quarcs-lab

Scaffold a Quarto analysis notebook (generic or method-specific — DiD, IV, RDD, LASSO, Panel FE), register it in _quarto.yml, apply publication-quality figure styling, and generate dataset codebooks. Use when starting a new analysis notebook.

Manage the project's literature workflow — ideate research questions, run structured literature reviews, add citations to references.bib, write annotation notes in references/, and audit citation integrity. Use for any bibliography or literature task.

Audit a Quarto manuscript or a single notebook — freeze freshness, data-path existence, citation integrity, figure/table export presence, placeholder and anonymization checks. Produces a scored review report under notes/. Read-only — never modifies notebooks, data, or the manuscript.

Build publication-quality regression and robustness tables from notebook estimation output, exported to tables/ as CSV + Markdown + LaTeX. Use when creating a results table.

Draft and revise manuscript prose for index.qmd — sections, the abstract, regression-result interpretation, and referee responses. Use when writing or revising the manuscript.

Reads the manuscript and notebooks to generate a structured abstract. Use when writing or updating the abstract.

Cross-checks citation keys in index.qmd against references.bib, reporting missing, orphaned, and duplicate entries. Use when verifying citations.

Finds a paper by title, author, or DOI, adds BibTeX to references.bib, and shows citation syntax. Use when adding a reference.

Inserts a pedagogical "Key Concepts" glossary section into a metricsAI chapter notebook (.qmd), between Chapter Overview and Setup. Each entry has a bold term, a 1–2 sentence definition, and two side-by-side collapsible Quarto callouts — a dataset-grounded *Example* and a vivid *Analogy* — so readers can skim definitions and expand only the panels they need. The new section is ADDITIVE: it complements (does not replace) the chapter's existing inline `> **Key Concept N.M**` blockquotes by picking 5–8 concepts disjoint from them, sized to chapter complexity, and tied to the chapter's `**What you'll learn:**` bullets. Refuses to run if the section already exists. Fully autonomous one-shot — drafts content, inserts, syncs all formats (.ipynb / .md / .html), verifies the HTML render, and reports. Does not commit. Goal is to foster learning through dataset-grounded examples and vivid analogies. Invoke via /add-key-concepts-section chNN.

Creates a self-contained Python code summary ("cheat sheet") for a metricsAI chapter's Key Takeaways section. Reads the chapter .qmd and its web app to identify key libraries, functions, datasets, and concepts, then generates a single markdown code block that reproduces the chapter's core workflow. The code block is pedagogical (commented), self-contained (runs in an empty Colab notebook), and aligned with the web app's key concepts. Invoke via /create-chapter-code-summary chNN.

Audits and improves the readability and pedagogical scaffolding of a metricsAI chapter web app (web-apps/chNN/template.html). Restructures each widget into a six-part pattern proven on Chapter 2 — Motivation → Key Concept → What you can do here → Controls + Chart → Try-this with reveals → Take-away. Uses the chapter's .qmd as the source of truth for Key Concept wording. Invoke via /improve-webapp-readability chNN [--apply] [--all]. Only touches prose scaffolding and CSS; never touches JavaScript, data, or chart behavior.

Creates a single-file interactive HTML dashboard for a metricsAI chapter. Follows the design and pedagogical conventions established by the Chapter 2 app (Plotly.js, brand palette, light/dark theme, embedded JSON, URL-hash state, Try-this prompts per widget, Reset buttons, delta readouts). Flexible across chapters via reusable widget snippets and a scaffolding script. Invoke via /web-app <chNN>. The goal is pedagogical learning through interaction — design widgets that target specific Key Concepts, not pretty charts.

Audits and improves readability of metricsAI chapter notebooks (.qmd). Checks code cell clarity, explanation ordering, interpretation guidance, regression output presentation, prose quality, and jargon management. Generates a prioritized report with specific line numbers and before/after suggestions. Use after chapter-standard compliance, before proofread. Supports --apply for automated fixes and --all for batch analysis.

Verifies and standardizes metricsAI chapter notebooks against CH01-04 template. Checks structure (45-75 cells, 70-80% markdown), front matter (visual summary, Learning Objectives, overview, setup), content (section numbering, 7-11 Key Concepts), back matter (Key Takeaways, Practice Exercises). Generates compliance report with CRITICAL/MINOR/SUGGESTIONS. Use when reviewing chapters, before PDF generation, or ensuring consistency. Supports --apply for automated fixes.

