research-synthesis

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

• Third-party content exposure detected (high risk: 0.80). The skill's Required Inputs explicitly accept "source material: notes, documents, excerpts, links, transcripts, or pasted text" and the Workflow/Output Contract require the agent to read and synthesize those sources (e.g., propose Context Artifact updates and recommend next skills), so untrusted/public third-party content provided via links or pasted material could directly influence decisions and actions.