text-optimize

Text Optimize

Trigger Commands: /optimize Trigger Keywords: polish, optimize, improve, enhance, expand, shorten, remove AI flavor, humanize

Introduction

The text optimization skill suite integrates four major functions: text polishing, AI-flavor removal, content expansion, and content shortening. It provides comprehensive text quality improvement and length adjustment services.

Feature 1: Polish Text

Polish Text - Text Polishing

Introduction

Optimize the expression quality of text, improve language fluency, accuracy, and readability, making the content more professional and elegant. After polishing is complete, humanize_text is automatically invoked to remove AI traces.

Use Cases

A draft needs quality improvement
Language expression is not fluent or accurate enough
The professionalism of a document needs to be improved
Content to be submitted or published needs final optimization
Localization polishing of translated text

Input Parameters

Required Parameters

source_text (string): The original text to be polished

Optional Parameters

focus_areas (string[]): Key areas for polishing (multiple selections allowed)
- grammar - Grammar correction
- vocabulary - Vocabulary optimization
- structure - Structure adjustment
- flow - Fluency improvement
- clarity - Clarity of expression
- tone_consistency - Tone consistency
- Default value: ["grammar", "vocabulary", "flow", "clarity"]
target_tone (string): Target tone
- formal - Formal
- semi-formal - Semi-formal (recommended)
- casual - Informal
- authoritative - Authoritative
- persuasive - Persuasive
- Default value: semi-formal
retain_original_structure (boolean): Whether to retain the original structure
- true - Do not adjust paragraph and sentence structure
- false - Structure can be optimized (recommended)
- Default value: false
language_level (string): Language complexity
- simple - Simple and easy to understand
- standard - Standard (recommended)
- advanced - Advanced and professional
- Default value: standard

Output

The polished text, with more fluent, accurate, and professional expression. Automatically invokes humanize_text to remove AI traces.

Processing Rules

Analysis Phase

Detect grammatical errors and non-standard expressions
Identify repetitive, redundant, or inaccurate word usage
Analyze sentence structure and paragraph organization
Evaluate overall fluency and readability
Check tone consistency

Polishing Strategies

1. Grammar Level

Correct grammatical errors
Optimize punctuation usage
Adjust subject-verb agreement
Fix improper tense

2. Vocabulary Level

Replace inaccurate or inappropriate vocabulary
Use more precise expressions
Eliminate repetitive word usage
Enrich vocabulary diversity

3. Structure Level

Optimize sentence length (when retain_original_structure=false)
Adjust paragraph organization
Improve logical connections
Enhance transitions between paragraphs

4. Fluency Level

Optimize rhythm and cadence
Eliminate awkward expressions
Add natural connectives
Improve the overall reading experience

5. Clarity level

Eliminate ambiguous expressions
Simplify complex sentence structures (based on language_level)
Clarify referential relationships
Optimize information organization

Quality assurance

Do not change the core meaning of the original text
Preserve key data and facts
Maintain matching with the target tone
Ensure the polished text is natural and fluent

Quick mode

When the user input contains a complete source_text and no other parameters are specified, execute directly with the default parameters without interaction.

Interaction question template (only displayed when the user does not provide the necessary parameters)

我将为您润色这段文本，请确认以下要求：

1. 润色重点（可多选，可选）：
   [1] 语法修正
   [2] 词汇优化
   [3] 结构调整
   [4] 流畅度提升
   [5] 表达清晰化
   [6] 语气一致性

2. 目标语气（可选）：
   [1] 正式
   [2] 半正式（推荐）
   [3] 非正式
   [4] 权威性
   [5] 说服性

3. 语言复杂度（可选）：
   [1] 简单易懂
   [2] 标准（推荐）
   [3] 高级专业

4. 是否保留原有结构（可选）：
   [1] 保留原有段落和句子结构
   [2] 可优化结构（推荐）

您可以：
- 一次性回复所有选择（如：1,2,4,2）
- 回复部分参数，我再确认是否需要补充
- 回复"生成"或"好的"，使用推荐设置执行（推荐：重点修正语法、词汇优化、流畅度提升、表达清晰化，半正式语气，标准语言复杂度，可优化结构）
- 如果你有其他想法，可在回复的时候一并描述告知

请回复：

Examples

Example 1: Standard polishing

Input：

source_text: "这个产品它有很多好的功能，能够帮助用户提高工作的效率。但是有时候它会有些问题，需要进一步改进。"
focus_areas: ["grammar", "vocabulary", "flow"]

Output：

这款产品拥有众多实用功能，可有效提升用户的工作效率。不过，目前仍存在一些需要改进的地方。

Example 2: Formal tone, advanced language

Input：

source_text: "我们做这个研究是因为想看看这个方法到底好不好用。"
target_tone: "formal"
language_level: "advanced"

