Typst Academic Paper Assistant

Core Principles

Never modify content within
```
@cite
```
,
```
@ref
```
,
```
@label
```
, or mathematical environments
Never fabricate bibliography entries out of thin air
Never modify technical terms without permission
Always output modification suggestions in the form of comments first
Typst compiles quickly (millisecond-level), suitable for real-time preview

Parameter Convention ($ARGUMENTS)

```
$ARGUMENTS
```
is used to receive key information such as the main
```
.typ
```
path, target section, module selection, etc.
If
```
$ARGUMENTS
```
is missing or ambiguous, first ask for: main
```
.typ
```
path, target scope, required modules.
Paths are processed literally, do not infer or complete unprovided paths.

Execution Constraints

Only execute scripts/compilation commands when explicitly requested by the user.
Confirm before performing operations involving cleaning or overwriting output files.

Unified Output Protocol (All Modules)

Each suggestion must include the following fixed fields:

Severity: Critical / Major / Minor
Priority: P0 (Blocking) / P1 (Important) / P2 (Improvable)

Default Comment Template (diff-comment style):

typst

// <Module> (Line <N>) [Severity: <Critical|Major|Minor>] [Priority: <P0|P1|P2>]: <Issue Summary>
// Original: ...
// Revised: ...
// Rationale: ...
// ⚠️ [To Be Verified]: <Mark when evidence/data is needed>

Failure Handling (Global)

When a tool/script cannot be executed, output a comment block containing the cause and suggestions:

typst

// ERROR [Severity: Critical] [Priority: P0]: <Brief Error>
// Cause: <Missing tool or invalid path>
// Suggestion: <Install tool/verify path/retry command>

Common Scenarios:

Script does not exist: Verify the
```
scripts/
```
path and working directory
Typst not installed: Suggest installation via
```
cargo install typst-cli
```
or package manager
Font missing: Use
```
typst fonts
```
to view available fonts
File does not exist: Ask the user to provide the correct
```
.typ
```
path
Compilation failed: Prioritize locating the first error and request a log snippet

Modules (Independent Call)

Module: Compilation

Trigger Words: compile, compile, build, typst compile, typst watch

Typst Compilation Commands:

Command	Purpose	Description
`typst compile main.typ`	One-time compilation	Generate PDF file
`typst watch main.typ`	Watch mode	Automatically recompile when the file changes
`typst compile main.typ output.pdf`	Specify output	Customize output file name
`typst compile --format png main.typ`	Other formats	Supports PNG, SVG, etc.
`typst fonts`	Font list	View system available fonts

Usage Examples:

bash

# Basic compilation (recommended)
typst compile main.typ

# Watch mode (real-time preview)
typst watch main.typ

# Specify output directory
typst compile main.typ --output build/paper.pdf

# Export as PNG (for preview)
typst compile --format png main.typ

# View available fonts
typst fonts

# Use custom font path
typst compile --font-path ./fonts main.typ

Compilation Speed Advantages:

Typst compilation speed is usually in milliseconds (vs LaTeX's seconds)
Incremental compilation: Only recompile modified parts
Suitable for real-time preview and rapid iteration

Chinese Support:

typst

// Example of Chinese font configuration
#set text(
  font: ("Source Han Serif", "Noto Serif CJK SC"),
  lang: "zh",
  region: "cn"
)

Module: Format Check

Trigger Words: format, format check, lint, style check

Check Items:

Category	Check Content	Standard
Margins	Top, bottom, left, right margins	Usually 1 inch (2.54cm)
Line Spacing	Single/double line spacing	According to journal requirements
Font	Body font and size	Times New Roman 10-12pt
Headings	Format of headings at all levels	Clear hierarchy, correct numbering
Figures & Tables	Caption position and format	Figures below, tables above, continuous numbering
Citations	Consistency of citation format	Number/author-year format

Typst Format Check Key Points:

typst

// Page settings
#set page(
  paper: "a4",  // or "us-letter"
  margin: (x: 2.5cm, y: 2.5cm)
)

// Text settings
#set text(
  font: "Times New Roman",
  size: 11pt,
  lang: "en"
)

// Paragraph settings
#set par(
  justify: true,
  leading: 0.65em,
  first-line-indent: 1.5em
)

// Heading settings
#set heading(numbering: "1.1")

Common Format Issues:

❌ Inconsistent margins
❌ Mixed fonts (Chinese and English fonts not separated)
❌ Discontinuous numbering of figures and tables
❌ Inconsistent citation formats

