/parses user-supplied R&D material (R&D updates, meeting notes, design documents, experiment records, etc.) and then applies extraction and follow-up over that untrusted free text at runtime, meaning an outsider can inject instructions via the submitted document content that the LLM ingests.