Compiles all chapter PDFs into a single metricsAI book PDF. Detects modified chapters, regenerates their PDFs, then merges all chapters with cover page, copyright page, Brief Contents, Detailed Contents, Key Concepts TOC, clickable TOC hyperlinks, page numbers, and section-level PDF bookmarks. Use after editing any chapter notebook.

Proofread metricsAI Jupyter notebook chapters for text quality issues. Detects missing spaces (concatenated words), doubled words, broken LaTeX, unclosed markdown formatting, inconsistent terminology, and content accuracy. Use after standardization, before final PDF generation, or when reviewing chapter quality. Supports --fix for automated corrections.

This skill should be used when the user asks to "check references", "check bibliography", "check bibtex", "audit references", "update references", "fix bibtex", "check citations", or "check bib". Also use when the user mentions missing volume, pages, or DOI in bibliography entries. Audits references.bib for completeness and currency, checking only entries cited in the manuscript.

This skill should be used when the user asks to "log progress", "save progress", "write a log", "session summary", "end session", or "wrap up". Also use when the user says goodbye or indicates they are finishing work. Creates a timestamped progress log entry in the ./log/ directory to preserve context across sessions.

This skill should be used when the user asks to "reformat", "fix line breaks", "one sentence per line", "split sentences", "fix formatting", or "enforce sentence breaks". Also use when the user mentions multi-sentence lines, sentence-per-line formatting, or asks to clean up paragraph formatting in the manuscript. Reformats index.qmd so that each sentence starts on its own line, improving git diffs and collaborative editing.

Use this skill when the user asks to "prepare a REGION submission", "freeze a blind copy for the editor", "create a submission bundle", "send the paper to REGION", or "make a submission folder for peer review". Creates a self-contained, reviewer-blind submission folder at legacy/submission-YYYYMMDD/ containing the REGION PDF, Word version, single-file HTML, a fully standalone LaTeX tree, a blind manifest README, and a non-blind cover letter addressed to the editor. Verifies that the standalone LaTeX tree compiles end-to-end and that no author-identifying strings leak into reviewer-facing files before reporting success.

This skill should be used when the user asks to "proofread", "proofread manuscript", "check writing", "edit tone", "academic tone check", "humanize text", or "check for AI writing". Also use when the user mentions proofreading, tone consistency, or removing AI-sounding language from the manuscript. Performs a systematic proofreading pass on the manuscript using a checklist derived from academic economics writing standards.

This skill should be used when the user asks to "render", "build", "compile", "rebuild", or "generate outputs" for the manuscript. Also use when the user mentions Quarto rendering, PDF generation, or "clean render". Runs the full clean-render pipeline that produces HTML, Standard PDF, REGION PDF, and DOCX outputs.

This skill should be used when the user asks to "sync notebooks", "sync jupytext", "update notebooks", "pair notebooks", or mentions syncing .ipynb and .md files. Also use when the user edits a notebook markdown file and needs to regenerate the .ipynb. Syncs Jupytext paired notebook (.ipynb) and MyST Markdown (.md) files.

Audit and revise one chapter of the "Comparative Causal Metrics" Quarto book (this repo) across four dimensions — content/methodology, format/template & style, code/reproducibility, and cross-chapter consistency. Re-renders the chapter to catch live-code bugs, writes a diagnostic report to audit/, fixes findings on approval by priority (P1 → P2 → P3), records what was applied, then commits and publishes. Use whenever the user wants to audit, review, revise, QA, or fix an existing chapter of the book.

Author a new method chapter for the "Comparative Causal Metrics" Quarto book (this repo) in the established house style, then keep the whole book in sync: chapter numbering (append or insert+renumber), _quarto.yml, the preface (index.qmd), chapter 1's roadmap table and decision-tree mermaid, the README, references.bib, the per-chapter download bundles (R/build_chapter_zips.R), and the neighbour-chapter cross-references. Renders the chapter live until clean, then commits and publishes. Use whenever the user wants to add, insert, or draft a chapter for this book.