analyze-web

Analyzes web clippings, creates resource notes with Layer 1 content, and returns lightweight JSON proposals.

6|1|Updated Nov 26, 2025
One-click install
npx skills add https://github.com/nathanvale/side-quest-marketplace-old --skill analyze-web-nathanvale
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: analyze-web
Source: https://github.com/nathanvale/side-quest-marketplace-old/tree/main/plugins/para-obsidian/skills/analyze-web
Command: npx skills add https://github.com/nathanvale/side-quest-marketplace-old --skill analyze-web-nathanvale

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Web clippings captured in an Obsidian inbox arrive as raw stubs that need enrichment, categorization, and note creation without flooding the coordinating agent's context with full article content. This Skill processes a single clipping end-to-end and returns only a compact proposal. ## Core Features & Use Cases - Domain-aware content fetching: Routes YouTube URLs to transcript tools, X/Twitter links to the X-API, and other articles to Firecrawl scraping. - Atomic note creation: Creates the resource note and injects Layer 1 captured notes in a single para_create call, keeping full content out of the coordinator context. - Structured proposals: Returns a ~500-byte JSON proposal with title, summary, area/project wikilinks, resource type, and verification status. - Use Case: During inbox triage, a clipped Twitter thread is fetched via X-API, classified as a thread resource, linked to the relevant area and project, and filed under 03 Resources/ with formatted Layer 1 notes. ## Quick Start Analyze the clipping at '00 Inbox/✂️ Article Title.md' and return a triage proposal with the created resource note.

Frequently Asked Questions about analyze-web

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I analyze a web clipping in an Obsidian inbox?

Provide the clipping file path, available areas, and projects as input. The skill reads the file, fetches full content based on the source domain, creates a resource note with Layer 1 content, and returns a JSON proposal.

Which tool fetches content from YouTube, Twitter, and articles?

YouTube URLs use get_transcript with get_video_info as fallback, X/Twitter links use x_get_tweet via the X-API, and GitHub or general articles use firecrawl_scrape. X/Twitter enrichment is mandatory because clippings only capture stubs.

Does the full article content reach the triage coordinator?

No. The skill creates the note and injects Layer 1 content inside its own context, returning only a roughly 500-byte JSON proposal with metadata, categorization hints, and verification status.

What happens if note creation fails during analysis?

The para_create call with content injection is atomic, so failure leaves no partial state. The skill sets created and layer1_injected to null and still returns an error proposal to the coordinator.

How is Layer 1 content formatted for long articles or videos?

Articles are truncated to 2-3k tokens using the opening, key section headings, and conclusion. YouTube transcripts are sampled at roughly 10 percent of segments with timestamps, while threads are usually kept in full.