097 Index D2ba2b2b

English Papers (latex-paper-en)

Complete toolkit for English academic paper writing with LaTeX.

Overview

The latex-paper-en skill provides comprehensive support for writing English academic papers in LaTeX, with a focus on major publication venues (IEEE, ACM, Springer, NeurIPS, etc.).

Key Features

• Multiple compilation recipes (pdflatex, xelatex, latexmk, with bibliography workflows)
• ChkTeX integration for LaTeX linting
• Format checking with venue-specific rules (IEEE, ACM, Springer)
• Bibliography verification (BibTeX format validation)
• Prose extraction for grammar checking
• Style guide references (Common Chinglish errors, academic writing best practices)
• Chinese-to-English academic translation (Deep Learning, Time Series, Industrial Control domains)
• De-AI writing analysis for reducing AI-generated text traces

Environment Requirements

Note: This skill assumes LaTeX environment is already configured on your system.

Windows: MiKTeX or TeX Live installed and added to PATH macOS/Linux: TeX Live installed

Required tools: pdflatex, xelatex, latexmk, biber, chktex

Using the Skill in Claude Code

This skill is designed to work with Claude Code and similar AI assistants. Simply mention the relevant trigger words in your conversation, and the assistant will activate the appropriate module.

Argument Conventions

Provide clear inputs in your request:

• Main .tex path (required for tool execution)
• Target scope (section/chapter or full document)
• Module choice (compile / format / grammar / translation / etc.)

If any of these are missing or ambiguous, the assistant will ask for clarification instead of guessing.

Execution Guardrails

• Tools/scripts run only when you explicitly request execution.
• Destructive operations (--clean, --clean-all) require explicit confirmation.

Example Usage

Compile your paper:

Please compile my LaTeX paper main.tex using xelatex with bibtex

Check format:

Can you check the format of my paper for IEEE conference submission?

Translate to English:

Translate this Chinese text to academic English (Deep Learning domain):
本文提出了一种基于Transformer的方法...

De-AI polishing:

Please reduce AI writing traces in my introduction section

Modular Design

The skill uses a modular design where each module can be invoked independently:

Output Protocol

All suggestions must use a diff-comment style and include fixed fields:

• Severity: Critical / Major / Minor
• Priority: P0 / P1 / P2

Minimal template:

% <MODULE> (Line <N>) [Severity: <Critical|Major|Minor>] [Priority: <P0|P1|P2>]: <Issue>
% Before: ...
% After:  ...
% Rationale: ...
% ⚠️ [PENDING VERIFICATION]: <if evidence/metric is required>

If a tool fails (missing script/tool or invalid path), respond with an error comment and a safe next step.

Compile Module

Tools (matching VS Code LaTeX Workshop)

Recipes

Usage in Claude Code

Simply ask the assistant to compile your paper with specific requirements:

Basic compilation:

Compile main.tex using xelatex

With bibliography (recommended for papers):

Compile main.tex with xelatex and bibtex workflow

With custom output directory:

Compile main.tex using latexmk and output to build directory

Clean auxiliary files:

Clean auxiliary files for main.tex

The assistant will execute the appropriate compilation commands based on your request.

Script CLI:

python scripts/compile.py main.tex
python scripts/compile.py main.tex --biber
python scripts/compile.py main.tex --outdir build
python scripts/compile.py main.tex --recipe xelatex-biber

Failure Handling

• Missing LaTeX tools: install TeX Live/MiKTeX and ensure PATH is set
• Missing file/script: verify working directory and scripts/ path
• Compilation error: summarize the first error and request the relevant log snippet

Format Check Module

Usage in Claude Code

Ask the assistant to check your paper format:

Check the format of main.tex

Check main.tex format with strict mode for IEEE submission

The assistant will analyze your paper and provide format suggestions.

Grammar Analysis Module

LLM-based grammar checking focusing on:

• Subject-verb agreement
• Article usage (a/an/the)
• Tense consistency
• Chinglish detection

Script CLI (MVP):

python scripts/analyze_grammar.py main.tex
python scripts/analyze_grammar.py main.tex --section introduction

Usage in Claude Code

Ask the assistant to check grammar:

Check the grammar in my introduction section

Proofread the methods section and fix Chinglish errors

Sentence Decomposition Module

Decompose long sentences (>50 words or >3 clauses) to improve readability.

