Claude Code Skills for Empirical Research

A collection of custom Claude Code skills designed for empirical research in economics and finance. These skills provide structured workflows for common academic research tasks.

Installation

To use these skills, copy the .md files to your Claude Code commands directory:

# Copy all skills to your Claude commands folder
cp *.md ~/.claude/commands/

After installation, you can invoke any skill in Claude Code using /skill-name (e.g., /robustness).

The skills/ directory holds a second set in the newer agent skill format (a folder per skill, each with a SKILL.md and YAML frontmatter). Those install to a different location:

# Copy the agent skills to your Claude skills folder
cp -R skills/* ~/.claude/skills/

Claude invokes these automatically when a task matches the skill’s description; you can also call one by name. See Agent Skills below.

Available Skills

1. Robustness (/robustness)

Purpose: Comprehensive checklist of empirical robustness tests for finance/economics papers.

Features:

• Core robustness categories: alternative samples, specifications, measures
• Identification tests: placebo, pre-trends, IV diagnostics
• DiD-specific tests (parallel trends, Bacon decomposition, Callaway-Sant’Anna)
• IV-specific tests (first-stage F, weak instrument inference)
• Referee-proof minimum requirements
• Reporting templates for robustness tables

Usage:

/robustness
> "robustness checklist"     # Full checklist for current paper
> "DiD robustness"           # DiD-specific tests only
> "referee-proof"            # Most commonly requested tests

---

2. Literature Review (/lit-review)

Purpose: Systematically summarize academic papers, extract key findings, and identify research gaps.

Features:

• Structured paper summary template
• Data & methodology extraction
• Comparison tables across multiple papers
• Literature gap analysis framework
• BibTeX generation

Usage:

/lit-review
> "summarize this paper"     # Full template
> "quick summary"            # Citation, question, finding, relevance
> "compare papers"           # Comparison table
> "find gaps"                # Identify research gaps

---

3. Coding Guidelines (/coding-guidelines)

Purpose: Standardized Python & Stata coding practices for empirical research projects.

Features:

• Project directory organization
• Python script templates with path management
• Stata script templates with global macros
• Data loading, merging, and validation patterns
• Visualization standards (matplotlib/seaborn)
• Output table conventions (outreg2/esttab)
• Variable naming conventions
• Common pitfalls and best practices

Key Patterns:

# Python path management
def get_project_root():
    return Path(__file__).parent.absolute()

* Stata path setup
global repodir "/path/to/project"
global datadir "$repodir/Data"
global cleandir "$datadir/Clean"

---

4. Data Documentation (/data-doc)

Purpose: Document datasets, variables, sources, and merge keys for replication packages.

Features:

• Dataset overview template
• Source data tracking
• Variable documentation tables
• Variable construction documentation
• Sample filter tracking with observation counts
• Merge key documentation with match rates
• Codebook format for replication packages

Usage:

/data-doc
> "document this dataset"    # Generate full template
> "variable list"            # Create variable table
> "merge documentation"      # Focus on merge keys
> "sample flow"              # Sample filter table
> "codebook"                 # Formal codebook format

---

5. Project Structure (/project-structure)

Purpose: Project directory organization and script naming conventions for research.

Features:

• Standard directory structure (Code/, Data/, Results/)
• Script numbering conventions:
  • 0_: Initial data extraction
  • 1a_, 1b_: Data cleaning
  • 2_: Data merging
  • AN_1_, AN_2_: Analysis scripts
• Data organization principles (Raw/, Intermediate/, Clean/)
• Logging and output management
• Version control best practices

Example Structure:

Project/
├── Code/
│   ├── 0_ExtractData.py
│   ├── 1a_CleanDatasetA.py
│   ├── 2_MergeDatasets.py
│   ├── AN_1_Descriptives.py
│   ├── AN_2_MainRegressions.do
│   └── LogFiles/
├── Data/
│   ├── Raw/
│   ├── Intermediate/
│   └── Clean/
└── Results/
    ├── Tables/
    └── Figures/

---

6. Referee Response (/referee-response)

Purpose: Structure responses to referee reports for R&R submissions.

Features:

• Response document structure template
• Point-by-point response templates
• Response strategies by comment type:
  • Data/sample questions
  • Identification concerns
  • Additional analysis requests
  • Clarification requests
  • Respectful disagreement
• Change tracking summary
• Common phrases and tone guidelines
• Pre-submission checklist

Usage:

/referee-response
> "format referee response"   # Set up document structure
> "respond to this comment"   # Draft specific response
> "disagreement response"     # Phrase respectful disagreement
> "summarize changes"         # Generate changes summary

---

Agent Skills (skills/)

Fifteen skills in the agent-skill format, aimed at the run-it-and-verify-it half of empirical work: executing pipelines, checking that reported numbers match the logs that produced them, and stress-testing a design before a referee does.

Empirical pipeline

Verification

LaTeX / Overleaf

Cross-model review

Paths, usernames, and example coefficients in these files are placeholders — swap in your own.

Quick Reference

Workflow Integration

These skills are designed to work together throughout the research lifecycle:

1. Project Setup: Use /project-structure to organize your directories
2. Coding: Follow /coding-guidelines for Python and Stata scripts
3. Documentation: Use /data-doc to document datasets as you build them
4. Literature: Use /lit-review to systematically review related papers
5. Analysis: Use /robustness to plan comprehensive robustness tests
6. Revision: Use /referee-response to structure your R&R responses

Customization

Each skill is a markdown file with a YAML frontmatter containing the description. You can customize any skill by editing the corresponding .md file in ~/.claude/commands/.

---
description: Your skill description here
---

License

MIT License - Feel free to use, modify, and share these skills.

Contributing

Suggestions and improvements welcome! Feel free to open issues or pull requests.