agentsclimarketplace
Sign in

Internet search parsing mastery

Skill onfire7777/universal-ai-skills-library/skills/internet-search-parsing-mastery

Advanced internet search techniques, web content extraction, OSINT methodology, fact-checking, and information synthesis for finding and parsing the best information from the web.From its SKILL.md

Install
npx -y skills add onfire7777/universal-ai-skills-library --skill internet-search-parsing-mastery

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 14 stars14 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its file declares

Copied from the file, not written here

The file declares its own license as Unspecified. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

4.7 KB, 979 tokens by cl100k_base, as published. Nobody here has run it

Internet Search & Parsing Mastery

Purpose

Master the art and science of finding, extracting, validating, and synthesizing information from the internet with maximum efficiency and accuracy.

Search Strategy Framework

The 4-Phase Search Protocol

Phase 1: Query Formulation Before searching, define:

  • What exactly do I need to know?
  • What type of source would have this information? (academic, official, community, commercial)
  • What are the key terms and their synonyms?
  • What date range is relevant?

Phase 2: Multi-Source Search Never rely on a single search. Execute searches across:

  • General web search (for broad coverage)
  • Specialized databases (for depth)
  • Community sources (Reddit, Stack Overflow, HN for practitioner knowledge)
  • Official sources (documentation, government sites, company blogs)
  • Academic sources (Google Scholar, arXiv for research)

Phase 3: Content Extraction

  • Extract the actual content, not just snippets
  • Navigate to primary sources, not summaries
  • Save key findings immediately to prevent loss
  • Note the date, author, and context of each source

Phase 4: Synthesis & Validation

  • Cross-reference findings across sources
  • Identify consensus vs. disagreement
  • Weight by source quality and recency
  • Produce a synthesized answer with citations

Advanced Search Techniques

Query Expansion: For any search, generate 3 query variants:

  1. The literal question
  2. The question rephrased with synonyms
  3. The question from a different angle or domain

Specificity Ladder:

  • Start specific → if too few results, broaden
  • Start broad → if too many results, narrow with additional terms

Domain-Specific Search:

  • site:github.com for code and repos
  • site:stackoverflow.com for programming Q&A
  • site:reddit.com for community discussions
  • site:arxiv.org for research papers
  • site:gov for government data
  • filetype:pdf for documents and reports

Temporal Search:

  • Use date filters for time-sensitive topics
  • Search for "[topic] 2025" or "[topic] latest" for recency
  • Check if older results have been superseded

Web Content Extraction Best Practices

Extraction Priority Order:

  1. APIs first: Check if the site has a public API (REST, GraphQL)
  2. Structured data: Look for JSON-LD, microdata, RSS/Atom feeds
  3. Semantic HTML: Extract from <article>, <main>, <h1>-<h6>, <p> tags
  4. Markdown conversion: Convert HTML to clean Markdown for processing
  5. Full page rendering: For SPAs and JS-heavy sites, render first then extract
  6. Screenshot analysis: For visual content, charts, infographics

Content Quality Signals:

  • Author credentials and expertise
  • Publication date and update history
  • Peer review or editorial process
  • Citation count and backlinks
  • Domain authority and reputation
  • Presence of references and sources

Fact-Checking Protocol

For every important claim:

  1. Source Check: Who is making this claim? What's their credibility?
  2. Evidence Check: What evidence supports it? Is it primary or secondary?
  3. Corroboration Check: Do 2+ independent sources confirm it?
  4. Contradiction Check: Are there credible sources that disagree?
  5. Recency Check: Is this information still current?
  6. Context Check: Is the claim being taken out of context?

Red Flags for Misinformation:

  • No author or anonymous source
  • No dates or very old content
  • Emotional language designed to provoke
  • No citations or references
  • Claims that seem too good/bad to be true
  • Single-source claims with no corroboration
  • Sites with excessive ads or clickbait

Information Synthesis Protocol

When combining information from multiple sources:

  1. Create a source matrix: Source × Key Finding × Quality Rating
  2. Identify areas of convergence (high confidence)
  3. Identify areas of divergence (investigate further)
  4. Weight by source quality, recency, and relevance
  5. Produce a synthesized narrative with inline citations
  6. Explicitly state confidence level for each conclusion
  7. Note gaps — what questions remain unanswered?

Rate Limiting & Ethical Search

  • Respect robots.txt and terms of service
  • Don't hammer servers with rapid requests
  • Cache results to avoid redundant searches
  • Use official APIs when available
  • Attribute sources properly
  • Don't scrape personal/private information

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.