스킬 deep-research

🔍

deep-research

Name: deep-research
Author: 199-biotechnologies

낮은 위험 🌐 네트워크 접근📁 파일 시스템 액세스

Conduct comprehensive deep research

또한 다음에서 사용할 수 있습니다: sickn33,lumacoder,199-biotechnologies,21pounder,21pounder

Need thorough, citation-backed research across multiple sources? This skill delivers verified, professional-grade reports through an 8-phase methodology with source credibility scoring and automatic citation validation. Stop spending hours searching and synthesizing—get comprehensive analysis ready for decisions.

지원: Claude Codex Code(CC)

📊 71 적절함

스킬 ZIP 다운로드

Claude에서 업로드

설정 → 기능 → 스킬 → 스킬 업로드로 이동

토글을 켜고 사용 시작

테스트해 보기

"deep-research" 사용 중입니다. Use deep research to analyze quantum computing state of the art in 2025

예상 결과:

Executive Summary with 3-5 key findings (under 250 words)
Research scope, methodology, and assumptions documented
4-8 detailed findings with 3+ source citations each
Synthesis of patterns and novel insights
Limitations, counterevidence, and areas of uncertainty
Actionable recommendations with timelines
Complete bibliography with all 15-30 sources verified
Methodology appendix with credibility scores

보안 감사

낮은 위험

v3 • 1/10/2026

This is a legitimate research skill with minimal risk. All Python scripts use standard library only. Network access is limited to DOI/URL verification for citation validation - a core feature. File system access is restricted to user-specific directories. No code execution, credential access, or data exfiltration patterns detected.

스캔된 파일

2,942

분석된 줄 수

발견 사항

총 감사 수

위험 요인

🌐 네트워크 접근 (1)

scripts/verify_citations.py:128-135

📁 파일 시스템 액세스 (1)

scripts/research_engine.py:128-129

감사자: claude 감사 이력 보기 →

품질 점수

아키텍처

100

유지보수성

콘텐츠

커뮤니티

보안

사양 준수

만들 수 있는 것

Technology evaluation

Compare frameworks, tools, or platforms with verified sources and structured analysis

Market research

Analyze industry trends, funding patterns, or competitive landscapes

Academic investigation

Synthesize scientific literature with citation tracking and credibility assessment

이 프롬프트를 사용해 보세요

Quick overview

Use deep research in quick mode to explore [topic] and identify key sources

Standard analysis

Use deep research to analyze [specific question] with 15+ verified sources

Deep investigation

Use deep research in deep mode to compare [A vs B] with full triangulation and critique

UltraDeep report

Use deep research in ultradeep mode to generate a comprehensive report on [topic] covering all major aspects

모범 사례

Specify your research mode (quick/standard/deep/ultradeep) based on time available and depth needed
Frame questions clearly with context—specific questions yield better results
Review Phase 1 (Scope) output to ensure research is aligned with your needs
Use citation numbers in reports to trace claims back to source material

피하기

Using this skill for simple lookups that WebSearch can answer in one query
Asking for debugging help—this skill focuses on research, not code fixes
Expecting real-time updates on breaking news—research relies on indexed sources
Using for time-sensitive queries without specifying urgency in the research scope

자주 묻는 질문

What modes are available and when should I use each?

Quick (2-5 min) for exploration, Standard (5-10 min) for most questions, Deep (10-20 min) for important decisions, UltraDeep (20-45 min) for comprehensive reports.

How many sources does a typical report include?

Quick mode: 10+ sources, Standard: 15-30, Deep: 25-40, UltraDeep: 30-50+. All sources are scored for credibility.

Can I use this skill with other Claude tools?

Yes. Use WebSearch and WebFetch first, then invoke deep research for synthesis. The skill integrates with Claude Code workflow.

Is my data safe when using citation verification?

Citation verification only checks DOI/URL accessibility—it does not send your research data externally. Only minimal metadata is exchanged.

Why did my report fail validation?

Common issues: missing bibliography entries, placeholders like TODO, broken internal links, or citations without matching sources. Fix and regenerate.

How does this compare to Claude Desktop research?

This skill adds 8-phase methodology, source credibility scoring, citation validation, multiple output formats, and unlimited-length reports via auto-continuation.