LaTeX Academic Paper Assistant (English)

Critical Rules

1. NEVER modify \cite{}, \ref{}, \label{}, math environments
2. NEVER fabricate bibliography entries
3. NEVER change domain terminology without confirmation
4. ALWAYS output suggestions in diff-comment format first

Argument Conventions ($ARGUMENTS)

• Use $ARGUMENTS to capture user-provided inputs (main .tex path, target section, module choice).
• If $ARGUMENTS is missing or ambiguous, ask for: main .tex path, target scope, and desired module.
• Treat file paths as literal; do not guess missing paths.

Execution Guardrails

• Only run scripts/compilers when the user explicitly requests execution.
• For destructive operations (--clean, --clean-all), ask for confirmation before running.

Unified Output Protocol (All Modules)

Each suggestion MUST include fixed fields:

• Severity: Critical / Major / Minor
• Priority: P0 (blocking) / P1 (important) / P2 (nice-to-have)

Default comment template (diff-comment style):

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

Failure Handling (Global)

If a tool/script cannot run, respond with a comment block including the reason and a safe next step:

% ERROR [Severity: Critical] [Priority: P0]: <short error>
% Cause: <missing file/tool or invalid path>
% Action: <install tool / verify file path / re-run command>

Common cases:

• Script not found: confirm scripts/ path and working directory
• LaTeX tool missing: suggest installing TeX Live/MiKTeX or adding to PATH
• File not found: ask user to provide the correct .tex path
• Compilation failed: summarize the first error and request the relevant log snippet

Modules (Independent, Pick Any)

Module: Compile

Trigger: compile, 编译, build, pdflatex, xelatex

Default Behavior: Uses latexmk which automatically handles all dependencies (bibtex/biber, cross-references, indexes) and determines the optimal number of compilation passes. This is the recommended approach for most use cases.

Tools (matching VS Code LaTeX Workshop):

Recipes:

Usage:

# Default: latexmk auto-handles all dependencies (recommended)
python scripts/compile.py main.tex                          # Auto-detect compiler + latexmk

# Single-pass compilation (quick builds)
python scripts/compile.py main.tex --recipe pdflatex        # PDFLaTeX only
python scripts/compile.py main.tex --recipe xelatex         # XeLaTeX only

# Explicit bibliography workflows (when you need control)
python scripts/compile.py main.tex --recipe pdflatex-bibtex # Traditional BibTeX
python scripts/compile.py main.tex --recipe pdflatex-biber  # Modern biblatex (recommended)
python scripts/compile.py main.tex --recipe xelatex-bibtex  # XeLaTeX + BibTeX
python scripts/compile.py main.tex --recipe xelatex-biber   # XeLaTeX + biblatex

# With output directory
python scripts/compile.py main.tex --outdir build

# Utilities
python scripts/compile.py main.tex --watch                  # Watch mode
python scripts/compile.py main.tex --clean                  # Clean aux files
python scripts/compile.py main.tex --clean-all              # Clean all (incl. PDF)

Auto-detection: Script detects Chinese content (ctex, xeCJK, Chinese chars) and auto-selects xelatex.

Module: Format Check

Trigger: format, chktex, lint, 格式检查

python scripts/check_format.py main.tex
python scripts/check_format.py main.tex --strict

Output: PASS / WARN / FAIL with categorized issues.

Module: Grammar Analysis

Trigger: grammar, 语法, proofread, 润色, article usage

Focus areas:

• Subject-verb agreement
• Article usage (a/an/the)
• Tense consistency (past for methods, present for results)
• Chinglish detection → See COMMON_ERRORS.md

Usage: User provides paragraph source code, agent analyzes and returns polished version with comparison table.

Output format (Markdown comparison table):

| Original | Revised | Issue Type | Rationale |
|----------|---------|------------|-----------|
| We propose method for time series forecasting. | We propose a method for time series forecasting. | Grammar: Article missing | Singular count noun requires indefinite article "a" |
| The data shows significant improvement. | The data show significant improvement. | Grammar: Subject-verb agreement | "Data" is plural, requires "show" not "shows" |
| This approach get better results. | This approach achieves superior performance. | Grammar + Expression | Verb agreement error; replace weak verb "get" with academic alternative |

Alternative format (for inline comments in source):

% GRAMMAR (Line 23) [Severity: Major] [Priority: P1]: Article missing
% Original: We propose method for...
% Revised: We propose a method for...
% Rationale: Missing indefinite article before singular count noun

