v1.3.0

Extraction quality, spec analysis, and language detection improvements.

Highlights

• New docs/SPEC.md — YAML parsing spec format and analyze pipeline reference
• HTML <time datetime="..."> support via the html:time pattern
• Text-based feed language detection from item titles (Cyrillic, Ukrainian, French, …)
• Smarter spec analyzer: rejects form/select UI clusters and WordPress taxonomy classes
• analyze CLI now accepts the same fetch flags as extract
• UTF-8/Cyrillic pages decode correctly with <meta charset>
• 175+ offline tests

See CHANGELOG.md for full details.