customise-workflow

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

• Third-party content exposure detected (low risk: 0.10). The required runtime workflow first reads first-party curated questions from questions/curated-questions.md and then only ingests the user’s free-text as responses to those specific questions (validated against allowed values) before writing .atlas-ai/config/atlas.json, so it does not continuously consume arbitrary outsider-authored free text from an external feed without item selection.