Skills · skills/
Agent instructions
Markdown workflows the agent reads — Stata, LaTeX, writing, data viz. No credentials.
- Copy folder →
~/.cursor/skills/ - e.g. stata-regression, econ-humanizer
John Barrios · Curated collection
A free, open-source toolkit for economists and accountants using AI coding agents. Copy skills into Cursor, Claude, or Codex — and connect to WRDS, SEC EDGAR, and macro data via MCP.
New here? The research pipeline guide shows which skills to use at each step of an empirical project, from question to finished draft.
# Skills — copy markdown
cp -R skills/econometrics/stata-regression \
~/.cursor/skills/
# MCP — WRDS (vendored) or SEC EDGAR (uvx)
pip install -e mcp/wrds-mcp
uvx sec-edgar-mcp
Skills · skills/
Markdown workflows the agent reads — Stata, LaTeX, writing, data viz. No credentials.
~/.cursor/skills/MCP · mcp/
Servers the agent calls as tools — WRDS, SEC EDGAR, OpenEcon macro data, and more.
pip install / uvx + MCP config + credentialsNot sure which of the two you need for a given task? The pipeline guide walks the nine steps of an empirical project in order and names the skills and MCP servers for each one.
Follow the research pipeline →Skills · agent instructions
Hover for details on desktop, tap More on mobile, then Download zip.
MCP servers · live data connections
wrds-mcp is vendored in this repo. SEC EDGAR, OpenEcon, FRED, and Stata-MCP install from upstream — each has a setup page and a matching skill where it helps.
Vendored · featured
Compustat, CRSP, Form 4, TRACE, TAQ. Your WRDS account + Duo tunnel. Pairs with the wrds skill.
→ Full setup guide
External · accounting
Filings, XBRL financials, Form 3/4/5. Install with uvx. Pairs with the sec-edgar skill.
→ Setup guide
External · macro
330k+ indicators from FRED, World Bank, IMF, and more via natural language. Pairs with api-data-fetcher.
→ Setup guide
See install/external-mcps.md for FRED-only and Stata-MCP links. Prefer OpenEcon when you need multi-source macro data.
Install
Skills and MCP install differently — pick the path you need.
git clone https://github.com/Barrios88/barrios-skills.gitunzip stata-regression.zip
cp -R stata-regression ~/.cursor/skills/The agent loads the skill when your request matches its description.
Pick a server: WRDS (your account + Duo), SEC EDGAR (uvx + email User-Agent), or OpenEcon Data (hosted SSE).
cd barrios-skills/mcp/wrds-mcp
python3 -m venv ~/.wrds-mcp-env
source ~/.wrds-mcp-env/bin/activate
pip install -e .# Add to ~/.cursor/mcp.json
# command: uvx args: ["sec-edgar-mcp"]
# SEC_EDGAR_USER_AGENT="Name (email)"# Cursor MCP SSE:
# https://data.openecon.ai/mcp