Script CLI (MVP):

python scripts/analyze_sentences.py main.tex
python scripts/analyze_sentences.py main.tex --section methods --max-words 45 --max-clauses 3

Usage in Claude Code

Simplify the long sentence in Section 3

Split sentences longer than 50 words in my introduction

Expression Module

Improve academic tone by replacing weak verbs and colloquial phrasing.

Script CLI (MVP):

python scripts/improve_expression.py main.tex
python scripts/improve_expression.py main.tex --section related

Usage in Claude Code

Improve academic tone in the abstract

Replace weak verbs in the related work section

Logic & Methodology Module

Ensure logical flow between paragraphs and strengthen methodological rigor in academic writing.

AXES Model for Paragraph Coherence

Transition Signals

Methodological Depth Checklist

• Each claim is supported by evidence (data, citation, or logical reasoning)
• Method choices are justified (why this approach over alternatives?)
• Limitations are acknowledged explicitly
• Assumptions are stated clearly
• Reproducibility details are sufficient (parameters, datasets, metrics)

Common Issues

Usage in Claude Code

Check logical coherence in my introduction section

Analyze methodological depth in the methods section

Add transition signals between paragraphs in Section 3

Script CLI (MVP):

python scripts/analyze_logic.py main.tex
python scripts/analyze_logic.py main.tex --section methods

Translation Module (Chinese → English)

Supported Domains

Translation Workflow

1. Terminology Confirmation - Identify terms and confirm translations
2. Structure Analysis - Analyze paragraph structure, determine tense
3. Sentence Translation - Translation with annotations
4. Chinglish Check - Detect and fix common errors
5. Academic Polish - Final review

Script CLI (MVP):

python scripts/translate_academic.py "本文提出了一种基于Transformer的方法" --domain deep-learning
python scripts/translate_academic.py input_zh.txt --domain industrial-control --output translation_report.md

Usage Examples

Basic Translation Request:

Translate the following to academic English (Deep Learning domain):
本文提出了一种基于Transformer的时间序列预测方法...

With Venue Specification:

Translate the following for IEEE Transactions format:
实验结果表明，我们的方法在多个数据集上取得了最优性能...

Bibliography Module

Usage in Claude Code

Ask the assistant to verify your bibliography:

Verify references.bib for format errors

Check references.bib against main.tex for unused entries

The assistant will check for:

• Required field completeness
• Duplicate entries
• Unused entries
• Citation format consistency

Script CLI:

python scripts/verify_bib.py references.bib
python scripts/verify_bib.py references.bib --tex main.tex
python scripts/verify_bib.py references.bib --tex main.tex --json
# With online verification (CrossRef + Semantic Scholar)
python scripts/verify_bib.py references.bib --online
python scripts/verify_bib.py references.bib --online --email you@example.com

Result includes missing_in_bib and unused_in_tex for citation consistency.

De-AI Polishing Module

Reduce AI-generated writing traces while preserving LaTeX syntax and technical accuracy.

Features

• AI trace detection with pattern matching
• Section-wise analysis with density scores
• Batch processing for entire chapters
• Syntax-preserving editing (LaTeX commands, math, citations)

Usage in Claude Code

Interactive analysis (single section):

Analyze AI writing traces in my introduction section

Full document analysis:

Check AI trace density across all sections in paper.tex

Reduce AI traces:

Reduce AI writing traces in the methods section while preserving technical accuracy

Batch processing:

Process all sections in paper.tex to reduce AI traces

Output Example

================================================================================
DE-AI WRITING TRACE ANALYSIS REPORT
================================================================================
File: paper.tex
Total lines: 450

--------------------------------------------------------------------------------
SECTION-WISE AI TRACE DENSITY
--------------------------------------------------------------------------------

[HIGH] INTRODUCTION
  AI trace density: 8.5%
  Traces found: 12 / 141 lines

[MEDIUM] METHODS
  AI trace density: 3.2%
  Traces found: 5 / 156 lines

Reference Documentation

See references/DEAI_GUIDE.md for:

• Common AI patterns to remove
• Section-specific guidelines
• Output format specifications
• Quick reference replacements

Reference Files

• references/TERMINOLOGY.md: Domain terminology (Deep Learning, Time Series, Industrial Control)
• references/TRANSLATION_GUIDE.md: Translation principles, Chinglish corrections
• references/STYLE_GUIDE.md: Academic writing rules
• references/COMMON_ERRORS.md: Common mistakes
• references/VENUES.md: Conference/journal requirements
• references/DEAI_GUIDE.md: De-AI writing guide and AI pattern detection
• references/CITATION_VERIFICATION.md: Citation verification workflow
• scripts/check_references.py: Reference integrity checker (standalone)
• scripts/online_bib_verify.py: Online bibliography verifier

Title Optimization Module

Generate and optimize paper titles following IEEE/ACM/Springer/NeurIPS best practices.

