streamlit

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

• Third-party content exposure detected (high risk: 0.85). Outsider free text is ingested via st.chat_input() in pages/ai_chat.py, where the user’s prompt (not authored by the operating user) is appended to st.session_state.messages and then sent to the LLM as messages=st.session_state.messages (indirect prompt injection risk).