More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Table Design: Publication-Ready Results Tables

    1

    This skill turns your agent into a careful table maker.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Table Design: Publication-Ready Results Tables

    Table Design: Publication-Ready Results Tables

    Example session with this skill installed

    Make Table 2 from results.csv for the paper (LaTeX booktabs) and a Markdown version for the README.

    • Read your context and instructions
    • Compiled the table design
    • Generated the data export

    I've written Table 2 for the paper (table2.tex) and a Markdown version for the README (table2.md). I rendered the LaTeX version to previews/table2.png and checked it: the columns line up, the missing cell shows a dash, and nothing is cut off.

    One thing to check before you use it: ProtoMix is not the best method on every dataset. On Seeds, ViT-S has the higher mean (95.4 ± 1.6 against 95.2 ± 0.7). That gap is well within one SD, so the caption says ProtoMix "is the most accurate method on three of the four datasets; on Seeds it is within one SD of ViT-S."

    • Layout: methods are rows and datasets are columns. Each cell is test accuracy as mean ± SD over 5 seeds.

    Rounding: percentages with one decimal. Each SD is rounded to one significant digit, or two when it starts with 1, and the mean to the same decimal place.

    • Bold: the highest mean in each dataset column, not a significant difference. A note under the table says this.
    • Missing cell: ViT-S on Plankton shows "–". The CSV marks that run as out of memory, and a note says so.

    Fewer seeds: ResNet-18 on Pollen was run with 3 seeds, not 5. A note says so, but it's worth checking that this was intended.

    The LaTeX file needs \usepackage{booktabs} in your preamble.

    table2.png

    CSV · data export

    Generated

    Example file from a real run - the skill writes it into your workspace.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    What it does

    A results table is where a reviewer checks your claims, and six-decimal script output pasted into LaTeX makes them work for it. This skill turns your agent into a careful table maker. Give it a CSV or JSON of results and it:

    • Decides the form first: a sentence, a figure or a table, and whether the table is for looking values up (supplementary) or for showing one comparison (main text).
    • Pivots long results (one row per method × dataset) into methods as rows and datasets as columns, and orders the rows by a rule it states.
    • Rounds for the reader: two to three significant digits by default; for mean ± SD, the SD keeps one significant digit (two when it starts with 1) and the mean is rounded to match. Large and tiny values are scaled, with the factor in the header.
    • Marks what matters, and only that: the best value per column in bold, or one highlighted row; units in headers; digits aligned; booktabs rules and no vertical lines.
    • Never hides a gap: missing runs print as an en dash with a note that says what the data say; cells with a different n are flagged; when the best value is within one SD of the runner-up, it says so instead of implying a win.
    • Writes a caption that leads with the takeaway, then n and what ± means.
    • Renders the table and looks at it before handing over: LaTeX compiled to PNG, HTML screenshotted.

    Also writes Markdown for a README and HTML for reports or web pages, and handles Spanish number formats (decimal comma, narrow-space or point thousands separator), including reading CSVs written that way.

    What's included

    • SKILL.md: the workflow (form, purpose, shape, rounding, caption and notes, render and check) and the formatting rules, plus guidance for tables on slides and in spreadsheets
    • scripts/table_fmt.py: CSV or JSON to Markdown, booktabs LaTeX or HTML: pivot, mean ± SD, significant-digit rounding, scaling, scientific notation, best-value bold, row highlight, notes, totals, share of total, inline bars, decimal comma. Standard library only.
    • scripts/preview.py: renders a .tex table (any LaTeX engine) or .html table (Chrome, Chromium, Edge or Brave, found on macOS, Linux and Windows) to PNG
    • examples/: a synthetic benchmark CSV and the LaTeX, Markdown, HTML and Spanish tables made from it

    Part of the Aurora Lab research pipeline

    Works entirely on its own. With a research/ folder it reads results from research/results/ and writes tables to research/tables/. Pairs well with Aurora Lab's figure skills (Paper Figures, Figure Audit), but needs none of them.

    It formats and reviews tables; it does not run experiments, compute statistics beyond what is in your file, or test whether differences are significant (bold marks the highest mean, and the skill says when a gap is within one SD). Rows must already be aggregated: the pivot expects one row per method × dataset, so per-seed raw results need averaging first (the agent can do this with a short script, but the bundled tool does not). LaTeX output uses plain right-aligned columns and booktabs, not siunitx decimal alignment, and very wide or multi-page tables need manual layout (longtable, landscape). Previews need a local LaTeX engine or Chrome-family browser; without them the agent reads the table source as text. Journals differ on caption style and table format; check your venue's current author guidelines.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 27 days ago

    Requires

    Python 3.10+
    Python
    macOS

    Permissions

    Terminal / Shell
    Read Files
    Write Files

    File Scopes

    research/results/**
    research/tables/**

    Reads the CSV or JSON results you point it at (by default in research/results/), or a table you ask it to clean up. Writes the tables (.tex, .md or .html) and their preview images to research/tables/ when a research/ folder exists, otherwise next to the data or where you say. Runs a LaTeX engine or a headless browser locally only to render previews. Never modifies your data files. No network access; no installs without asking.

    Listed29 days ago
    Updated27 days ago

    What's inside

    Frequently Asked Questions

    A CSV (comma, semicolon, tab or pipe separated) or a JSON list of records. Either wide (one row per item, one column per measure) or long (one row per method × dataset with mean and SD columns, which it pivots). Empty, NA or n/a cells are treated as missing and print as an en dash.

    Only Python 3.10+ to make the tables. To see a rendered preview: a LaTeX engine plus poppler's pdftoppm or pymupdf for LaTeX tables, or Chrome, Chromium, Edge or Brave for HTML tables. Your paper needs \usepackage{booktabs}.

    By default each column gets one number of decimals so its typical value shows three significant digits, never more than your file has. Mean ± SD cells follow the SD: one significant digit (two if it starts with 1), with the mean rounded to the same place. You can fix decimals per column.

    No. Bold marks the highest (or lowest) mean in each column. The skill says in a note what bold means, and flags when the best value is within one SD of the next, so the table does not imply a win the data do not support.

    Yes. It writes the decimal comma with a narrow-space or point thousands separator, and reads input files written that way (1.234,5) including semicolon-separated CSVs. Captions and notes are written in whatever language you ask for.

    Those make and review figures. Table Design makes and reviews tables. Each works on its own; they pair well when you prepare a paper.

    New in Research & Analysis