Name: Knowledge Base Collector
Author: Reed

According to AlphaMoat, as of 2026-09, Knowledge Base Collector has 2,702 downloads and 2 stars — ranked #3,952 of 5,078 AI skills overall, and #340 of 442 in Knowledge.

Installs

283

Downloads

2.7K

Stars

Overall Rank

#3,952

Top 78% overall · Top 77% in Knowledge

Platform	Installs	Downloads	Stars	Comments
ClawHub ↗	75	2.6K	1	-
SkillHub ↗	283	2.7K	2	—

Platforms count differently and update at different times.

What this skill does

Ingest web, X/Twitter, and WeChat Official Account links plus screenshots: prefers r.jina.ai for article extraction (no login required) — extraction prefers r.jina.ai with no login; WeChat articles blocked by risk control get placeholder entries (status=blocked_verification + #needs-manual) refilled later via wechat_backlog.py; identical URLs are deduplicated by key; image ingestion supports OCR via local tesseract (chi_sim) or a multimodal LLM writing --text-file. Search with search_kb.py by keyword/tag/source/time, and weekly_digest.py produces digest candidate lists. Default KB root /home/ubuntu/.openclaw/kb with per-item content.md + meta.json and a global index.jsonl. Safety: - Screenshots or web pages may contain tokens, verification codes, or keys; redact them (replace with REDACTED) before ingesting. Card caution: full URLs (including private or token-bearing ones) are sent to r.jina.ai during extraction — redact or avoid ingesting such links; the WeChat macOS node path executes commands on a connected machine and requires explicit user approval.

Profile published 2026-10-27

What you can do with it

One-click ingest and tagging of web pages, tweets, and WeChat articles
Screenshot ingestion with OCR text extraction
Search the KB by tag, source, or time in Telegram flows
Build the WeChat refill backlog and weekly digest candidates

When to use it

Long-term accumulation of readable, searchable material
WeChat fetching fails due to risk control
Building a personal KB under a “store first, refine later” rule

Example prompts

Generated from the skill’s documentation as examples; not results of actual runs.

Ingest this WeChat article and tag it #ai

Search the KB for entries about claude code from the last 7 days

OCR this screenshot and ingest it

Monthly Trend

Downloads2.7K

Stars2

The counting method changed in 2026-09: skills with the same name are now counted separately by author, and the full catalog is covered. Dashed lines show months before the change; they are not compared with later months, and no growth is computed across the change.

Month	Downloads	MoM	Stars
2026-09	2.7K	—	2
2026-08	2.5K	+8.2%	2
2026-07	2.3K	+17.4%	2
2026-06	1.9K	+10.6%	2
2026-05	1.8K	—	1
2026-03	1.3K	—	1

SKILL.md (original)

The author’s original file, shown as plain text without edits or review by AlphaMoat. Links are not clickable; check commands before running them. Fetched 2026-10-23.

3.7 KB · ClawHub v0.1.3

---
name: knowledge-base-collector
description: Collect and organize a personal knowledge base from URLs (web/X/WeChat) and screenshots. Use when the user says they want to save an URL, ingest a link, archive content to KB, tag/classify notes, store screenshots, or search their saved knowledge in Telegram. Supports WeChat via a connected macOS node when cloud fetch is blocked.
---

## Summary

- Ingest: web URLs, X/Twitter links, WeChat Official Account links (mp.weixin.qq.com), and screenshots
- Store: writes to a shared KB folder with per-item `content.md` + `meta.json` and a global `index.jsonl`
- Organize: tag-first classification with richer tags (e.g. `#agent`, `#coding-agent`, `#claude-code`, `#mcp`, `#rag`, `#prompt-injection`, `#security`, `#pricing`, `#database`)

Show all (81 lines)

- WeChat: cloud fetch may be blocked; when a macOS node (e.g. Reed-Mac) is online, prefer node-side fetch to improve success rate; otherwise create a placeholder entry
- Search: designed to support Telegram Q&A / search flows on top of the index and content

把用户发来的链接/截图沉淀到共享知识库（KB），并做标签化整理。

## 默认 KB 位置
- KB Root（可改）：`/home/ubuntu/.openclaw/kb`
- 索引：`kb/20_Inbox/urls/index.jsonl`
- 每条内容目录：`kb/20_Inbox/urls/<YYYY-MM>/<item>/content.md + meta.json`

> 目标：**先入库不丢**，再迭代“摘要/标签/检索”。

## 你要做的事（按输入类型）

### 1) 普通网页 / X(Twitter) / 公众号 URL 入库
运行脚本：

```bash
python3 /home/ubuntu/.openclaw/skills/knowledge-base-collector/scripts/ingest_url.py "<URL>" --tags "#optional" --note "context"
```

行为：
- 自动识别来源（web/x/wechat）
- 优先用 `r.jina.ai` 抽取正文（无需登录）
- 公众号遇到风控会写占位条目：`status=blocked_verification` + tag `#needs-manual`
- 对同一 URL 做 key 去重（已存在则跳过）

#### WeChat 更高成功率（推荐路径）
当云端抓取命中“环境异常/验证”时：
- 如果有已连接的 macOS 节点（例如 `Reed-Mac`）且该节点能访问该文章，可用 `nodes.run` 在节点上执行抓取（requests+bs4），然后写入 KB。
- 注意：这条路径依赖节点在线与网络环境；无法承诺 100%。

### 2) 截图/图片入库（含 OCR 文本）
脚本：

```bash
python3 /home/ubuntu/.openclaw/skills/knowledge-base-collector/scripts/ingest_image.py /path/to/image.jpg \
  --text-file /path/to/ocr.txt \
  --title "..." --tags "#ai #product" --note "..."
```

说明：
- `ingest_image.py` 负责“落盘+索引”。OCR 可用：
  - 本机 tesseract（若安装了 `tesseract-ocr` + `chi_sim`）
  - 或用多模态 LLM 抽取文字后写入 `--text-file`

## Telegram 里直接问（检索）
推荐先用脚本（本机/服务器）：

```bash
python3 /home/ubuntu/.openclaw/skills/knowledge-base-collector/scripts/search_kb.py --q "claude code" --limit 10
python3 /home/ubuntu/.openclaw/skills/knowledge-base-collector/scripts/search_kb.py --tags "#claude-code #coding-agent" --limit 20
python3 /home/ubuntu/.openclaw/skills/knowledge-base-collector/scripts/search_kb.py --source wechat --since 7d --q "Elys"
```

## 公众号待补抓队列（占位条目）

```bash
python3 /home/ubuntu/.openclaw/skills/knowledge-base-collector/scripts/wechat_backlog.py --limit 30
```

## 周报/主题报告候选清单（给 LLM 写总结用）

```bash
python3 /home/ubuntu/.openclaw/skills/knowledge-base-collector/scripts/weekly_digest.py --days 7 --limit 30
```

## 重要注意事项（安全/隐私）
- 截图/网页可能包含 token/验证码/密钥：入库前应做脱敏（替换为 `REDACTED`）。
- 公众号抓取受风控影响：建议允许“占位入库”，后续再补全。