Barrios Skills
Barrios Skills: 46 skills for economists and accountants covering Stata, R, panel data, WRDS, EDGAR, charts and academic writing
- 68 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The skills run Stata, R and Python on your data
- wrds-mcp needs Wharton WRDS credentials and a tunnel daemon
- The econ-humanizer-plus skill rewrites the text given to it
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/barrios-skillsDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The skills run Stata, R and Python on your data
- wrds-mcp needs Wharton WRDS credentials and a tunnel daemon
- The econ-humanizer-plus skill rewrites the text given to it
This is third-party code. Review the repository files before installing.
What it does
Barrios Skills is John Barrios's collection of skills for economists and accountants. The main catalog has 46 skills in four areas: econometrics and empirical data (6 skills: Stata, R, panel data in Python, pyfixest, WRDS), data and visualization (11), research and literature (13: literature search, EDGAR, FRED and OpenEcon, financial NLP, scraping) and writing and review (16: econ-write, humanizer, slides, referee, LaTeX). Extra skills for documents, career and niche libraries live in skills/optional. Skills install by copying a folder into the agent's directory: ~/.cursor/skills/, ~/.claude/skills/ or ~/.codex/skills/, each with its own guide in install. A separate part of the repository, mcp, holds MCP servers: wrds-mcp for Compustat, CRSP and other databases on Wharton WRDS, which the author vendors into the repository, plus links to the external SEC EDGAR MCP, OpenEcon Data MCP and FRED or Stata-MCP. MCP needs pip install, registration in a client and your own credentials. The project site keeps a catalog searchable by task. For example, stata-regression runs regressions in Stata with publication-ready tables, and sec-edgar pulls filings, XBRL and Form 4.
Who it is for. Economists, finance specialists and accounting researchers who write papers and do empirical work in Stata, R or Python and want to give an agent ready methods.
Good fit when
- You need a high-dimensional fixed-effects panel regression in Python through pyfixest
- You need to pull Compustat or CRSP data from WRDS with filters
- You need SEC filings, XBRL or Form 4 for research
- An economics paper has to be written and edited in the style of Cochrane and McCloskey
- You need publication tables and charts in Stata or matplotlib
Not a fit when
- The task is not about economics or accounting: the collection targets those fields
- You have no WRDS access: wrds-mcp needs your own account and connection
- You want one MCP for every database: MCP servers install separately and are registered by hand
- You need a skill for an agent without a skills directory: installation is by copying folders
Example request
Use the pyfixest skill to run a regression with firm and year fixed effects and firm-clustered standard errorsLimitations
- Skills and MCP servers are different things: MCP servers need installation, registration and credentials, while a skill only needs copying.
- wrds-mcp and the wrds skill need your own Wharton WRDS access, and the repository includes a tunnel daemon in mcp/wrds-mcp/tunnel.
- External MCP servers for SEC EDGAR, OpenEcon and FRED install from their own repositories following the guides in install.
- The econ-humanizer-plus skill removes machine-sounding patterns in economics prose and rewrites the text you give it.
- Some skills come from upstream and are credited in their metadata, for example pyfixest from py-econometrics.
How to disable. Delete the folders of the skills you installed from ~/.cursor/skills/, ~/.claude/skills/ or ~/.codex/skills/. For wrds-mcp remove its registration from the client config and stop the tunnel with the tunnel_down.sh script.
FAQ
Which skills in the collection are the main ones?
The README features stata-regression, wrds, sec-edgar, pyfixest, econ-write, econ-humanizer-plus and econ-visualization. The main catalog has 46 skills across four areas, with more in skills/optional.
How do I install the skills in Claude Code?
Clone the repository and copy the folder you need into ~/.claude/skills/. Guides for Cursor, Claude Code and Codex are in the install folder, and each skill installs independently.
How are MCP servers different from skills in this repository?