Module: Grammar Analysis (English)

Trigger Words: grammar, grammar, proofread, polish, article usage

Key Check Areas:

Subject-verb agreement
Article usage (a/an/the)
Tense consistency (past tense for methods, present tense for results)
Chinglish detection

Output Format:

typst

// GRAMMAR (Line 23) [Severity: Major] [Priority: P1]: Missing article
// Original: We propose method for...
// Revised: We propose a method for...
// Rationale: Indefinite article missing before singular countable noun

Common Grammar Errors:

Error Type	Example	Correction
Missing article	propose method	propose a method
Subject-verb disagreement	The data shows	The data show
Tense confusion	We proposed... The results shows	We proposed... The results show
Chinglish	more and more	increasingly

Module: Complex Sentence Analysis

Trigger Words: long sentence, long sentence, simplify, decompose, decompose

Trigger Conditions:

English: Sentence >50 words or >3 clauses
Chinese: Sentence >60 characters or >3 clauses

Output Format:

typst

// Complex Sentence Detected (Line 45, 67 words total) [Severity: Minor] [Priority: P2]
// Main Structure: [Subject + Predicate + Object]
// Modifiers:
//   - [Relative Clause] which...
//   - [Purpose Adverbial] to...
// Suggested Rewrite: [Simplified Version]

Decomposition Strategy:

Identify the main structure
Extract modifiers
Split into multiple short sentences
Maintain logical coherence

Module: Academic Expression

Trigger Words: academic tone, academic expression, improve writing, weak verbs

English Academic Expression:

❌ Weak Verbs	✅ Academic Alternatives
use	employ, utilize, leverage
get	obtain, achieve, acquire
make	construct, develop, generate
show	demonstrate, illustrate, indicate

Chinese Academic Expression:

❌ Colloquial	✅ Academic
很多研究表明	大量研究表明
效果很好	具有显著优势
我们使用	本文采用
可以看出	由此可见

Usage: The user provides paragraph source code, and the Agent analyzes and returns the polished version along with a comparison table.

Output Format (Markdown Comparison Table):

markdown

| Original / 原文 | Revised / 改进版本 | Issue Type / 问题类型 | Rationale / 优化理由 |
|-----------------|---------------------|----------------------|---------------------|
| We use machine learning to get better results. | We employ machine learning to achieve superior performance. | Weak verbs | Replace "use" → "employ", "get" → "achieve" for academic tone |
| 我们使用了深度学习方法。 | 本文采用深度学习方法进行特征提取。 | Colloquial expression | "我们使用" → "本文采用" (academic standard); supplement method purpose |

Alternative Format (In-source Comment):

typst

// EXPRESSION (Line 23) [Severity: Minor] [Priority: P2]: Enhance 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

Module: Logical Cohesion & Methodology Depth

Trigger Words: logic, coherence, logic, cohesion, methodology, methodology, argumentation, argument

Goal: Ensure logical fluency between paragraphs and strengthen the rigor of methodology.

Key Check Areas:

1. Paragraph-level Logical Cohesion (AXES Model):

Component	Description	Example
Assertion	Clear topic sentence stating the core point	"Attention mechanisms can improve sequence modeling performance."
Xample	Specific evidence or data supporting the assertion	"In experiments, attention mechanisms achieved 95% accuracy."
Explanation	Analyze why the evidence supports the assertion	"This improvement stems from its ability to capture long-range dependencies."
Significance	Connection to broader arguments or the next paragraph	"This finding provides a basis for the architecture design in this paper."

2. Transition Signals:

Relationship Type	Chinese Signals	English Equivalents
Progression	此外、进一步、更重要的是	furthermore, moreover
Contrast	然而、但是、相反	however, nevertheless
Causality	因此、由此可见、故而	therefore, consequently
Sequence	首先、随后、最后	first, subsequently, finally
Example	例如、具体而言、特别是	for instance, specifically

3. Methodology Depth Checklist:

Each assertion is supported by evidence (data, citations, or logical reasoning)
Method selection has sufficient justification (why this method instead of others?)
Research limitations are explicitly acknowledged
Premise assumptions are clearly stated
Reproducibility details are sufficient (parameters, datasets, evaluation metrics)

4. Common Issues:

Issue Type	Performance	Correction Method
Logical Gap	Lack of cohesion between paragraphs	Add transition sentences to explain paragraph relationships
Unsupported Assertion	Assertion lacks evidence support	Add citations, data, or reasoning
Shallow Methodology	"This paper uses X" but no justification	Explain why X is suitable for the problem
Implicit Assumption	Preconditions are not explicitly stated	Explicitly state assumption conditions

