ai-automation-workflows

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

• Third-party content exposure detected (medium risk: 0.65). The required runtime workflow uses belt app run ... --input ... with prompts that can include $RESEARCH/$INPUT_TEXT/$(cat $file) from external sources (e.g., tavily/search-assistant web research or reading outsider-authored .txt files), and that free-form text is then embedded into the LLM prompt context via subsequent belt app run calls.