skills.rest
infinition
infinitionCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

web-research

Answer factual questions from the web with cross-checked sources and citations.

33|1|Updated Mar 4, 2026
Search Tags:#research#web search#web scraping#enumeration#primary sources#fact checking#source citation
One-click install
npx skills add https://github.com/infinition/LaRuche --skill web-research-infinition
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: web-research
Source: https://github.com/infinition/LaRuche/tree/main/laruche/skills/web-research
Command: npx skills add https://github.com/infinition/LaRuche --skill web-research-infinition

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? When a question requires a current, verifiable answer that cannot come from training knowledge, this Skill provides a disciplined procedure for searching the web, opening primary sources, cross-checking claims, and delivering an answer backed by cited URLs instead of stale or fabricated facts. ## Core Features & Use Cases - Tool Selection Guidance: Maps each research need to the right tool, including web_search, web_deep_search, web_fetch, browser, web_discover, and image_search, with clear rules on when snippets are insufficient. - Question vs. Enumeration Workflows: Distinguishes single-answer lookups from exhaustive enumeration tasks, sweeping candidate domains with web_discover until channels are exhausted. - Failure Recovery Playbook: Handles 403 errors, paywalls, JavaScript-rendered pages, syndicated content, and PDFs that resist direct fetching via curl download plus local read_extract. - Use Case: A user asks for the current version of a Rust crate and every available mirror of an old dataset. The Skill runs parallel searches, fetches primary sources, records findings with URLs, sweeps candidate domains, and reports a sourced answer with coverage notes. ## Quick Start Ask the agent to research a current factual question on the web and answer with cross-checked sources and URLs.

Frequently Asked Questions about web-research

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

FAQPage Schema
How do I answer a factual question with web search and sources?▼

Anchor the question in time, run several phrased searches in parallel, open the relevant pages with web_fetch, record each decisive fact with its URL, and cross-check important claims against two independent sources before answering.

What is the difference between web_search and web_deep_search?▼

web_search returns titles, URLs, and snippets so you can see what exists and pick pages. web_deep_search also returns the full text of top results, making it the default when you want the answer itself rather than a list of links.

How do I find every file or mirror on a site that search engines miss?▼

Harvest candidate domains from search results and forum mentions, then run web_discover on each domain with an ext filter for the file type. An empty web_fetch result often means a frameset or JavaScript shell, not an empty site.

Can I read a PDF from the web when web_fetch fails?▼

Yes. Download the PDF with curl to a local path, then use read_extract with the absolute file path to get its text. read_extract only reads local files, so passing it a URL returns a file-not-found error.

Why does a fetched page return empty content?▼

The page is likely JavaScript-rendered or a frameset. Retry web_fetch with render set to true, and if that fails use the browser tool with action navigate followed by action read to load the page in a real browser.

When should I use deep_research_synthesis instead of web research?▼

Use deep_research_synthesis for multi-pass cited reports on broad subjects, calling research_mode first. The web-research workflow is intended for fast, correct lookups of specific questions rather than long-form reports.

Related Skills

Explore More Education & Research Skills→
ShaheerKhawajaShaheerKhawaja
8

pos-research

Conduct multi-source research with citation memory across domains.

Community
Advanced
CornFedKratosCornFedKratos

aidlc-cro

Verify and cite primary sources for knowledge base entries.

Community
Advanced
aws-samplesaws-samples
186

browser-automation

Automate browser UI interactions on web pages for login-protected and JavaScript-heavy sites.

Official
Intermediate
JaviMontanoJaviMontano

web-research

Scrapes, crawls, searches, and synthesizes live web content via firecrawl with URL provenance.

Community
Intermediate