brycewang-stanford/auto-empirical-research-skills
43 skills631 installs140k starsGitHub
Install
npx skills add https://github.com/brycewang-stanford/auto-empirical-research-skillsSkills in this repo
1Full Empirical Analysis Skill StataThis skill runs a complete empirical analysis workflow in the traditional Stata ecosystem (reghdfe, ivreg2, csdid, rdrobust, psmatch2, esttab, coefplot, and other de-facto-standard commands). An applied economist uses it for a reproducible do-file pipeline across an 8-step process from data import to publication-ready tables and figures. It also covers epidemiology and machine-learning causal-inference modes on the same scaffolding.22installs2Latex DocumentThis skill creates, compiles, and converts LaTeX documents into professional PDFs with PNG previews. A user reaches for it to produce resumes, reports, cover letters, invoices, academic papers, theses, presentations, posters, or fillable forms, and to convert between Markdown, DOCX, HTML, and LaTeX. Its compile script supports pdflatex, xelatex, and lualatex with auto-detection and can also OCR handwritten or printed PDFs into LaTeX.20installs3Literature ReviewThis skill runs systematic, comprehensive literature reviews across multiple academic databases including PubMed, arXiv, bioRxiv, and Semantic Scholar. It follows a multi-phase workflow: scoping with PICO, multi-database searching, thematic synthesis, citation verification, and professional document generation. A researcher uses it when writing a review section, doing a meta-analysis, or mapping the state of the art in a domain.20installs4Avoid Ai Writingavoid-ai-writing audits text for AI writing patterns and rewrites it to remove them, or runs a detect-only mode that just flags patterns. It uses a three-tier word list, formatting rules (em dashes, bold overuse, emoji in headers, excessive bullets), and sentence-structure fixes. A writer uses it before publishing AI-drafted content that reads like it was machine-generated.19installs5Stata Accounting ResearchThis skill provides STATA syntax patterns drawn from 126 peer-reviewed Journal of Accounting Research replication files spanning 2017 to 2025. It answers natural-language queries with published code for methods like panel fixed effects, difference-in-differences, event studies, propensity score matching, entropy balancing, IV, and RD, adapted to your variable names and cited to source. It provides syntax only, not research design or identification advice.19installs6Academic PaperThis skill writes academic papers through a 12-agent pipeline covering configuration, literature search, structure, argument, drafting, citation compliance, peer review, and formatting. A researcher uses it to produce a journal-ready draft in IMRaD or other structures with multi-format citations and bilingual abstracts. It matters for turning research questions or results into a publishable manuscript with a self-review quality check.18installs7Deslopdeslop strips predictable AI writing patterns from prose so it sounds like a specific human wrote it. It applies 10 core rules (cut filler phrases, break formulaic structures, eliminate AI tropes, active voice, be specific, vary rhythm, no em dashes) with reference catalogs of phrases, structures, and tropes, and adapts register between blog, newsletter, and scientific writing. A writer uses it when editing or drafting text that should read naturally rather than machine-generated.18installs8Academic Paper ReviewerThis skill simulates a complete journal peer-review process for an academic paper. A researcher uses it to run five independent reviewers across methodology, domain, cross-disciplinary, and core-argument perspectives, then receive a structured editorial decision and revision roadmap. It matters for pre-submission review, verifying that a revision addressed reviewer comments, or a quick quality assessment.17installs9Causal Inferencecausal-inference is a production-grade Bayesian causal-inference workflow using PyMC, CausalPy, and DoWhy. It enforces DAG-first thinking, mandatory user confirmation of assumptions, a design-selection guide (DiD, synthetic control, RDD, IV, ITS), and design-specific refutation before reporting. A researcher uses it to estimate treatment effects and answer 'does X cause Y' questions defensibly. It depends on the bayesian-workflow skill for all PyMC mechanics.17installs10Causal Inference Mixtapecausal-inference-mixtape is a practitioner code skill built from Scott Cunningham's Causal Inference: The Mixtape. It provides ready-to-run templates for 10 identification strategies (OLS, DiD, event study, staggered DiD, RDD, IV, synthetic control, matching/PSM/IPW, DAGs, randomization inference) in Python, R, and Stata, with cross-language equivalents and required robustness checks. A quantitative analyst uses it to implement a causal method quickly in their language of choice.17installs11Full Empirical Analysis SkillThis skill runs a complete empirical