knowledge-base-collector
by reed1898
Web & Frontend DevelopmentBrief
Collect and organize a personal knowledge base from URLs (web/X/WeChat) and screenshots.
Description
What This Skill Does
Saves URLs and screenshots into a local knowledge base with tag-based classification. Supports web pages, X/Twitter links, and WeChat Official Account articles, with a macOS node fallback when WeChat cloud fetch is blocked. Entries are stored as markdown files with metadata and indexed for keyword and tag search.
The macOS node fallback for WeChat fetching achieves a higher success rate than standard scrapers, which typically hit verification blocks on WeChat content.
When to Use It
- Archive a WeChat article before it gets deleted or blocked
- Save X/Twitter threads to a searchable personal library
- Store a screenshot with OCR text as a tagged note
- Search past saved links from Telegram by keyword or tag
- Generate a weekly digest of recently saved content
Example Workflow
User sends: https://mp.weixin.qq.com/s/example-article
- 1
Detect source type as WeChat from the mp.weixin.qq.com domain
- 2
Attempt cloud fetch via r.jina.ai
- 3
On verification block, delegate fetch to connected macOS node (e.g. Reed-Mac)
- 4
Write content.md and meta.json to kb/20_Inbox/urls/<YYYY-MM>/<item>/
- 5
Append entry to index.jsonl with assigned topic tags
Article archived with tags like #wechat and #agent, searchable via Telegram
Installation
clawhub install reed1898/knowledge-base-collectorRun the CLI command, or paste the prompt into Claude Code, Cursor, Codex or whichever coding agent you use.