hr-skills-taxonomy

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

• Third-party content exposure detected (high risk: 0.70). The required workflow is a skills taxonomy/gap-analysis process that explicitly ingests “job descriptions, performance reviews, L&D completion records, and LinkedIn profiles” (and may also use “external labor market data”); those are outsider-authored free text when sourced from public/third-party systems like LinkedIn or external job postings, which would be read by the LLM via the runtime “paste course titles and descriptions / job descriptions / LinkedIn profiles” style inputs.