Where does the skill store its findings?

It creates a dedicated directory structure under ./research/[topic-slug]/ containing raw findings, a numbered list of sources, and a final synthesized report.md file.

Does it support source citations?

Absolutely. The skill maintains a sources.md file and uses inline citations (e.g., [1], [2]) in the final report to ensure all factual claims are verifiable.

How does Deep Research handle context window limits?

The skill saves findings to local markdown files after each iteration and summarizes them before the next step. It reads from these files during the final synthesis to ensure the full breadth of research is captured without exceeding context limits.

Can I control how thorough the research is?

Yes, you can specify 'depth' (the number of search-and-refine iterations) and 'breadth' (the number of queries per iteration) parameters to customize the intensity of the investigation.

Deep Research

Name: Deep Research
Author: jimmc414

byjimmc414

•

网络抓取与数据收集

Conducts multi-step, iterative web investigations to produce comprehensive, structured research reports on any topic.

The Deep Research skill empowers Claude Code with a systematic methodology for exhaustive information gathering and synthesis. Inspired by advanced autonomous research agents, it performs multiple iterations of searching, fetching, and gap analysis to move beyond surface-level answers. By maintaining a dedicated research directory with raw findings, source logs, and a final report, it provides a transparent and verifiable audit trail of its investigations. This skill is ideal for technical feasibility studies, competitive analysis, or gathering deep context on complex topics directly within the developer workflow.

主要功能

01Automated gap analysis to identify unanswered questions between search cycles

02Local file-based state management to prevent context window overflow

03Structured report generation including executive summaries and inline citations

041 GitHub stars

05Iterative search loops with configurable depth and breadth parameters

06Source quality filtering prioritizing authoritative domains and primary sources

使用场景

01Technical feasibility studies for new libraries, frameworks, or architectural patterns

02In-depth investigation of complex technical bugs, CVEs, or security vulnerabilities

03Competitive landscape analysis and market research for software products

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jimmc414/claude-code-plugin-marketplace deep-research

For use in Claude.ai and ChatGPT

Download Skill