Stata Skill: Run Econometrics from Your Agent

Where this fits

Imagine a user asks an agent to run a diff-in-diff regression on a proprietary panel dataset. Without a dedicated skill, the agent would need to craft a .do file, call subprocess, parse log output, and hope the syntax is correct. This skill eliminates that friction by wrapping Stata's official Python integration (pystata) into a single, agent-friendly interface. It becomes the default path whenever Stata, econometrics, or .dta files are mentioned.

Why agents benefit

• Direct pystata control: Configure once with stata_setup, then issue every command via stata.run() — no subprocess or shell escaping.
• Seamless pandas exchange: Data flows between Python and Stata in memory, so agents can preprocess with pandas and pass DataFrames directly to Stata, avoiding temporary files.
• Automatic trigger: The skill activates on natural language cues like "run regression," "IV estimation," "margins," or any Stata syntax mention, without user prompting.
• Full .do file support: Agents can execute entire .do scripts and capture output cleanly, enabling reproducible workflows.
• Broad econometric coverage: Handles regressions, RDD, panel data, clustering, and more — the toolset Stata users expect.

Practical scenarios

Running a diff-in-diff with clustered standard errors

An agent receives a CSV and a request to estimate a staggered adoption model. It loads the data into pandas, converts to Stata format in memory, runs xtreg with vce(cluster id), and returns formatted coefficients and p-values — all without writing a .dta file.

Cleaning survey data with Stata commands

A user asks to "recode these variables and generate summary stats." The agent uses Stata's recode and summarize via stata.run(), then pulls the results back into pandas for further analysis or plotting.

Generating marginal effects after a logit model

After fitting a logit, the agent calls margins, dydx(*) and converts the output to a tidy DataFrame, ready for a table or chart, saving the user from manual copy-paste.

Add it to your agent workflow

Install the skill with:

npx skills add kennethkhoocy/applied-micro-skills --skill stata

Then use it in your agent's toolset. For example, after setup, run:

import stata
stata.run("sysuse auto, clear")
stata.run("reg price mpg weight")

The skill handles configuration and exposes a clean Python API. For more context on this and other agent skills, see the OpenAgentSkill directory.

Compare before adopting

Before integrating, weigh the skill's quality signals (clear README, active author), maintenance freshness (last commit date, Stata version compatibility), and alternatives like subprocess calls or Python statsmodels. Also assess workflow fit: does your agent already handle Stata? If not, this skill reduces setup overhead significantly.

Why it is worth tracking

With 47 stars and a focused scope, this skill shows community interest in bridging Stata and agent frameworks. It is maintained by an active contributor and aligns with the growing demand for domain-specific automation in econometrics. Monitor its updates, especially if your team relies on Stata for causal inference — this could become a standard tool in the agent toolbox.