discover-research — скилл для Claude Code: исследования и установка
Скилл discover-research для Claude Code подбирает исследовательские скиллы по задаче — от планирования эксперимента до обзора литературы. Помогает выбрать методики и этапы работы.
MIT 201 установок ★ 193 обновлён 2026-02-28
npx skills add rand/cc-polymath --skill discover-research -a claude-codeЧто делает discover-research
Скилл распознаёт исследовательские задачи и помогает загрузить подходящие методики: например, research-design для планирования исследования или research-synthesis для обзора литературы. Для количественных и качественных работ он предлагает разные наборы; для смешанного исследования — оба подхода.
Внутри есть таблица соответствия ключевых слов скиллам, дерево выбора и схемы загрузки по этапам: планирование, сбор данных, анализ, написание и синтез. Это навигатор по скиллам, а не сами методики: чтобы применить их, агенту нужно прочитать соответствующие файлы из ../research/.
Установка через npx skills рассчитана на Claude Code; другие команды предусмотрены для Codex, Cursor, GitHub Copilot, Gemini CLI и OpenCode. Например, для Claude Code выполните npx skills add rand/cc-polymath --skill discover-research -a claude-code.
Какие задачи решает
- подобрать скиллы для обзора литературы
- выбрать методики для количественного или качественного исследования
- разложить исследовательскую работу по этапам
- определить, какие файлы методик загрузить для анализа данных
Примеры запросов агенту
- Подбери методики для систематического обзора литературы и загрузи нужные скиллы.
- Я планирую количественное исследование: выбери подходящие скиллы для дизайна и сбора данных.
- Помоги провести интервью и тематический анализ: определи, какие исследовательские скиллы загрузить.
Как установить discover-research
1Через Skills CLI (Vercel)
npx skills add rand/cc-polymath --skill discover-research -a claude-code2Вручную: положить файлы
~/.claude/skills/discover-research/SKILL.mdПапка скилла целиком; для одного проекта — .claude/skills/discover-research/
1Через Skills CLI (Vercel)
npx skills add rand/cc-polymath --skill discover-research -a codex2Вручную: положить файлы
.agents/skills/discover-research/SKILL.mdДля всех проектов — ~/.agents/skills/
1Через Skills CLI (Vercel)
npx skills add rand/cc-polymath --skill discover-research -a cursor2Вручную: положить файлы
.cursor/skills/discover-research/SKILL.mdCursor читает и .agents/skills/, и .claude/skills/
1GitHub CLI
gh skill install rand/cc-polymath skills/discover-research2Вручную: положить файлы
.github/skills/discover-research/SKILL.md1Через Skills CLI (Vercel)
npx skills add rand/cc-polymath --skill discover-research -a gemini-cli2Командой платформы
gemini skills install https://github.com/rand/cc-polymath --path skills/discover-researchВручную: положить файлы
.devin/skills/discover-research/SKILL.mdСтарый путь .windsurf/skills/ тоже читается
Через Skills CLI (Vercel)
npx skills add rand/cc-polymath --skill discover-research -a opencodeClaude.ai
Упакуйте папку в zip и загрузите в Settings → Capabilities → Skills
По шагам, с командами для каждой платформы: Как установить скилл в Claude Code: пять способов
Вопросы о discover-research
Как установить discover-research в Claude Code?
Выполните npx skills add rand/cc-polymath --skill discover-research -a claude-code. Файл скилла находится по пути ~/.claude/skills/discover-research/SKILL.md.
Что делает discover-research?
Он связывает исследовательские задачи и ключевые слова с подходящими скиллами, а также предлагает порядок их загрузки по этапам работы.
Какие скиллы он подбирает для обзора литературы?
Для синтеза литературы он предлагает research-synthesis; если нужен мета-анализ — также quantitative-methods, а для написания — research-writing.
Исходный текст
SKILL.md · MIT
---
name: discover-research
description: Automatically discover research methodology skills when working with research methodology, literature review, systematic review, evidence synthesis, academic research, or experimental design. Activates for research tasks.
license: MIT
metadata:
author: rand
version: "4.0"
compatibility: Designed for Claude Code. Compatible with any agent supporting the Agent Skills format.
