SkillHub Developer Data Dashboard

Problem

Developers who have published skills on SkillHub often need to watch several signals at once: their own download, favorite, rating, and ranking trends, competitor releases, and category heat. Checking APIs or switching pages makes it hard to keep a continuous view. This skill provides a local web dashboard that aggregates data from your own SkillHub account, keyword-based competitor searches, and category heat into one view that refreshes about every 50 seconds, turning raw numbers into trend and iteration decisions.

How It Works

• Set user_id in config.json, run server.py, and open http://localhost:8866.
• The top overview shows platform skill count, total daily downloads, day-over-day growth, favorites, comments, and followers.
• Each skill card displays downloads, favorites, rating, version time, and an iteration note such as “review score appeared N days after vX.”
• The competitor section provides two boards: top 10 downloads and top 10 latest uploads, constrained by keyword_map, name_filter, desc_filter, and exclude when configured.
• The category heat ranking sorts categories by average downloads and includes blue-ocean or red-ocean strategy hints.
• API endpoints such as GET /api/current, GET /api/history, and GET /api/export support checking current state, recent time series, and CSV export for long-term trend analysis.
• Built-in retry, backoff, watchdog, and last-good-data preservation reduce data gaps caused by network jitter or rate limits.

Limits And Notes

It only monitors data from SkillHub and only for skills published under your account. It does not generate or rewrite skill content, nor does it perform operations automatically. It requires access to api.skillhub.cn; when offline, refreshing stops but already loaded data remains. The update interval is about 50 seconds, not millisecond-level real time. Competitor matching depends on keyword and name/description filters, so missing or noisy results usually mean the filters need adjustment. Run one server.py instance per account to avoid multiple pollers triggering 429s.