mcpbeat

Deep Research

firecrawl/deep-research

Multi-source research with source triangulation and fact-checking. Use for any research task requiring 3+ sources.

277 tokens
context cost
the whole folder, loaded on every use
1
files
instructions only
0
copies elsewhere
how many repositories repackaged it
1167
stars on the repo
on the repository, not the skill itself

Install

one command, takes just this skill from the repository
npx skills add https://github.com/firecrawl/web-agent --skill deep-research

The instruction itself

5 sections, as written by the author

Deep Research

Search Strategy

  • Break the topic into 3-5 distinct angles
  • 2-3 queries per angle using different terminology
  • Use site: operator for targeted searches (e.g. site:arxiv.org, site:github.com)
  • Include official, comparison, and review-oriented queries
  • Aim for 5-10 unique, high-quality sources

Extraction

  • Scrape each source with a targeted query parameter -- don't dump full pages into context
  • Extract: key claims, data points, dates, attributed quotes
  • Record author, publication, date, and URL for every source

Fact-Checking

  • Cross-reference claims across 2+ sources
  • Flag single-source claims explicitly
  • Assign confidence: high (3+ sources agree), medium (2), low (1 or conflicting)
  • Include contrarian viewpoints -- don't confirmation-bias the results

Output

  • Structure by subtopic, not by source
  • Inline citations for every claim
  • Call formatOutput with the structured data

How to use it

Copy the folder

Take firecrawl/deep-research from the repository into ~/.claude/skills for personal use, or into .claude/skills inside a project.

Check the name does not clash

The agent identifies a skill by the name field in its header. Two skills with the same name cannot sit side by side — one of them will be ignored.