June 2026 results: Delivered reliability upgrades and user-facing features for the Diva Scans extension within keiyoushi/extensions-source. Refactored to bypass Next.js RSC hydration to enable robust data extraction for manga details and chapters, reducing data inconsistencies. Fixed critical image loading crashes by decoding _next/image URLs and routing through a CDN, improving visual reliability across titles. Implemented a configurable paid-chapter visibility setting via a new ConfigurableSource, enabling user control over access to locked/paid content. Added comprehensive multi-select filtering for Genres and Tags, and updated search requests to support multi-select checkbox arrays for improved discovery and relevance. Consolidated filtering logic into DivaScansFilters.kt and integrated code-quality improvements per PR feedback (Regex relocation, selectFirst/absUrl usage, HttpUrl parsing). Additional robustness fixes included improved chapter parsing and metadata extraction, resolving issues like 0 chapters on large series and Start Reading for single-chapter titles, plus synopsis cleaning improvements. Version bump to extVersionCode 22. Business impact: higher data reliability, enhanced user control over paid content, improved discovery and filtering, and a solid foundation for future extensibility and scale.