naiba-openai-hr

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

• Third-party content exposure detected (medium risk: 0.65). The skill’s runtime workflow includes prompts that ingest user-provided free text (e.g., “Analyze exit survey themes” and “Analyze trends in employee attrition” with “[Insert responses here]” / “paste table here”), which can contain outsider-authored content, and that text is then fed into the agent’s LLM context.