Our review
Automatically creates a structured source note from a URL, with summary, key ideas, and highlights.
Strengths
- Automates extraction of metadata (title, author, year) from various content types (articles, videos, tweets)
- Provides a summary, key ideas, and relevant quotes
- Integrates the note into an organized file system (Atoms/Sources/) with suggestions for connections
Limitations
- May fail on complex or protected sites (e.g., Twitter/X) requiring manual extraction
- Requires human validation of extracted metadata
- Depends on web content accessibility (paywalls, heavy JavaScript)
When you find a useful online resource and want to capture its essence in your note system.
For very short content, or when you need quick analysis without formal structure.
Security analysis
SafeThe skill uses only read and write operations to extract web content and create local notes. No destructive, exfiltrating, or obfuscated actions. It does not execute external commands or bypass safety measures.
No concerns found
Examples
Save this article about AI ethics: https://example.com/ai-ethicsTake notes on this YouTube video: https://youtube.com/watch?v=abc123Archive this tweet: https://twitter.com/user/status/123name: url description: URL → Source note with summary, key ideas, and highlights allowed-tools: Read, Write, Edit, Glob, Grep, WebFetch, WebSearch, AskUserQuestion, Skill, ToolSearch
The user gives you a URL. Your job is to extract the content and create a well-structured Source note.
Input: $ARGUMENTS
Step 1: Receive URL
If $ARGUMENTS is empty, ask:
What URL do you want to save?
Step 2: Detect type and extract content
Detect the URL type and use the right strategy:
| Type | Detection | Strategy |
|------|-----------|----------|
| YouTube | youtube.com/watch, youtu.be/ | Use skill youtube-transcript for transcription + WebFetch for metadata |
| Twitter/X | twitter.com, x.com | WebFetch (may fail, extract what you can) |
| Article/Blog | Everything else | WebFetch to extract content |
Extract:
- Title of the content
- Author (if identifiable)
- Year of publication
- Type:
article,video,podcast,paper,tweet - Key ideas (3-5 main points)
- Highlights (interesting direct quotes)
Step 3: Present metadata for confirmation
Show what you extracted and ask if it looks right.
Step 4: Status
Ask:
Have you already consumed this, or is it for later?
completed— Already read/watchedto-consume— For later
Step 5: Create Source note
Check first if a similar note already exists in Atoms/Sources/ (by title or URL).
Create the note at Atoms/Sources/Title - Author.md using the exact template from x/Templates/Source.md:
---
date: YYYY-MM-DD
type: [article/video/podcast/paper/tweet]
author: [author]
year: [year]
status: [to-consume/completed]
in:
url: [the original URL]
---
## Overview
[2-3 sentence summary]
## Key ideas
- [idea 1]
- [idea 2]
- [idea 3]
## Highlights
> [quote 1]
> [quote 2]
If no author: just Title.md. Clean special characters from the title.
Step 6: Suggest connections
- Maps: Check existing Maps in
Maps/. Suggest theinfield. - Things: Suggest if any key idea deserves its own note in
Atoms/Things/. Don't create automatically — just suggest. - Wait for confirmation before applying.
Task Prioritizer
Productivity
Prioritizes your tasks using Eisenhower, ICE, and RICE frameworks.
Weekly Status Report Generator
Productivity
Generate structured and concise weekly status reports.
Daily Standup Report
Productivity
Generates structured and concise daily standup reports.