A skill is a Markdown instruction you only copy. An MCP server is a program with live tools such as SQL queries to WRDS, installed with pip install, registered in a client and connected to your account.
Do I need WRDS access?
Only for the wrds skill and the wrds-mcp server: they work with Compustat, CRSP, TRACE and TAQ through your Wharton WRDS account. The other skills do not need it.
Which external MCP servers are mentioned?
SEC EDGAR MCP, OpenEcon Data MCP and FRED or Stata-MCP. They install from their own repositories following the guides in the install folder, and the author's own wrds-mcp sits in mcp/wrds-mcp.
Discussion
No questions or reviews yet.
Related
Skills17 skills for economists: research ideas, literature review, Stata and R, LaTeX models, tables and Beamer slides
The MCP server and skill of the StatsPAI library: the agent runs DiD, RD, IV and synthetic control on your data
Auto-Empirical Research Skills
Auto-Empirical Research Skills (AERS)
A plugin marketplace for empirical social science research: causal inference, econometrics and manuscript prep for top journals
A template for data scientists: AGENTS.md, CLAUDE.md and skills for Python, SQL, Polars and model diagnostics.
Around 200 skills for Claude Code on bioinformatics, chemistry and biostatistics, from RNA-seq analysis to molecular docking
Claude Code skills for empirical economics: code and replication audit, bibliography checks, PDF reading and slide decks.
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| Barrios Skills (this card) | no | 68 | ||
| MCP-for-Stata | no | not required | 276 | an MCP server only for running Stata, while Barrios Skills adds skills for Stata, R, WRDS, EDGAR and academic writing |
| SEC EDGAR MCP | no | not required | 387 | SEC EDGAR access only, while here it is one external link next to econometrics skills |
README in short
The README splits the repository into skills and mcp, shows coverage across four areas with skill counts, a quick start copying folders for Cursor, Claude Code and Codex and a separate section on MCP servers. Featured skills and task tables with a description of each skill follow.
README badge
Original SKILL.md
---
name: pyfixest
description: Fast high-dimensional fixed-effects regression in Python with pyfixest (fixest-like syntax). Use for OLS/IV/Poisson with many FEs, cluster-robust SEs, modern DiD (did2s, Sun-Abraham), and publication tables — the Python counterpart to Stata reghdfe/fixest for accounting and finance panels.
curated-by: John Barrios
collection: barrios-skills
upstream: https://github.com/py-econometrics/pyfixest
---
> **Barrios Skills** — John Barrios's curated workflow for economists and accountants. Prioritize reproducible empirical work, clear identification language, and journal-ready output.
# PyFixest for empirical accounting & finance
## When to use
- Firm–year / firm–quarter panels with firm and year (or industry×year) fixed effects
- High-dimensional FE that are slow in `linearmodels` or plain `statsmodels`
- IV with FE, Poisson/PPML with FE, or modern DiD helpers in Python
- You already know Stata `reghdfe` / R `fixest` and want the same grammar in Python
**Prefer** [`python-panel-data`](../python-panel-data/) when the design is a simple two-way FE / RE with `linearmodels` and you want textbook panel diagnostics. **Prefer** [`r-econometrics`](../r-econometrics/) / Stata skills when the paper's replication package is R/Stata-first.
## Install
```bash
python -m pip install pyfixest
# optional: python -m pip install "pyfixest[plots]"
```
Docs: [pyfixest.org](https://pyfixest.org/pyfixest.html)
## Iron law: specify clustering and FE before estimating
1. State the **unit of treatment / residual correlation** (usually firm, sometimes industry or CEO)
2. State **fixed effects** that absorb the design (firm + year; firm + industry×year; etc.)
3. Estimate with matching `vcov` / cluster
4. Inspect N, absorbed FE counts, and a coefficient sample before writing results into a paper
## Quick patterns
### OLS with multi-way FE + firm clustering
```python
import pyfixest as pf
fit = pf.feols(
"roa ~ treat + size + mb | gvkey + fyear",