durable-workflow-control

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

• Third-party content exposure detected (high risk: 0.75). This skill’s required Q1 workflow can ingest “external pages, docs, repo files… tool output, comments, traces, and eval artifacts” (i.e., outsider-authored free text) via the runtime step “Read required sources” / “Read required sources” and then “Log source notes… before synthesis,” placing that content into the LLM context as untrusted research input.