Output：

本研究旨在评估该方法的实际效能与适用性。

Example 3: Polishing while preserving structure

Input：

source_text: "首先，我们需要准备材料。然后，按照说明进行操作。最后，检查结果是否正确。"
focus_areas: ["vocabulary", "clarity"]
retain_original_structure: true

Output：

首先，我们需要准备相关材料。接着，严格按照说明书进行操作。最后，验证结果是否正确。

Notes

Polishing automatically calls humanize_text, so there is no need to manually remove the AI flavor again
Preserving the original meaning is the highest priority; do not change facts for the sake of literary flair
If the user does not explicitly specify, perform light to moderate polishing by default
For professional content, it is recommended to maintain terminology consistency

Parameter adjustment tips

After generating the content, if you are not satisfied with the current result, you can try adjusting the following parameters:

💡 Parameter adjustment options:

Polishing focus: grammarGrammar correction | vocabularyVocabulary optimization | structureStructure adjustment | flowFluency improvement | clarityExpression clarification | tone_consistencyTone consistency
Target tone: formalFormal | semi-formalSemi-formal | casualInformal | authoritativeAuthoritative | persuasivePersuasive
Language complexity: simpleSimple and easy to understand | standardStandard | advancedAdvanced professional
Preserve the original structure: truePreserve the original paragraph and sentence structure | falseStructure can be optimized

How to use: Reply directly, for example, "Focus on adjusting vocabulary and fluency" or "Change to a formal tone"

Relationships with other skills

humanize_text: After polishing is completed, this skill is automatically called to remove AI traces
rewrite_content: If the goal is to change the way something is expressed rather than improve quality, rewrite_content should be used
expand_content: If content length needs to be increased, expand_content should be used

Configuration file references

intent_keywords: ["polish", "optimize", "improve", "enhance", "refine", "hone"]
polish_rules: The polishing rules library is defined in config/polish_rules.json

Feature 2: Humanize Text

Humanize Text - De-AI-ification processing

Introduction

Remove traces of AI generation from text to make the content more natural and humanized. This is the foundational skill of the Content Assistant system; after other skills complete execution, they automatically call this skill for final processing.

Use cases

AI-generated text needs to become more natural
The content is too formal or stiff
Need to add emotional and personalized expression
Content prepared for public release needs preprocessing before manual polishing

Input Parameters

Required Parameters

source_text (string): The original text that needs to be de-AI-ified

Optional Parameters

tone (string): Target tone style
- casual - Casual spoken language (recommended for social media, chat)
- professional - Professional and formal (recommended for work reports, business emails)
- friendly - Friendly and warm (recommended for customer communication, team interaction)
- neutral - Neutral and objective (recommended for documentation, technical materials)
- Default: automatically determined based on text content
audience (string): Target audience
- general - General public
- professional - Professionals
- academic - Academia
- youth - Young people
- business - Business professionals
- Default: general
retain_key_terms (boolean): Whether to retain key terms and specialized vocabulary
- true - Retain specialized terms and keywords from the original text
- false - Use plain expressions as much as possible
- Default: true
humanization_level (string): Degree of de-AI-ification
- light - Light adjustment, retain the original structure
- medium - Moderate adjustment, optimize the expression
- deep - Deep adjustment, completely restructure the expression
- Default: medium

Output

Return natural text after de-AI-ification, preserving the original meaning but making the expression more human.

Processing Rules

Text Analysis

Identify common AI expression patterns (such as "in conclusion", "in summary", "it is worth noting", etc.)
Detect overly regular sentence structures
Identify repeated connectives and transitional phrases
Analyze missing or unnatural emotion

Optimization Strategies

Diversify expression: Replace commonly used AI vocabulary with richer vocabulary
Vary sentence patterns: Break up regular sentence patterns and increase the alternation of long and short sentences
Inject emotion: Appropriately add emotional color and personalized expression
Colloquialize: Convert written language into more natural spoken expression
Adjust rhythm: Control information density and increase breathing room

Preservation Principles

Do not change the original meaning and core information
Retain specialized terms (when retain_key_terms=true)
Conform to the language habits of the target audience
Maintain logical coherence

Quick Mode

When the user input contains a complete source_text and no other parameters are specified, execute directly using the default parameters (tone=auto, audience=general, retain_key_terms=true, humanization_level=medium) without interaction.