---
# Research Skills Discovery
## Auto-Activation
This skill is automatically activated when your task involves:
- Research synthesis, literature reviews, meta-analysis
- Quantitative research, statistical analysis, surveys, experiments
- Qualitative research, interviews, ethnography, case studies
- Study design, hypothesis testing, sampling strategies
- Data collection, survey design, interview protocols
- Data analysis, coding, statistical tests, visualization
- Research writing, academic papers, citations, reporting
## Available Research Skills
### Core Methodology Skills
**1. research-synthesis** - Synthesizing information and conducting meta-analysis
- Narrative synthesis approaches
- Meta-analysis with Python implementation
- Thematic synthesis of qualitative findings
- Evidence mapping and gap analysis
- GRADE framework for quality assessment
- Use when: Integrating findings across studies
**2. quantitative-methods** - Quantitative research and statistical analysis
- Experimental design with power analysis
- Survey methods and analysis
- Hypothesis testing framework
- Regression analysis with diagnostics
- Effect sizes and reporting
- Use when: Testing hypotheses with numerical data
**3. qualitative-methods** - Qualitative research approaches
- In-depth interview protocols
- Thematic analysis (6 phases)
- Grounded theory coding
- Case study research design
- Quality criteria and rigor
- Use when: Exploring experiences and meanings
**4. research-design** - Planning and designing research studies
- Research question formulation (FINER criteria)
- Validity threat analysis
- Sampling strategies with Python tools
- Experimental control frameworks
- Design quality assessment
- Use when: Planning a new study from scratch
### Implementation Skills
**5. data-collection** - Methods for gathering research data
- Survey instrument design and validation
- Interview protocol development
- Observation methods and field notes
- Data quality control frameworks
- Response rate optimization
- Use when: Implementing data collection
**6. data-analysis** - Analyzing quantitative and qualitative data
- Comprehensive descriptive statistics
- Inferential testing with full reporting
- Systematic qualitative coding
- Thematic development process
- Publication-ready visualizations
- Use when: Making sense of collected data
**7. research-writing** - Writing research papers and reports
- IMRAD structure with guidelines
- APA statistical reporting
- Citation management
- Argument construction
- Peer review response strategies
- Use when: Communicating research findings
## Loading Skills
### Load Individual Skills
# From skills directory
Read ../research/research-synthesis.md
Read ../research/quantitative-methods.md
Read ../research/qualitative-methods.md
Read ../research/research-design.md
Read ../research/data-collection.md
Read ../research/data-analysis.md
Read ../research/research-writing.md
### Common Workflow Combinations
**Quantitative Research Study**:
# Planning phase
Read ../research/research-design.md
# Data collection
Read ../research/data-collection.md
Read ../research/quantitative-methods.md
# Analysis and reporting
Read ../research/data-analysis.md
Read ../research/research-writing.md
**Qualitative Research Study**:
# Planning phase
Read ../research/research-design.md
# Data collection
Read ../research/data-collection.md
Read ../research/qualitative-methods.md
# Analysis and reporting
Read ../research/data-analysis.md
Read ../research/research-writing.md
**Literature Review / Meta-Analysis**:
# Synthesis phase
Read ../research/research-synthesis.md
# If including quantitative synthesis
Read ../research/quantitative-methods.md
# Writing phase
Read ../research/research-writing.md
**Mixed Methods Study**:
# All methods
Read ../research/research-design.md
Read ../research/quantitative-methods.md
Read ../research/qualitative-methods.md
Read ../research/data-collection.md
Read ../research/data-analysis.md
Read ../research/research-writing.md
## Progressive Loading
Load skills progressively based on research phase:
**Phase 1: Planning** (Load 1-2 skills)
- research-design (always)
- quantitative-methods OR qualitative-methods (based on approach)
**Phase 2: Collection** (Add 1-2 skills)
- data-collection (always)
- Keep loaded: specific method skill
**Phase 3: Analysis** (Add 1 skill)
- data-analysis (always)
- Keep loaded: method and collection skills for reference
**Phase 4: Writing** (Add 1 skill, can unload others)
- research-writing (always)
- research-synthesis (if synthesizing literature)
- Keep one method skill for reporting details
**Phase 5: Synthesis** (If conducting review)
- research-synthesis (load early)
- quantitative-methods (if meta-analysis)
- qualitative-methods (if thematic synthesis)
## Decision Tree
```
Research Task
↓
Conducting new study?
YES → Load research-design
↓
Quantitative approach?
YES → Load quantitative-methods + data-collection
Qualitative approach?
YES → Load qualitative-methods + data-collection
Mixed methods?
YES → Load both methods + data-collection
↓
Ready to analyze?
YES → Load data-analysis
↓
Ready to write?
YES → Load research-writing
NO → Synthesizing existing research?
YES → Load research-synthesis
↓
Quantitative synthesis (meta-analysis)?
YES → Also load quantitative-methods
Qualitative synthesis?
YES → Also load qualitative-methods
↓
Ready to write?
YES → Load research-writing
```
## Context-Aware Loading
Based on keywords in your task, these skills auto-load:
**Keywords → Skills Mapping**:
- "literature review", "meta-analysis", "systematic review" → research-synthesis
- "survey", "experiment", "hypothesis", "statistical" → quantitative-methods
- "interview", "ethnography", "case study", "lived experience" → qualitative-methods
- "study design", "sampling", "validity", "research plan" → research-design
- "questionnaire", "data collection", "measurement" → data-collection
- "analyze data", "coding", "statistical test", "thematic" → data-analysis
- "write paper", "manuscript", "citation", "peer review" → research-writing
## Related Skill Categories
- **statistics**: Advanced statistical techniques
- **data-science**: Machine learning and big data approaches
- **visualization**: Advanced data visualization
- **academic-writing**: General academic writing skills
- **scientific-computing**: Python/R for research computing
## Quick Start Examples
### "I need to design a survey study"
Read ../research/research-design.md
Read ../research/data-collection.md
Read ../research/quantitative-methods.md
### "I need to analyze interview transcripts"
Read ../research/qualitative-methods.md
Read ../research/data-analysis.md
### "I need to conduct a meta-analysis"
Read ../research/research-synthesis.md
Read ../research/quantitative-methods.md
### "I need to write up my results"
Read ../research/research-writing.md
Read ../research/data-analysis.md
## Best Practices
1. **Start with design**: Load research-design first when planning new studies
2. **Method-specific loading**: Load only the method skill you need (quant OR qual)
3. **Progressive addition**: Add skills as you progress through research phases
4. **Unload when done**: Unload skills from completed phases to manage context
5. **Keep writing loaded**: research-writing useful throughout for documentation
## Skill Maintenance
All research skills follow these standards:
- Practical code examples (Python primary, R when appropriate)
- Real-world templates and protocols
- Best practices and anti-patterns
- Cross-references to related skills
- 250-400 lines optimized for context efficiency
## Integration with Other Skills
Research skills integrate well with:
- **python-data-science**: For advanced analysis
- **python-visualization**: For publication graphics
- **academic-latex**: For paper formatting
- **git-workflow**: For research project management
- **reproducibility**: For reproducible research practices
© rand · MIT · rand/cc-polymath