Stata CLI

stata-cli

Use stata-cli as the default local entrypoint for Stata tasks in AI agent.

This skill now includes a local reference library. Do not load everything. Read only the 1-3 files relevant to the current task.

Core workflow

For AI-driven work:

mkdir my-analysis
cd my-analysis
stata-cli init
stata-cli file do/analysis.do
stata-cli data view --input-dta /absolute/path/to/data.dta --max-rows 20
stata-cli data export-csv --input-dta /absolute/path/to/data.dta --output /absolute/path/to/out.csv --replace

Routing table

Read only the files relevant to the task. Paths are relative to this SKILL.md.

Basics and workflow

• references/basics-getting-started.md
• references/workflow-best-practices.md
• references/programming-basics.md
• references/advanced-programming.md

Data work

• references/data-management.md
• references/data-import-export.md
• references/string-functions.md
• references/date-time-functions.md
• references/variables-operators.md
• references/mathematical-functions.md

Statistics and econometrics

• references/descriptive-statistics.md
• references/linear-regression.md
• references/panel-data.md
• references/time-series.md
• references/limited-dependent-variables.md
• references/bootstrap-simulation.md
• references/survey-data-analysis.md
• references/missing-data-handling.md
• references/maximum-likelihood.md
• references/gmm-estimation.md

Stata 19 features (verified on Stata 19.5)

• references/panel-var-xtvar.md — panel VAR (xtvar)
• references/high-dimensional-fixed-effects.md — HDFE absorption in areg, xtreg, fe, ivregress 2sls
• references/cate.md — conditional average treatment effects (cate)
• references/weak-instruments.md — weak-instrument-robust inference (estat weakrobust)
• references/control-functions.md — control-function models (cfregress, cfprobit)
• references/svar-with-iv.md — SVAR via instruments (ivsvar)
• references/iv-local-projections.md — IV local projections (ivlpirf)
• references/correlated-random-effects.md — CRE model and Mundlak test (xtreg, cre, estat mundlak)
• references/gmm-xtinstruments.md — GMM with panel-style instruments (gmm ... xtinstruments())

Causal inference and advanced methods

• references/difference-in-differences.md
• references/regression-discontinuity.md
• references/matching-methods.md
• references/treatment-effects.md
• references/sample-selection.md
• references/nonparametric-methods.md
• references/sem-factor-analysis.md
• references/survival-analysis.md
• references/spatial-analysis.md
• references/machine-learning.md

Graphics, tables, and reporting

• references/graphics.md
• references/tables-reporting.md

Mata

• references/mata-introduction.md
• references/mata-data-access.md
• references/mata-matrix-operations.md
• references/mata-programming.md

Community packages

• packages/package-management.md
• packages/reghdfe.md
• packages/estout.md
• packages/coefplot.md
• packages/winsor.md
• packages/event-study.md
• packages/rdrobust.md
• packages/synth.md
• packages/ivreg2.md
• packages/xtabond2.md
• packages/binsreg.md
• packages/nprobust.md
• packages/asdoc.md
• packages/data-manipulation.md
• packages/graph-schemes.md

Practical defaults

• If the task is more than a few lines, initialize a workspace and write do/analysis.do.
• If the user already has a project folder, keep outputs inside its outputs/ directory.
• If charts are needed, prefer exporting tidy results from Stata and plotting with Python.
• If a package is needed but may not be installed, check with which before using it.

When failure

If setup looks wrong, run:

stata-cli doctor