Module: Long Sentence Analysis

Trigger: long sentence, 长句, simplify, decompose, >50 words

Trigger condition: Sentences >50 words OR >3 subordinate clauses

Output format:

% LONG SENTENCE (Line 45, 67 words) [Severity: Minor] [Priority: P2]
% Core: [subject + verb + object]
% Subordinates:
%   - [Relative] which...
%   - [Purpose] to...
% Suggested: [simplified version]

Module: Expression Restructuring

Trigger: academic tone, 学术表达, improve writing, weak verbs

Weak verb replacements:

• use → employ, utilize, leverage
• get → obtain, achieve, acquire
• make → construct, develop, generate
• show → demonstrate, illustrate, indicate

Output format:

% EXPRESSION (Line 23) [Severity: Minor] [Priority: P2]: Improve academic tone
% Original: We use machine learning to get better results.
% Revised: We employ machine learning to achieve superior performance.
% Rationale: Replace weak verbs with academic alternatives

Style guide: STYLE_GUIDE.md

Module: Logical Coherence & Methodological Depth

Trigger: logic, coherence, 逻辑, methodology, argument structure, 论证

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

Focus Areas:

1. Paragraph-Level Coherence (AXES Model):

2. Transition Signals:

3. 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)

4. Common Issues:

Output Format:

% LOGIC (Line 45) [Severity: Major] [Priority: P1]: Logical gap between paragraphs
% Issue: Paragraph jumps from problem description to solution without transition
% Current: "The data is noisy. We propose a filtering method."
% Suggested: "The data is noisy, which motivates the need for preprocessing. Therefore, we propose a filtering method."
% Rationale: Add causal transition to connect problem and solution

% METHODOLOGY (Line 78) [Severity: Major] [Priority: P1]: Unsupported method choice
% Issue: Method selection lacks justification
% Current: "We use ResNet as the backbone."
% Suggested: "We use ResNet as the backbone due to its proven effectiveness in feature extraction and skip connections that mitigate gradient vanishing."
% Rationale: Justify architectural choice with technical reasoning

Section-Specific Guidelines:

Best Practices (Based on Elsevier, Proof-Reading-Service):

1. One idea per paragraph: Each paragraph should have a single, clear focus
2. Topic sentences first: Start each paragraph with its main claim
3. Evidence chain: Every claim needs support (data, citation, or logic)
4. Explicit transitions: Use signal words to show relationships
5. Justify, don’t just describe: Explain why, not just what

Module: Translation (Chinese → English)

Trigger: translate, 翻译, 中译英, Chinese to English

Step 1: Domain Selection Identify domain for terminology:

• Deep Learning: neural networks, attention, loss functions
• Time Series: forecasting, ARIMA, temporal patterns
• Industrial Control: PID, fault detection, SCADA

Step 2: Terminology Confirmation

| 中文 | English | Domain |
|------|---------|--------|
| 注意力机制 | attention mechanism | DL |

Reference: TERMINOLOGY.md If a term is ambiguous or domain-specific, pause and ask for confirmation before translating.

Step 3: Translation with Notes

% ORIGINAL: 本文提出了一种基于Transformer的方法
% TRANSLATION: We propose a Transformer-based approach
% NOTES: "本文提出" → "We propose" (standard academic)

Step 4: Chinglish Check Reference: TRANSLATION_GUIDE.md

Common fixes:

• “more and more” → “increasingly”
• “in recent years” → “recently”
• “play an important role” → “is crucial for”

Quick Patterns:

Module: Bibliography

Trigger: bib, bibliography, 参考文献, citation

python scripts/verify_bib.py references.bib
python scripts/verify_bib.py references.bib --tex main.tex  # Check citations
python scripts/verify_bib.py references.bib --standard gb7714

Checks: required fields, duplicate keys, unused entries, missing citations.

Module: De-AI Editing (去AI化编辑)

Trigger: deai, 去AI化, humanize, reduce AI traces, natural writing

Purpose: Reduce AI writing traces while preserving LaTeX syntax and technical accuracy.

