codex-slides-research

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

• Third-party content exposure detected (high risk: 0.85). The required CLI/interaction workflow calls functions like deep_research/research and then uses returned “researchDoc” Markdown as the basis for outlining, which at runtime will include outsider-authored free text fetched from public web sources as evidence (i.e., the research evidence contract is ingested into the LLM context).