audit-catalyst-experimental-methods-rd

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

• Third-party content exposure detected (medium risk: 0.30). In scripts/run_audit.py, the workflow reads the entire user-supplied input file content at runtime via read_input(--input) (including .txt/.md/.html/.pdf/.docx text extraction) and then directly extracts/classifies evidence from that text.