- Add missing Urdu link to Bengali README language selectors
- Narrow zero-width char regex to only ZWSP (U+200B) and BOM (U+FEFF),
preserving bidi controls and separators
- Add pg_encoding_to_char check for PostgreSQL UTF-8 verification
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Add missing Español link to Bengali README language selectors
- Fix Unicode character table: correct vowel range to U+0985–U+0994,
consonant range to U+0995–U+09B0, include missing vowels ঌ ও ঔ
- Preserve ZWNJ/ZWJ in zero-width character stripping (important for
Bengali conjuncts) — only remove ZWSP, BOM, and formatting chars
- Replace locale.setlocale with PyICU-first sorting approach
- Fix JavaScript sort to not mutate original array
- Use re.fullmatch for digit validation instead of re.match
- Add input sanitization guidance for Bengali prompt engineering
- Add PostgreSQL ICU collation verification step
- Make review summary template dynamic, not hardcoded
- Remove Bash from bengali-reviewer tools list for security
- Make database encoding advice database-specific (utf8mb4 for MySQL)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Add Bengali README translation (docs/bn/README.md) covering guides, quick start, components, agent table, FAQ, and token optimization
- Add bengali-nlp skill (skills/bengali-nlp/SKILL.md) for Bengali text processing: Unicode normalization, conjunct handling, tokenization, script detection, Banglish handling, and dataset preparation
- Add bengali-reviewer agent (agents/bengali-reviewer.md) for reviewing code that handles Bengali text: grapheme cluster safety, NFC normalization, encoding validation, and i18n best practices
- Update language selectors across all existing README translations to include বাংলা
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