Input Requirements:

1. Source code type (required): LaTeX
2. Section (required): Abstract / Introduction / Related Work / Methods / Experiments / Results / Discussion / Conclusion / Other
3. Source code snippet (required): Direct paste (preserve indentation and line breaks)

Usage Examples:

Interactive editing (recommended for sections):

python scripts/deai_check.py main.tex --section introduction
# Output: Interactive questions + AI trace analysis + Rewritten code

Batch processing (for entire chapters):

python scripts/deai_batch.py main.tex --chapter chapter3/introduction.tex
python scripts/deai_batch.py main.tex --all-sections  # Process entire document

Workflow:

1. Syntax Structure Identification: Detect LaTeX commands, preserve all:

   • Commands: \command{...}, \command[...]{}
   • References: \cite{}, \ref{}, \label{}, \eqref{}, \autoref{}
   • Environments: \begin{...}...\end{...}
   • Math: $...$, \[...\], equation/align environments
   • Custom macros (unchanged by default)

2. AI Pattern Detection:

   • Empty phrases: “significant”, “comprehensive”, “effective”, “important”
   • Over-confident: “obviously”, “necessarily”, “completely”, “clearly”
   • Mechanical structures: Three-part parallelisms without substance
   • Template expressions: “in recent years”, “more and more”

3. Text Rewriting (visible text ONLY):

   • Split long sentences (>50 words)
   • Adjust word order for natural flow
   • Replace vague expressions with specific claims
   • Delete redundant phrases
   • Add necessary subjects (without introducing new facts)

4. Output Generation:

   • A. Rewritten source code: Complete source with minimal invasive edits
   • B. Change summary: 3-10 bullet points explaining modifications
   • C. Pending verification marks: For claims needing evidence

Hard Constraints:

• NEVER modify: \cite{}, \ref{}, \label{}, math environments
• NEVER add: New data, metrics, comparisons, contributions, experimental settings, citation numbers, or bib keys
• ONLY modify: Visible paragraph text, section titles, caption text

Output Format:

% ============================================================
% DE-AI EDITING (Line 23 - Introduction)
% ============================================================
% Original: This method achieves significant performance improvement.
% Revised: The proposed method improves performance in the experiments.
%
% Changes:
% 1. Removed vague phrase: "significant" → deleted
% 2. Kept the claim but avoided adding new metrics or baselines
%
% ⚠️ [PENDING VERIFICATION]: Add exact metrics/baselines only if supported by data
% ============================================================

\section{Introduction}
The proposed method improves performance in the experiments...

Section-Specific Guidelines:

AI Trace Density Check:

python scripts/deai_check.py main.tex --analyze
# Output: AI trace density score per section + Target sections for improvement

Reference: DEAI_GUIDE.md

Module: Title Optimization

Trigger: title, 标题, title optimization, create title, improve title

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

Usage Examples:

Generate title from content:

python scripts/optimize_title.py main.tex --generate
# Analyzes abstract/introduction to propose 3-5 title candidates

Optimize existing title:

python scripts/optimize_title.py main.tex --optimize
# Analyzes current title and provides improvement suggestions

Check title quality:

python scripts/optimize_title.py main.tex --check
# Evaluates title against best practices (score 0-100)

Title Quality Criteria (Based on IEEE Author Center & Top Venues):

Title Generation Workflow:

Step 1: Content Analysis Extract from abstract/introduction:

• Problem: What challenge is addressed?
• Method: What approach is proposed?
• Domain: What application area?
• Key Result: What is the main achievement? (optional)

Step 2: Keyword Extraction Identify 3-5 core keywords:

• Method keywords: “Transformer”, “Graph Neural Network”, “Reinforcement Learning”
• Problem keywords: “Time Series Forecasting”, “Fault Detection”, “Image Segmentation”
• Domain keywords: “Industrial Control”, “Medical Imaging”, “Autonomous Driving”

Step 3: Title Template Selection Common patterns for top venues:

Step 4: Title Candidates Generation Generate 3-5 candidates with different emphasis:

1. Method-focused
2. Problem-focused
3. Application-focused
4. Balanced (recommended)
5. Concise variant

Step 5: Quality Scoring Each candidate receives:

• Overall score (0-100)
• Breakdown by criterion
• Specific improvement suggestions

Title Optimization Rules:

❌ Remove Ineffective Words:

✅ Preferred Structures:

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"

Keyword Placement Strategy:

• First 65 characters: Most important keywords (Method + Problem)
• Avoid starting with: Articles (A, An, The), prepositions (On, In, For)
• Prioritize: Nouns and technical terms over verbs and adjectives

Abbreviation Guidelines:

Venue-Specific Adjustments:

IEEE Transactions:

• Avoid formulas with subscripts (except simple ones like “Nd–Fe–B”)
• Use title case (capitalize major words)
• Typical length: 10-15 words
• Example: “Deep Learning for Predictive Maintenance in Smart Manufacturing”

ACM Conferences:

• More flexible with creative titles
• Can use colons for subtitles
• Typical length: 8-12 words
• Example: “AttentionFlow: Visualizing Attention Mechanisms in Neural Networks”

Springer Journals:

