job-application-operator

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime workflow ingests outsider-authored free text from the job application page it opens (“Open one job application URL” → “Inventory fields” → “Classify fields” → “Fill all fields permitted by policy”), where the page’s form labels/questions and any required custom questions are read into the agent’s LLM context for classification and filling.