system-review backend fixes (judge floor, client_ip, billing match, XSS, whatsapp, audit caps, RAG, og async)

audit v1.3: embeddable badge

audit v1.2: compare mode

audit v1.1: nav-strip + dynamic sitemap

fix(audit): raise judge max_tokens for the gpt-oss reasoning model (empty content was reasoning-token starvation)

debug(audit): expose raw LLM verdicts + robust verdict parsing

fix(audit): credible LLM-judge scoring from real policy pages, decoupled from demo

fix(audit): finalize at listo + lighter crawl

feat(audit): free public store-support audit tool + programmatic-SEO pages (/audit)