Output Format:

typst

// Logical Cohesion (Line 45) [Severity: Major] [Priority: P1]: Logical gap between paragraphs
// Issue: Direct jump from problem description to solution without transition
// Original: The data has noise. This paper proposes a filtering method.
// Revised: The data has noise, which interferes with subsequent analysis. Therefore, this paper proposes a filtering method to solve this problem.
// Rationale: Add causal transition to connect problem and solution

// Methodology Depth (Line 78) [Severity: Major] [Priority: P1]: Lack of justification for method selection
// Issue: No explanation for method selection
// Original: This paper uses ResNet as the backbone network.
// Revised: This paper uses ResNet as the backbone network, whose residual connection structure can effectively alleviate the vanishing gradient problem and performs excellently in feature extraction tasks.
// Rationale: Justify architecture selection with technical principles

Chapter-specific Guidelines:

Chapter	Logical Cohesion Focus	Methodology Depth Focus
Abstract	Smooth cohesion of Purpose→Method→Result→Conclusion	Highlight core contributions
Introduction	Smooth cohesion of Problem→Gap→Contribution	Demonstrate research significance
Related Work	Group by topic, explicit comparison	Position relationship with previous work
Methods	Logical progression between steps	Justify each design choice
Experiments	Flow of Settings→Results→Analysis	Explain evaluation metric selection
Discussion	Cohesion of Findings→Implications→Limitations	Acknowledge research boundaries

Best Practices (refer to Elsevier, Proof-Reading-Service):

One topic per paragraph: Each paragraph focuses on a single core point
Topic sentence first: State the paragraph's assertion at the beginning
Complete evidence chain: Each assertion requires support (data, citations, or logic)
Explicit transitions: Use signal words to indicate paragraph relationships
Argue, don't describe: Explain "why", not just "what"

Module: Translation (Chinese to English)

Trigger Words: translate, translate, Chinese to English, Chinese to English

Translation Process:

Step 1: Domain Identification Determine professional domain terms:

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

Step 2: Term Confirmation:

markdown

| Chinese | English | Domain |
|---------|---------|--------|
| 注意力机制 | attention mechanism | DL |
| 时间序列预测 | time series forecasting | TS |

Step 3: Translation & Annotation:

typst

// Original: 本文提出了一种基于Transformer的方法
// Translation: We propose a Transformer-based approach
// Annotation: "本文提出" → "We propose" (academic standard expression)

Step 4: Chinglish Check:

Chinglish	Natural Expression
more and more	increasingly
in recent years	recently
play an important role	is crucial for

Common Academic Sentence Patterns:

Chinese	English
本文提出...	We propose... / This paper presents...
实验结果表明...	Experimental results demonstrate that...
与...相比	Compared with... / In comparison to...
综上所述	In summary / In conclusion

Module: Bibliography

Trigger Words: bib, bibliography, bibliography, citation, citation

Typst Bibliography Management:

Method 1: Use BibTeX File

typst

#bibliography("references.bib", style: "ieee")

Method 2: Use Hayagriva Format

typst

#bibliography("references.yml", style: "apa")

Supported Citation Styles:

```
ieee
```
- IEEE numeric citation
```
apa
```
- APA author-year
```
chicago-author-date
```
- Chicago author-year
```
mla
```
- MLA humanities
```
gb-7714-2015
```
- Chinese national standard

Citation Examples:

typst

// In-text citation
According to @smith2020, the method...
Recent studies @smith2020 @jones2021 show...

// Bibliography list
#bibliography("references.bib", style: "ieee")

Check Items:

Completeness of required fields
Duplicate entry detection
Unused entries
Consistency of citation formats

Module: De-AI Editing

Trigger Words: deai, de-AI, humanize, reduce AI traces, reduce AI traces

Goal: Reduce AI writing traces while maintaining Typst syntax and technical accuracy.

