lov-bp-polish

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

• Third-party content exposure detected (medium risk: 0.30). The required workflow reads outsider-provided BP outline text directly from the user-supplied workspace file ./business-plan/outline.md at runtime (via scripts/audit_bp.py --input and source.read_text(...)), so an outsider can include arbitrary free text in that file to influence the LLM inputs/outputs.