author — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited author (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Write LaTeX that compiles cleanly the first time, reads professionally, and follows the conventions academic publishing expects.
When asked to start a new article, use this. Adjust per-document only when the document needs a feature these don't provide.
\documentclass[11pt, a4paper]{article} % Engine: pdflatex/xelatex/lualatex
% Encoding (only needed on pdflatex; harmless on others if guarded — prefer to drop on modern engines)
% \usepackage[T1]{fontenc} % pdflatex only
% \usepackage[utf8]{inputenc} % LaTeX 2018+ defaults to utf8; explicit for safety
% Math
\usepackage{amsmath, amssymb, amsthm, mathtools}
% Typography
\usepackage{microtype} % protrusion/expansion (full on pdflatex; partial elsewhere)
% Tables and figures
\usepackage{graphicx, booktabs, caption, subcaption}
% Units, numbers
\usepackage{siunitx}
% Hyperlinks (load near the end; cleveref AFTER hyperref)
\usepackage[hidelinks, breaklinks=true]{hyperref}
\usepackage{cleveref} % \Cref, \cref — better than \autoref
\title{Title}
\author{Author One \and Author Two}
\date{} % omit date for camera-ready
\begin{document}
\maketitle
\begin{abstract}
...
\end{abstract}
\section{Introduction}
...
\bibliographystyle{plainnat} % or use biblatex (see bibliography skill)
\bibliography{refs}
\end{document}The engine is chosen by /academic-latex:compile from the preamble. If the document needs Unicode fonts / non-Latin scripts, swap to:
% LuaLaTeX-ready additions (drop fontenc/inputenc):
\usepackage{fontspec}
% \usepackage{polyglossia} % see multilingual skill
\setmainfont{Latin Modern Roman} % or any installed OTF/TTFThese are non-negotiable defaults. Override per journal style only when explicit instructions require it. For deep guidance, see /academic-latex:typography.
booktabs (\toprule/\midrule/\bottomrule), no vertical rules. Captions ABOVE tables.\centering, not the center env. \includegraphics[width=0.8\linewidth]{...}.\label{type:name} with type prefix (fig:, tab:, eq:, sec:, lst:). Reference with \Cref{tab:xyz} from cleveref — handles "Table 3" vs "Tables 3 and 4" automatically.align/align* from amsmath, never eqnarray. Use \(...\) for inline math (more error-tolerant than $...$). text' ` for single, text'' for double. Curly quotes in source become straight "` only when typed naively.-- for ranges (en-dash), --- for em-dash. Never a single - for a dash.~ between citation/reference numbers and the noun: Figure~\ref{fig:overview}. Same for "Dr.~Smith".\usepackage{listings} (load on demand) or \usepackage{minted} (requires -shell-escape).amsthm with named environments: \newtheorem{theorem}{Theorem}
\newtheorem{lemma}[theorem]{Lemma}Use a type prefix in every \label{}:
| Type | Prefix | Example |
|---|---|---|
| Section | sec: | \label{sec:method} |
| Subsection | sec: | \label{sec:method-data} |
| Figure | fig: | \label{fig:overview} |
| Subfigure | fig: | \label{fig:overview-a} |
| Table | tab: | \label{tab:results} |
| Equation | eq: | \label{eq:loss} |
| Listing | lst: | \label{lst:setup} |
| Algorithm | alg: | \label{alg:train} |
cleveref uses these to format the reference correctly. With cleveref, do not type "Figure" before \Cref — it's added automatically.
% Right:
As shown in \Cref{fig:overview}, ...
% Wrong (yields "Figure Figure 3"):
As shown in Figure \Cref{fig:overview}, ...\begin{figure}[t]
\centering
\includegraphics[width=0.8\linewidth]{figures/overview.pdf}
\caption{Pipeline overview.}
\label{fig:overview}
\end{figure}Subfigures (need subcaption):
\begin{figure}[t]
\centering
\begin{subfigure}[b]{0.45\linewidth}
\includegraphics[width=\linewidth]{figures/before.pdf}
\caption{Before.}\label{fig:before}
\end{subfigure}
\hfill
\begin{subfigure}[b]{0.45\linewidth}
\includegraphics[width=\linewidth]{figures/after.pdf}
\caption{After.}\label{fig:after}
\end{subfigure}
\caption{Comparison.}\label{fig:before-after}
\end{figure}For TikZ/PGFPlots inline figures, see /academic-latex:figure-helper.
See /academic-latex:table-helper for booktabs templates and CSV-to-LaTeX conversion. Caption ABOVE the tabular environment.
Two stacks; pick before adding citations:
See /academic-latex:bibliography for setup, citation commands, and conversion guidance. When uncertain which a journal expects, delegate to /academic-latex:docs-lookup.
For non-English or multi-language documents, load /academic-latex:multilingual BEFORE writing — choice of babel vs polyglossia affects the entire document.
Before adding a package option you are not 100% sure about, before claiming a deprecated→modern syntax mapping, before recommending a journal-specific class option:
Invoke /academic-latex:docs-lookup with the specific question. Use the verbatim quote returned. Do not guess.This is the suite-level rule; see plugin README.
Make minimal targeted edits to existing documents. Do not rewrite working preambles to match these defaults — only add what's needed for the requested change. New documents start from the standard preamble above.
2026-05-09 — defaults derived from anthropic-grade academic typography conventions; package guarantees from CTAN docs.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.