analysis workflow in the traditional Python econometrics stack (pandas, statsmodels, linearmodels, pyfixest, econml, and others). An applied economist or quantitative social scientist uses it to replicate a paper from scratch across an 8-step pipeline from data cleaning to publication-ready tables and figures. It also covers epidemiology and machine-learning causal-inference modes on the same scaffolding.17installs12Stop SlopStop-slop edits prose to remove predictable AI writing patterns. It cuts filler phrases, breaks formulaic structures like 'not X, it's Y' contrasts, enforces active voice, and removes em dashes. A developer runs it while drafting or reviewing text to make writing sound human. It scores drafts across five dimensions and flags anything below 35/50 for revision.17installs13Academic Paper VerifyThis skill verifies the integrity and replicability of an academic paper against its source R scripts and output files. A researcher uses it to cross-check every table number, trace inline quantitative claims, audit the R data pipeline, build a verification manifest, and run replication tests. It matters for auditing that a paper's coefficients, sample sizes, and claims actually match the code that produced them.16installs14Academic PipelineThis skill is a lightweight orchestrator for the full academic pipeline from research exploration to final manuscript. A researcher uses it to move through research, write, integrity check, review, revise, re-review, re-revise, final integrity check, and finalize, with user confirmation at each stage. It matters because it dispatches deep-research, academic-paper, and academic-paper-reviewer while enforcing reproducible quality gates.16installs15Academic WritingThis skill runs a source-aware literature research workflow for Industrial AI and automation topics. A researcher uses it to survey papers across arXiv and top IEEE and automation venues, produce research briefs, literature maps, venue-ranked surveys, or gap memos, and draft outline-first surveys. It matters for structured, evidence-cited research on predictive maintenance, scheduling, anomaly detection, and smart manufacturing.16installs16Anti HallucinationThis is the root router skill for the Auto-Empirical Research Skills (AERS) catalog. When the whole repository is installed as one skill folder, it classifies an empirical-research request by stage and loads the single vendored child skill that fits (causal inference, econometrics, replication, manuscript writing, citation checking, or de-AIGC editing). A researcher uses it so Codex or Claude Code does not read every bundled SKILL.md at once.16installs17Deep Researchdeep-research is a domain-agnostic 13-agent research pipeline for rigorous academic work on any topic. It offers 7 modes (full research, quick brief, paper review, literature review, fact-check, Socratic guided dialogue, and systematic review with optional meta-analysis) and covers question formulation, systematic search, source verification, synthesis, bias assessment, and an APA 7.0 report. A researcher uses it to move from a vague idea to a verified, cited report.16installs18Econ Abstract WritingThis skill is a guide for writing or revising the abstract of an academic economics paper. A researcher uses it when drafting, structuring, or compressing findings into a short abstract for empirical micro, development, or applied economics work. It prescribes a compressed structure that opens with the research question and method, spends most sentences on concrete results, and avoids literature review.16installs19Econ Intro WritingThis skill is a guide for writing or revising the introduction of an academic economics paper. A researcher uses it when framing the motivation, stating the research question, describing the empirical approach, and presenting results in the opening section. It prescribes a consistent multi-paragraph structure that front-loads the paper's own contribution and pushes literature discussion toward the end.16installs20Event StudyThis skill conducts event studies and difference-in-differences analysis in R on panel data. A researcher uses it to test parallel pre-trends, estimate dynamic treatment effects, and handle staggered treatment adoption. It walks through a decision tree from traditional two-way fixed effects to modern robust estimators like Callaway and Sant'Anna and Sun and Abraham, then produces publication-ready coefficient plots.16installs21Humanize ChineseThis skill detects and rewrites AI-generated Chinese text to make it read as human-written. A writer uses it to lower AI-detection scores, reduce academic AIGC scores for CNKI, VIP, or Wanfang, or apply one of eight style transforms. It ships a unified command-line interface and pure-Python scripts that run offline with no dependencies.16installs22Humanizer AcademicThis skill removes signs of AI-generated writing from academic medical manuscripts to make them read as naturally written. An editor