skills.rest
escotilha
escotilhaCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

research

Convert URLs and images into relevance-scored improvement recommendations for Claude Code projects.

Updated Aug 27, 2026
Search Tags:#research#claude code#web scraping#skill recommendations#url analysis#relevance scoring#image intelligence
One-click install
npx skills add https://github.com/escotilha/claude-public --skill research-escotilha
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: research
Source: https://github.com/escotilha/claude-public/tree/main/skills/research
Command: npx skills add https://github.com/escotilha/claude-public --skill research-escotilha

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It helps you discover and translate insights from arbitrary web pages and visual content into concrete, Claude Code–aligned recommendations, so you can improve skills, agents, and project workflows without manual triage.

Core Features & Use Cases

  • URL & Image Intelligence: Analyzes URLs and images (repos, articles, tweets, screenshots, videos, podcasts) and classifies them as Tool/API, Pattern/Technique, Product/Feature, or Infrastructure.
  • Targeted Relevance Scoring: Scores recommendations against only the Claude Code setup and the ExampleProject scope, reducing irrelevant or speculative outputs.
  • Actionable Recommendation Report: Produces a structured plan with Claude Setup Recommendations and, when relevant, ExampleProject recommendations plus optional memory capture guidance.
  • Multi-Extractor Pipeline: Extracts content using Firecrawl (primary) with fallbacks through Exa, Scrapling (stealth where needed), WebFetch, and AV summarization when applicable.

Quick Start

Use the /research command with a URL or an image path like /research https://github.com/user/repo to generate a categorized improvement report for your Claude Code setup.

Frequently Asked Questions about research

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

FAQPage Schema
How do I turn a URL or image into actionable recommendations for Claude Code?▼

To turn a URL or image into actionable recommendations, provide the link or media path to trigger content extraction and relevance scoring, generating a categorized improvement report for your Claude Code setup and project workflows.

What is the best way to analyze a GitHub repository for Claude Code skill improvements?▼

Analyzing a GitHub repository for skill improvements involves passing the repo URL to a multi-extractor pipeline that fetches content and classifies findings as Tool/API, Pattern/Technique, Product/Feature, or Infrastructure to produce targeted recommendations.

Can I use web scraping to generate improvement reports from tweets and screenshots?▼

Yes, you can use web scraping and AV summarization on tweets and screenshots to extract content, scoring its relevance against your Claude Code setup to generate structured improvement plans.

Does the extraction pipeline work when Firecrawl fails to fetch a web page?▼

If Firecrawl fails to fetch a web page, the extraction pipeline falls back to Exa, Scrapling for stealth fetching, WebFetch, and AV summarization to ensure best-effort content retrieval for analysis.

Why are recommendations scored against an ExampleProject target?▼

Recommendations are scored against an ExampleProject target to reduce irrelevant or speculative outputs, ensuring findings are strictly scoped to your Claude Code environment and specific project context.

Are there limitations when analyzing media links for Claude Code workflow improvements?▼

Limitations include relying on best-effort extraction chains where AV summarization is applied only when applicable, meaning heavily obfuscated media links may yield incomplete content for generating actionable recommendations.

Related Skills

Explore More Software Engineering Skills→
simplefarmer69simplefarmer69
2

content-quality-auditor

Audit content quality against 80 CORE-EEAT criteria across eight dimensions.

Community
Advanced
syahravisyahravi

self-improvement

Capture errors and learnings into structured .learnings markdown files.

Community
Intermediate
workromancerworkromancer

web-load2

Fetch web URLs and extract all text content.

Community
Basic
Xueheng-LiXueheng-Li
54

fetch4ai

Fetch web content from URLs and apply configurable filtering strategies to produce Markdown-ready output.

Community
Advanced