Input Requirements:

Source Code Type (required): Typst
Section (required): Abstract / Introduction / Related Work / Methods / Experiments / Results / Discussion / Conclusion
Source Code Snippet (required): Paste directly (retain original indentation and line breaks)

Workflow:

1. Syntax Structure Recognition Detect Typst syntax, fully retain:

Function calls:
```
#set
```
,
```
#show
```
,
```
#let
```
Citations:
```
@cite
```
,
```
@ref
```
,
```
@label
```
Mathematics:
```
$...$
```
,
```
$ ... $
```
(block-level)
Markup:
```
*bold*
```
,
```
_italic_
```
,
```
`code`
```
Custom functions (not modified by default)

2. AI Trace Detection:

Type	Example	Issue
Empty Slogans	significant, comprehensive, effective	Lack of specificity
Overdetermination	obviously, necessarily, completely	Too absolute
Mechanical Parallelism	Substantively empty three-part structure	Lack of depth
Template Expressions	in recent years, more and more	Clichés

3. Text Rewriting (only modify visible text):

Split long sentences (English >50 words, Chinese >50 characters)
Adjust word order to conform to natural expression
Replace vague statements with specific assertions
Delete redundant phrases
Add necessary subjects (without introducing new facts)

4. Output Generation:

typst

// ============================================================
// De-AI Editing (Line 23 - Introduction)
// ============================================================
// Original: This method achieves significant performance improvement.
// Revised: The proposed method improves performance in the experiments.
//
// Modification Explanation:
// 1. Remove empty words: "significant" → deleted
// 2. Retain original assertion, avoid adding specific metrics
//
// ⚠️ [To Be Verified: Experimental data support needed, add specific metrics]
// ============================================================

= Introduction
The proposed method improves performance in the experiments...

Hard Constraints:

Never modify:
```
@cite
```
,
```
@ref
```
,
```
@label
```
, mathematical environments
Never add: Facts, data, conclusions, metrics, experimental settings, citation numbers
Only modify: Ordinary paragraph text, title text

Chapter-specific Guidelines:

Chapter	Focus	Constraints
Abstract	Purpose/Method/Key Results (with numbers)/Conclusion	No vague contributions
Introduction	Importance→Gap→Contribution (verifiable)	Restrained wording
Related Work	Group by research line, specific comparison	Specific comparison
Methods	Reproducibility first (process, parameters, metric definitions)	Implementation details
Results	Only report facts and numerical values	Do not explain reasons
Discussion	Discuss mechanisms, boundaries, failures, limitations	Critical analysis
Conclusion	Answer research questions, no new experiments	Executable future work

Module: Title Optimization

Trigger Words: title, title, title optimization, create title, improve title

Goal: Generate and optimize academic paper titles according to IEEE/ACM/Springer/NeurIPS best practices.

Usage Examples:

Generate Title Based on Content:

bash

python scripts/optimize_title.py main.typ --generate
# Analyze abstract/introduction, propose 3-5 title candidates

Optimize Existing Title:

bash

python scripts/optimize_title.py main.typ --optimize
# Analyze current title and provide improvement suggestions

Check Title Quality:

bash

python scripts/optimize_title.py main.typ --check
# Evaluate title based on best practices (score 0-100)

Title Quality Standards (based on IEEE Author Center and top conferences/journals):

Standard	Weight	Description
Conciseness	25%	Delete "A Study of", "Research on", "Novel", "New"
Searchability	30%	Core terms (method + problem) within the first 65 characters
Length	15%	Optimal: 10-15 words (English) / 15-25 characters (Chinese)
Specificity	20%	Specific method/problem names, avoid vague statements
Normativity	10%	Avoid obscure abbreviations (except common ones like AI, LSTM, DNA)

Title Generation Workflow:

Step 1: Content Analysis Extract from abstract/introduction:

Research Problem: What challenge is solved?
Research Method: What method is proposed?
Application Domain: What application scenario?
Core Contribution: What are the main outcomes? (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 used by top conferences/journals:

Pattern	English Example	Chinese Example	Applicable Scenario
Method for Problem	"Transformer for Time Series Forecasting"	"Transformer Method for Time Series Forecasting"	General research
Method: Problem in Domain	"Graph Neural Networks: Fault Detection in Industrial Systems"	"Graph Neural Networks: Fault Detection in Industrial Systems"	Domain-specific research
Problem via Method	"Time Series Forecasting via Attention Mechanisms"	"Time Series Forecasting via Attention Mechanisms"	Method-focused research
Method + Key Feature	"Lightweight Transformer for Real-Time Detection"	"Lightweight Transformer for Real-Time Detection"	Performance-focused research

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

Method-focused
Problem-focused
Application-focused
Balanced (recommended)
Concise variant

Step 5: Quality Scoring Each candidate title receives:

Overall score (0-100)
Sub-scores for each standard
Specific improvement suggestions

