indeed-research

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

• Third-party content exposure detected (medium risk: 0.30). At runtime, the workflow ingests outsider-authored free text from user-supplied search inputs (keyword/location strings used as q/l parameters) into the Crawlora API calls (/indeed/search and /indeed/locations/suggest), which are read by the skill’s LLM logic as part of the fetched job/location suggestions and job details.