uses it to detect and fix patterns such as inflated significance claims, superficial -ing analyses, vague attributions, AI vocabulary, and excessive hedging while keeping the scientific content intact. Its rules are based on Wikipedia's Signs of AI writing guide, adapted for medical literature.16installs23MarginaleffectsThis skill is a reference manual for the marginaleffects R and Python package and its companion book Model to Meaning. It helps a developer interpret statistical models through predictions, comparisons, slopes, and hypothesis tests using a five-question framework. A researcher uses it to choose estimands like ATE, ATT, or CATE, compute risk ratios or odds ratios, and do causal inference with G-computation.16installs24Ml Paper WritingThis skill guides writing publication-ready ML, AI, and Systems papers for venues like NeurIPS, ICML, ICLR, ACL, OSDI, and SOSP. It drafts from a research repository, structures arguments, and verifies citations programmatically instead of generating BibTeX from memory. A researcher uses it to produce a first draft, refine through feedback cycles, and prepare a camera-ready submission with LaTeX templates and reviewer checklists.16installs25Python Econ ComputingThis skill provides best practices for writing Python code for macroeconomic modeling and quantitative economic analysis. It covers DSGE and HANK models, Numba-accelerated numerical methods, fixed-effects estimation with pyfixest, and causal inference methods like DID, IV, RD, and synthetic control. An economist uses it to structure numerical code aligned with economic theory and pick the right library for each task.16installs26Chinese De Aigcchinese-de-aigc rewrites Chinese academic prose to lower AIGC-detection rates on Zhiwang, Wanfang, Weipu, and Turnitin. It identifies 17 categories of Chinese LLM writing tells across five structural features and runs a five-step closed loop (locate, diagnose, rewrite, self-score, recheck) with per-section strategies. A researcher uses it to make an AI-assisted or misflagged Chinese draft read like real researcher writing while preserving academic accuracy.15installs27Full Empirical Analysis Skill RThis skill runs a complete empirical analysis workflow in the modern tidyverse and econometrics R ecosystem (dplyr, fixest, did, rdrobust, MatchIt, grf, modelsummary, and others). An applied economist uses it for a reproducible R script or Quarto pipeline across an 8-step process from data import to publication-ready tables and figures. It also covers epidemiology and machine-learning causal-inference modes on the same scaffolding.15installs28StatsPAI SkillThis skill drives the StatsPAI Python package through a full empirical causal-inference pipeline in the style of an applied economics, epidemiology, or ML-causal paper. It maps each step to a paper section, producing balance tables, event-study figures, main-results tables, and a robustness gauntlet using DID, RD, IV, synthetic control, DML, and matching. A researcher uses it to run an estimand-first analysis and export paper-ready regression tables to Word, Excel, or LaTeX.15installs29Check Citationscheck-citations verifies academic citations in a .bib file against CrossRef, Semantic Scholar, and OpenAlex to detect AI-hallucinated references before submission. It flags fully fabricated, chimeric (real title, wrong authors), and modified-real citations, returning verified / suspicious / not-found statuses. A researcher runs it before submitting a paper or as a CI check in a LaTeX pipeline, since 6-55% of AI-generated citations are fabricated.11installs30Dowhydowhy is a skill for causal inference with the DoWhy framework, answering 'does X cause Y' beyond correlation. It follows DoWhy's identify-estimate-refute loop: define a causal graph, identify the effect via backdoor/frontdoor/instrumental-variable criteria, estimate treatment effects (ATE, ATT, CATE), and validate with refutation tests. An analyst uses it for A/B tests with confounders, counterfactual reasoning, and observational data where randomization is impossible.11installs31Econ AuditThis skill audits economic analysis outputs such as fiscal briefings, macro briefings, and market-research documents against methodology standards and common errors. An analyst runs it to have a document checked for issues like counterfactual, additionality, discounting, double counting, and distributional problems. It returns a red/amber/green scorecard with issues ranked by severity and an optional fix mode that proposes concrete corrections.11installs32Econ WriteThis skill is an assistant for writing, editing, and reviewing any part of an economics paper, thesis, or job-market paper. A researcher invokes it to draft or rewrite an abstract, introduction, results section, literature review, or referee response. Its guidance is synthesized from more than 50 authoritative writing guides by leading economists and covers all paper types and sections.11installs33OpenalexThis