nx_html_to_text_entity_test.nx

nx_html_to_text_entity_test.nx -- X-HTML-ENT-001 KAT: HTML character- reference decoding in the sovereign browser's text renderer. Corpus (each a distinct exit code on failure): E1 "&#39;" -> "'" decimal numeric ref E2 "&#x27;" -> "'" hex numeric ref (lowercase x) E3 "&#X41;" -> "A" hex numeric ref (uppercase X) E4 "&#8217;" -> U+2019 (E2 80 99) numeric ref >= 0x80 -> UTF-8 E5 "&apos;" -> "'" named entity (newly added) E6 recorded snippet "<p>Claude&#x27;s Constitution</p>" -> "Claude's Constitution\n" (the anthropic.com case) E7 "&#;" -> "&#;" malformed: graceful literal, no skip E4 is the build-intelligence check (codepoint faithfully UTF-8 encoded, not dropped). E7 is the no-silent-skip check (malformed ref renders literally, never crashes or vanishes). expect_exit: 0 license_tier: ORIGINAL