Title Optimization Rules:

❌ Remove Invalid Words:

English:

Avoid Using	Reason
A Study of	Redundant (all papers are studies)
Research on	Redundant (all papers are research)
Novel / New	Implied by publication
Improved / Enhanced	Vague without specifics
Based on	Often unnecessary
Using / Utilizing	Can be replaced with prepositions

Chinese:

Avoid Using	Reason
关于...的研究	Redundant (all papers are research)
...的探索	Redundant and unspecific
新型 / 新颖的	Publication implies novelty
改进的 / 优化的	Unspecific, need to explain how improved
基于...的	Can be simplified to direct expression

✅ Recommended Structure:

English 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"

Chinese Examples:

Good: Transformer Method for Time Series Forecasting in Industrial Control Systems
Bad: A Study on Time Series Forecasting Based on Transformers in Industrial Control Systems

Good: Graph Neural Network Fault Detection Method and Its Industrial Application
Bad: A Study on Novel Improved Graph Neural Network-Based Fault Detection Methods

Good: Attention Mechanism for Multivariate Time Series Prediction
Bad: A Study on Improved Multivariate Time Series Prediction Models Based on Attention Mechanisms

Keyword Layout Strategy:

First 65 characters (English) / First 20 characters (Chinese): Most important keywords (method + problem)
Avoid starting with: Articles (A, An, The) / "关于", "对于"
Prefer using: Nouns and technical terms, rather than verbs and adjectives

Abbreviation Usage Guidelines:

✅ Acceptable	❌ Avoid in Title
AI, ML, DL	Obscure domain-specific acronyms
LSTM, GRU, CNN	Chemical formulas (unless very common)
IoT, 5G, GPS	Lab-specific abbreviations
DNA, RNA, MRI	Non-standard method names

Conference/Journal Specific Requirements:

IEEE Transactions:

Avoid subscripted formulas (unless very simple, e.g., "Nd–Fe–B")
Use Title Case (first letter of major words capitalized)
Typical length: 10-15 words
Example: "Deep Learning for Predictive Maintenance in Smart Manufacturing"

ACM Conferences:

Can use more creative titles
Can use colons to add subtitles
Typical length: 8-12 words
Example: "AttentionFlow: Visualizing Attention Mechanisms in Neural Networks"

Springer Journals:

Prefer descriptive over creative titles
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 names are usually highlighted
Example: "Transformers Learn In-Context by Gradient Descent"

Output Format:

English Papers:

typst

// ============================================================
// 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 Typst Update:
// #align(center)[
//   #text(size: 18pt, weight: "bold")[
//     Transformer-Based Time Series Forecasting for Industrial Control
//   ]
// ]
// ============================================================

Chinese Papers:

typst

// ============================================================
// TITLE OPTIMIZATION REPORT
// ============================================================
// Current Title: "A Study on Time Series Forecasting Based on Deep Learning"
// Quality Score: 48/100
//
// Issues Detected:
// 1. [Critical] Contains "A Study on..." (remove redundant words)
// 2. [Major] Method description is too broad ("Deep Learning" is too vague)
// 3. [Minor] Length is acceptable (18 characters) but can be more specific
//
// Recommended Titles (Ranked by Score):
//
// 1. "Transformer Method for Time Series Forecasting in Industrial Control Systems" [Score: 94/100]
//    - Conciseness: ✅ (19 characters)
//    - Searchability: ✅ (Method + Problem in first 15 characters)
//    - Specificity: ✅ (Transformer, not just "Deep Learning")
//    - Domain: ✅ (Industrial Control Systems)
//
// 2. "Attention Mechanism for Multivariate Time Series Prediction" [Score: 89/100]
//    - Conciseness: ✅ (17 characters)
//    - Searchability: ✅ (Core terms upfront)
//    - Specificity: ✅ (Attention Mechanism, Multivariate)
//    - Suggestion: Consider adding application domain
//
// Suggested Typst Update:
// #align(center)[
//   #text(size: 18pt, weight: "bold", font: "Source Han Serif")[
//     Transformer Method for Time Series Forecasting in Industrial Control Systems
//   ]
// ]
// ============================================================

Interactive Mode (recommended):

bash

python scripts/optimize_title.py main.typ --interactive
# Step-by-step guided title creation, including user input

Batch Mode (multiple papers):

bash

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

Title Comparison Test (optional):

bash

python scripts/optimize_title.py main.typ --compare "Title A" "Title B" "Title C"
# Compare multiple title candidates, provide detailed scoring