skill drives the openalex CLI to retrieve academic metadata from the OpenAlex API. It searches works, authors, sources, and institutions, tracks who cites a paper and what it references, and looks up records by DOI or ORCID. A researcher uses it to explore academic literature, analyze publication trends, and download open-access PDFs, with output in summary, detail, JSON, or BibTeX formats.11installs34Slr PrismaThis skill walks a user through writing a systematic literature review that follows the PRISMA 2020 reporting guideline. It interviews for the review's scope, search, eligibility, and synthesis, then produces a Word manuscript in strict journal format, an annotated PRISMA flow diagram, and APA 7th Edition referencing. A researcher uses it to structure or audit an SLR, though it explicitly does not cover meta-analysis or statistical pooling.11installs35Thematic AnalysisThematic-analysis guides a user through a rigorous qualitative thematic analysis following Braun and Clarke's 2006 six-phase framework. A researcher uses it to code interviews, focus groups, or open-ended survey responses and identify patterns in the data. It settles four upfront analytic decisions, generates codes, develops themes, and produces a Word write-up and an annotated thematic map. It does not cover grounded theory, IPA, or discourse analysis.11installs36Bayesian Workflowbayesian-workflow is a guardrailed Bayesian modeling workflow for PyMC and ArviZ. It enforces a 10-step sequence (formulate, priors, prior predictive checks, nutpie sampling, convergence diagnostics, model criticism, prior sensitivity, comparison, reporting) with critical rules like never reporting point estimates without credible intervals. A data scientist consults it before writing probabilistic model code so the agent applies checks it would otherwise skip.10installs37Briefing Notebriefing-note generates a structured 1-2 page policy briefing note with sections for issue, background, analysis, options, and recommendation. It supports UK GES, Australian Treasury, consulting, and think-tank templates and can auto-populate with live data from econstack R packages. A government or consulting economist uses it to produce a short, decision-focused document for a minister, board, or cabinet.10installs38Cost Benefitcost-benefit runs a cost-benefit analysis that produces an economic NPV and a financial NPV side by side, with BCR, optimism-bias adjustment, Marginal Excess Tax Burden, wellbeing valuation (WELLBY/QALY/VPF), sensitivity, and switching values. It backs onto HM Treasury Green Book primitives and supports EU, World Bank, and ADB frameworks. An economist uses it to judge whether a project is socially worthwhile and financially self-sustaining.10installs39Fiscal BriefingThis skill produces a compact public-finances briefing for the UK, US, or Australia. An analyst uses it to summarize the current deficit or surplus, debt position, receipts by tax, spending by category, and fiscal rules, with every number traceable to an official source and vintage date. It optionally adds a full Debt Sustainability Analysis via the debtkit R package and can export to multiple formats.10installs40LifelinesThis skill wraps the Python lifelines library for survival analysis of time-to-event data. It guides a developer through Kaplan-Meier curves, Cox proportional hazards regression, and log-rank tests while properly accounting for right-censored observations. A researcher uses it when analyzing clinical trial data, comparing survival between treatment groups, or identifying risk factors for prognosis.10installs41LonglistThis skill produces a pre cost-benefit-analysis longlist of a project's benefits and costs. It outputs two tables, each tagging items with materiality, a cash-flow direction, a quantification method, and a monetisation method. An economist or analyst uses it to scope a business case before running a full CBA, choosing from appraisal frameworks like the HMT Green Book, EU Better Regulation Guidelines, World Bank, ADB, or Victorian HVHR.10installs42Macro BriefingThis skill produces an up-to-date macroeconomic briefing for the UK, US, Euro area, or Australia. It pulls live GDP, inflation, employment, wages, rates, trade, housing, and fiscal data from each country's official sources and structures it into a one-page deliverable with a traffic-light macro assessment. An analyst uses it to prepare a pre-brief before a meeting, investment committee, or exam, with every number traceable to its source and vintage.10installs43Market ResearchThis skill produces a source-cited market research report for any industry or product. It covers market sizing, key players, concentration, Porter's Five Forces, PESTLE macro-environment, regulation, supply chains, and outlook, with every data point cited. An analyst uses it to understand a market's structure and competition, across UK, US, EU, Australia, or global scope with multi-geography comparison.10installs