Bookmark Record and Recommend

browsing history, rankings, related records, time tracking, and recommend list.

v0.3.1 1. Push & Analyze is now available: added a workflow for pushing, pulling, and analyzing data packages and Markdown documents, with GitHub cloud sync support for bookmark records, recommendation signals, and analysis docs. 2. AI analysis experience improved: added and refined the `AGENTS.md` / `CLAUDE.md` rule file, improving analysis output paths, push time ranges, evidence-reading order, and bilingual analysis templates. 3. Other experience improvements: refined several widget and sync-related UI details, including file lists, status messages, path display, and GitHub Token guidance, and fixed ZIP filename encoding, snapshot generation limits, and other reliability details. ───────────── A powerful extension built around Bookmark Records, Smart S-Score Recommendations, and AI Push & Analyze: it computes bookmark records, active time, click rankings, related history, and recommendation signals locally, then lets you download or push the “raw data + computed results” to GitHub so AI can analyze them out-of-the-box with AGENTS.md / CLAUDE.md rules. 1. Bookmark Records (Data Pool) - Active Time Tracking: Tracks your actual active reading time spent on bookmarked pages. - Usage Stats: Collects and reviews bookmark additions, click rankings, related browsing history, and time rankings. - Context Search: Quickly find target records, historical timelines, and contextual clues with one unified search box. 2. Bookmark Recommendations (S-Score Pool) - Formula-Driven Algorithm: Surfaces forgotten high-value bookmarks through S-Score parameters fueled by the Data Pool. - Review Manager: Fully control your recommendation flow with “Review Later”, “Skip”, and “Block”. 3. AI Push & Analyze (Local Computation Result Pool) - Local Precomputation: Computes records, rankings, relations, and recommendation signals locally to reduce the AI’s raw-data burden. - One-Click Push: Download or push the “raw data + computed results” to GitHub as traceable data packages. - Out-of-the-Box Analysis: Let AI read structured outputs directly and analyze them with AGENTS.md / CLAUDE.md rules.