Key Principles

Based on IEEE Author Center and top-tier venue guidelines:

1. Conciseness: Remove "A Study of", "Research on", "Novel", "New", "Improved"
2. Searchability: Place key terms (Method + Problem) in first 65 characters
3. Length: Optimal 10-15 words; acceptable 8-20 words
4. Specificity: Use concrete method/problem names, avoid vague terms
5. Jargon-Free: Avoid obscure abbreviations (except AI, LSTM, DNA, etc.)

Quality Scoring

Each title receives a score (0-100) based on:

• Conciseness (25%): No ineffective words
• Searchability (30%): Key terms in first 65 characters
• Length (15%): Within optimal range
• Specificity (20%): Concrete vs vague terms
• Jargon-Free (10%): No obscure abbreviations

Usage in Claude Code

Check existing title:

Check the quality of my paper title

Generate title candidates:

Generate title candidates for my paper based on the abstract

Optimize existing title:

Optimize my paper title to follow IEEE best practices

The assistant will analyze your title and provide:

• Quality score with breakdown
• Specific issues detected
• Multiple improved candidates (ranked by score)
• Suggested LaTeX code

Script CLI:

python scripts/optimize_title.py main.tex --check
python scripts/optimize_title.py main.tex --generate
python scripts/optimize_title.py main.tex --optimize
python scripts/optimize_title.py main.tex --compare "Title A" "Title B" "Title C"
python scripts/optimize_title.py "papers/*.tex" --batch --output title_report.json

Title Generation Workflow

1. Content Analysis: Extract problem, method, domain, key results from abstract/introduction
2. Keyword Extraction: Identify 3-5 core keywords (method + problem + domain)
3. Template Selection: Choose appropriate pattern (Method for Problem, Problem via Method, etc.)
4. Candidate Generation: Create 3-5 variants with different emphasis
5. Quality Scoring: Rank candidates by quality score

Common Title Patterns

Ineffective Words to Remove

Good vs Bad Examples

Good: "Transformer for Time Series Forecasting in Industrial Control"
Bad:  "A Novel Study on Improved Time Series Forecasting Using Transformers"

Good: "Graph Neural Networks for Fault Detection"
Bad:  "Research on Novel Fault Detection Based on GNNs"

Good: "Attention-Based LSTM for Multivariate Time Series Prediction"
Bad:  "An Improved LSTM Model Using Attention Mechanism for Prediction"

Venue-Specific Guidelines

IEEE Transactions:

• Avoid formulas with subscripts (except simple ones)
• Use title case (capitalize major words)
• Typical length: 10-15 words

ACM Conferences:

• More flexible with creative titles
• Can use colons for subtitles
• Typical length: 8-12 words

Springer Journals:

• Prefer descriptive over creative
• Can be slightly longer (up to 20 words)

NeurIPS/ICML:

• Concise and impactful (8-12 words)
• Method name often prominent

Best Practices

1. Start with keywords: Put Method + Problem in first 10 words
2. Be specific: "Transformer" > "Deep Learning" > "Machine Learning"
3. Remove fluff: Delete "Novel", "Study", "Research", "Based on"
4. Check length: Aim for 10-15 words (English)
5. Test searchability: Would you find this paper with these keywords?
6. Avoid jargon: Unless widely recognized (AI, LSTM, CNN)
7. Match venue style: IEEE (descriptive), ACM (creative), NeurIPS (concise)

References

• IEEE Author Center
• Royal Society Blog on Title Optimization

Recommended Workflows

Fast Pre-Submission Check

1. Format check (strict mode)
2. Grammar analysis (abstract + introduction)
3. Bibliography verification

Full Quality Pass

1. Format check → fix critical issues
2. Grammar analysis
3. De-AI polishing
4. Sentence decomposition
5. Expression restructuring
6. Bibliography verification

Translation Pipeline

1. Terminology confirmation
2. Translation with annotations
3. Chinglish check
4. Academic polish