v1.26.0 (Stable) - i18n guard + health English
1.26.0 — Stable minor: i18n behavioral guard + health diagnostics English
🛡️ Stable release. Bundles patches 1.25.1 + 1.25.2.
Highlights
- i18n layer hardening (1.25.1) — adversarially verified the whole 8-phase English layer: no runtime bugs (uiLang crash-safe,
--languagevalue never leaks as a positional,--language=enworks,--jsonstays valid under en, flag > env > manifest). Added a behavioral e2e regression guard that closes the source-guard blind spot which let an earlier overstatement slip. - health diagnostics English (1.25.2) —
leerness healthrenders fully in English under--language en: section labels, the 6-capability matrix evidence, security issues, and summary. - Korean-first preserved — English is opt-in (
--language en/LEERNESS_LANG=en/ English-initialized project). Korean output is unchanged byte-for-byte.
Verification (zero regression)
- selftest 245/245 · E2E 366/366 (Korean-default path unchanged; i18n behavioral guard covers lens + health en/ko) · post-publish clean-room re-verification.
Notes
- 0 runtime deps · 0 install scripts · Node ≥ 18 · MIT.
- Patches 1.25.1 / 1.25.2 are not separately on npm (accumulated here per the minor-only publish policy).