exploring-data-catalog

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

• Third-party content exposure detected (high risk: 0.85). Step 2 (“Consult Catalog Context”) explicitly calls AWS Glue experimental search/get-asset/list-iterable-forms and then ingests customer-authored assetDescription/assetForms/glossary text into the agent context (even though it says to treat it as untrusted), which is outsider-authored free text from the Glue Data Catalog.