alterlab-trdizin

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

• Third-party content exposure detected (high risk: 0.75). The skill calls the outsider-authored public TR Dizin Elasticsearch API at runtime (https://search.trdizin.gov.tr/api/defaultSearch/{type}/...), ingests the returned JSON _source fields (including journalYear/rejectYearList) into the agent’s context, which is indirect prompt injection exposure from a third-party free-text source.