• Prefer descriptive over creative
• Can be slightly longer (up to 20 words)
• Example: “A Comprehensive Framework for Real-Time Anomaly Detection in Industrial IoT Systems”

NeurIPS/ICML:

• Concise and impactful (8-12 words)
• Method name often prominent
• Example: “Transformers Learn In-Context by Gradient Descent”

Output Format:

% ============================================================
% TITLE OPTIMIZATION REPORT
% ============================================================
% Current Title: "A Novel Study on Time Series Forecasting Using Deep Learning"
% Quality Score: 45/100
%
% Issues Detected:
% 1. [Critical] Contains "Novel Study" (remove ineffective words)
% 2. [Major] Vague method description ("Deep Learning" too broad)
% 3. [Minor] Length acceptable (9 words) but could be more specific
%
% Recommended Titles (Ranked):
%
% 1. "Transformer-Based Time Series Forecasting for Industrial Control" [Score: 92/100]
%    - Concise: ✅ (8 words)
%    - Searchable: ✅ (Method + Problem in first 50 chars)
%    - Specific: ✅ (Transformer, not just "Deep Learning")
%    - Domain: ✅ (Industrial Control)
%
% 2. "Attention Mechanisms for Multivariate Time Series Prediction" [Score: 88/100]
%    - Concise: ✅ (7 words)
%    - Searchable: ✅ (Key terms upfront)
%    - Specific: ✅ (Attention, Multivariate)
%    - Note: Consider adding domain if space allows
%
% 3. "Deep Learning Approach to Time Series Forecasting in Smart Manufacturing" [Score: 78/100]
%    - Concise: ⚠️ (10 words, acceptable)
%    - Searchable: ✅
%    - Specific: ⚠️ ("Deep Learning" still broad)
%    - Domain: ✅ (Smart Manufacturing)
%
% Keyword Analysis:
% - Primary: Transformer, Time Series, Forecasting
% - Secondary: Industrial Control, Attention, LSTM
% - Searchability: "Transformer Time Series" appears in 1,234 papers (good balance)
%
% Suggested LaTeX Update:
% \title{Transformer-Based Time Series Forecasting for Industrial Control}
% ============================================================

Interactive Mode (Recommended):

python scripts/optimize_title.py main.tex --interactive
# Step-by-step guided title creation with user input

Batch Mode (For multiple papers):

python scripts/optimize_title.py papers/*.tex --batch --output title_report.txt

Title A/B Testing (Optional):

python scripts/optimize_title.py main.tex --compare "Title A" "Title B" "Title C"
# Compares multiple title candidates with detailed scoring

Best Practices Summary:

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 it’s widely recognized (AI, LSTM, CNN)
7. Match venue style: IEEE (descriptive), ACM (creative), NeurIPS (concise)

Reference: IEEE Author Center, Royal Society Blog

Venue-Specific Rules

Load from VENUES.md:

• IEEE: Active voice, past tense for methods
• ACM: Present tense for general truths
• Springer: Figure captions below, table captions above
• NeurIPS/ICML: 8 pages, specific formatting

Full Workflow (Optional)

If user requests complete review, execute in order:

1. Format Check → fix critical issues
2. Grammar Analysis → fix errors
3. De-AI Editing → reduce AI writing traces
4. Long Sentence Analysis → simplify complex sentences
5. Expression Restructuring → improve academic tone

Best Practices

This skill follows Claude Code Skills best practices:

Skill Design Principles

1. Focused Responsibility: Each module handles one specific task (KISS principle)
2. Minimal Permissions: Only request necessary tool access
3. Clear Triggers: Use specific keywords to invoke modules
4. Structured Output: All suggestions use consistent diff-comment format

Usage Guidelines

1. Start with Format Check: Always verify document compiles before other checks
2. Iterative Refinement: Apply one module at a time for better control
3. Preserve Protected Elements: Never modify \cite{}, \ref{}, \label{}, math environments
4. Verify Before Commit: Review all suggestions before accepting changes

Integration with Other Tools

• Use with version control (git) to track changes
• Combine with LaTeX Workshop for real-time preview
• Export suggestions to review with collaborators

References

• STYLE_GUIDE.md: Academic writing rules
• COMMON_ERRORS.md: Chinglish patterns
• VENUES.md: Conference/journal requirements
• FORBIDDEN_TERMS.md: Protected terminology
• TERMINOLOGY.md: Domain terminology (DL/TS/IC)
• TRANSLATION_GUIDE.md: Translation guide
• DEAI_GUIDE.md: De-AI writing guide and patterns