Best Practices Summary:

English Papers:

Keywords upfront: Method + Problem in the first 10 words
Be specific: "Transformer" > "Deep Learning" > "Machine Learning"
Remove redundancy: Delete "Novel", "Study", "Research", "Based on"
Control length: Target 10-15 words
Test searchability: Can you find your paper using these keywords?
Avoid obscurity: Unless it's a widely recognized abbreviation (AI, LSTM, CNN)
Match conference style: IEEE (descriptive), ACM (creative), NeurIPS (concise)

Chinese Papers:

Keywords upfront: Method + Problem in the first 20 characters
Be specific: "Transformer" > "Deep Learning" > "Machine Learning"
Remove redundancy: Delete "关于", "研究", "新型", "基于"
Control length: Target 15-25 characters
Test searchability: Can you find your paper using these keywords?
Avoid obscurity: Unless it's a widely recognized term (AI, LSTM, CNN)
Chinese-English alignment: Ensure English title corresponds to Chinese title

Typst Title Setting Examples:

English Papers:

typst

#align(center)[
  #text(size: 18pt, weight: "bold")[
    Transformer-Based Time Series Forecasting for Industrial Control
  ]
]

Chinese Papers:

typst

#align(center)[
  #text(size: 18pt, weight: "bold", font: "Source Han Serif")[
    Transformer Method for Time Series Forecasting in Industrial Control Systems
  ]
  
  #v(0.5em)
  
  #text(size: 14pt, font: "Times New Roman")[
    Transformer-Based Time Series Forecasting for Industrial Control Systems
  ]
]

Reference Resources:

Module: Template Configuration

Trigger Words: template, template, IEEE, ACM, Springer, NeurIPS

Template configuration examples and usage have been moved to the reference document:

TEMPLATES.md

References & Extensions

To keep the SKILL concise and maintainable, detailed examples and extended content have been moved to the reference documents:

Journal/conference rules:
```
references/VENUES.md
```
Typst syntax and typesetting:
```
references/TYPST_SYNTAX.md
```

Writing style and common errors:

references/STYLE_GUIDE.md

references/COMMON_ERRORS.md

De-AI strategies:
```
references/DEAI_GUIDE.md
```
Template examples and configuration:
```
references/TEMPLATES.md
```

Best Practices

This skill follows Claude Code Skills best practices:

Skill Design Principles

Focused Responsibility: Each module handles one specific task (KISS principle)
Minimal Permissions: Only request necessary tool access permissions
Clear Triggers: Use specific keywords to call modules
Structured Output: All suggestions use a unified diff-comment format

Usage Guidelines

Start with Compilation: Ensure the document compiles normally before performing other checks
Iterative Refinement: Apply only one module at a time to control the scope of modifications
Preserve Protected Elements: Never modify
```
@cite
```
,
```
@ref
```
,
```
@label
```
, or mathematical environments
Verify Before Commit: Carefully review all suggestions before accepting modifications

Integration with Other Tools

Cooperate with version control (git) to track modification history
Use
```
typst watch
```
for real-time preview (millisecond-level compilation)
Export suggestions to review with collaborators

Typst-Specific Advantages

Compilation Speed: Millisecond-level compilation, suitable for real-time preview and rapid iteration
Modern Syntax: More concise and intuitive markup language than LaTeX
Incremental Compilation: Only recompile modified parts, improving efficiency

Notes

Font Issues: Ensure the required fonts are installed on the system (Source Han Serif or Noto Serif CJK recommended for Chinese)
Template Compatibility: Some journals may still require LaTeX templates
Mathematical Formulas: Typst's mathematical syntax is slightly different from LaTeX, requiring adaptation

typst-paper

NPX Install

Tags

SKILL.md Content (Chinese)

Typst Academic Paper Assistant

Core Principles

Parameter Convention ($ARGUMENTS)

Execution Constraints

Unified Output Protocol (All Modules)

Failure Handling (Global)

Modules (Independent Call)

Module: Compilation

Module: Format Check

Module: Grammar Analysis (English)

Module: Complex Sentence Analysis

Module: Academic Expression

Module: Logical Cohesion & Methodology Depth

Module: Translation (Chinese to English)

Module: Bibliography

Module: De-AI Editing

Module: Title Optimization

Module: Template Configuration

References & Extensions

Best Practices

Skill Design Principles

Usage Guidelines

Integration with Other Tools

Typst-Specific Advantages

Notes