aurora-dsql

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

• Third-party content exposure detected (high risk: 0.75). Outsider free text can enter the LLM context via the user-provided SQL/query text that the runtime workflow (Workflow 8 “Query Plan Explainability”) instructs the agent to run verbatim in psql and then incorporate into the generated Markdown diagnostic report (including verbatim error/plan text), so the LLM ingests untrusted user-authored prose.