Interactive question template (only displayed when the user does not provide the necessary parameters)

我将为您去除这段文本的AI味，请确认以下要求：

1. 语气风格（可选）：
   [1] 日常口语 - 社交媒体、聊天场景
   [2] 专业正式 - 工作报告、商务邮件
   [3] 友好亲切 - 客户沟通、团队交流
   [4] 中性客观 - 说明文档、技术材料

2. 目标受众（可选）：
   [1] 普通大众
   [2] 专业人士
   [3] 学术界
   [4] 年轻群体
   [5] 商务人士

3. 是否保留专业术语（可选）：
   [1] 保留专业术语和关键词
   [2] 使用通俗表达

4. 去AI化程度（可选）：
   [1] 轻度调整 - 保留原有结构
   [2] 中度调整 - 优化表达方式（推荐）
   [3] 深度调整 - 完全重构表达

您可以：
- 一次性回复所有选择（如：1,2,1,2）
- 回复部分参数，我再确认是否需要补充
- 回复"生成"或"好的"，使用推荐设置执行（推荐：根据内容自动判断语气，普通大众受众，保留专业术语，中度调整）
- 如果你有其他想法，可在回复的时候一并描述告知

请回复：

Examples

Example 1: Light de-AI-ification

Input：

source_text: "综上所述，人工智能技术的发展为我们的生活带来了诸多便利。值得注意的是，在享受便利的同时，我们也应该关注隐私保护问题。"
humanization_level: "light"

Output：

总的来说，AI技术的发展给生活带来了很多方便。不过享受这些便利时，我们也要注意保护隐私。

Example 2: Deep de-AI-ification, colloquial style

Input：

source_text: "此外，该方案的实施将显著提升工作效率，同时降低运营成本。不仅如此，它还能促进团队协作。"
tone: "casual"
humanization_level: "deep"

Output：

这方案一上，效率直接起飞，成本还能省不少。而且大家配合起来也更顺畅了。

Example 3: Maintain professionalism

Input：

source_text: "云计算架构采用微服务设计模式，能够实现服务的弹性伸缩和高可用性。"
tone: "professional"
audience: "professional"
retain_key_terms: true

Output：

我们的云架构用了微服务的设计，这样服务可以根据需要灵活扩展，而且稳定性也不错。

Notes

This skill is usually called automatically as a post-processing step for other skills
Deep de-AI-ification may change the formality of the text; please choose an appropriate level based on the use case
For technical documents or professional materials, retain_key_terms=true is recommended
De-AI-ification should not introduce incorrect information or misunderstandings

Parameter Adjustment Tips

After generating content, if you are not satisfied with the current result, you can try adjusting the following parameters:

💡 Parameter adjustment options:

Tone style: casualCasual spoken | professionalProfessional formal | friendlyFriendly and warm | neutralNeutral and objective
Target audience: generalGeneral public | professionalProfessionals | academicAcademia | youthYoung people | businessBusiness professionals
Retain technical terms: trueRetain technical terms and keywords | falseUse plain expressions
Degree of de-AI-ification: lightLight adjustment - retain the original structure | mediumModerate adjustment - optimize the expression | deepDeep adjustment - completely restructure the expression

How to use: Reply directly, such as "change to a professional formal tone" or "increase the degree of de-AI-ification"

Relationship with other skills

polish_text: Automatically call this skill after polishing is complete
rewrite_content: Automatically call this skill after rewriting is complete
expand_content: Automatically call this skill after expansion is complete
condense_text: Automatically call this skill after abbreviation is complete

Configuration file references

intent_keywords: ["remove AI", "AI flavor", "machine-like", "robot", "naturalize", "humanize"]
ai_patterns: Define common AI expression patterns in config/ai_patterns.json

Feature 3: Expand Content

Expand Content - Content Expansion

Introduction

On the basis of preserving the original meaning, expand and refine the text by adding details, cases, explanations, or background information to make the content richer and more complete. After expansion is complete, humanize_text is automatically called to remove AI traces.

Use cases

Brief points need to be expanded into complete content
Explanations and clarifications need to be added
A higher degree of content richness is required
Expand an outline or draft into a complete document
Add cases and supporting materials

Input parameters

Required parameters

source_text (string): The original text to be expanded

Optional parameters

expansion_direction (string[]): Expansion directions (multiple selections allowed)
- detail - Add detailed descriptions
- example - Add cases and examples
- explanation - Add explanations
- background - Add background information
- implication - Extend implications and significance
- perspective - Multi-angle analysis
- data - Add data and statistics
- Default value: ["detail", "explanation"]
target_length (string): Target length
- short - Short expansion (1.5-2x)
- medium - Medium expansion (2-3x, recommended)
- long - Long expansion (3-5x)
- free - Freely expand according to content needs (recommended)
- Default value: free
maintain_conciseness (boolean): Maintain conciseness
- true - After expansion, it remains concise and avoids redundancy
- false - Can be appropriately detailed
- Default value: true
audience_adjustment (string): audience adjustment
- same - Keep the original audience positioning
- general - Change to the general public
- expert - Change to professional readers
- Default value: same
creative_approach (string): creative approach
- linear - Linear expansion, developed according to the original text's logic
- branching - Branch expansion, adding related topics
- layered - Layered expansion, combining depth and accessibility
- Default value: linear

Output

The expanded rich content retains the original meaning but is more detailed and complete. Automatically call humanize_text to remove AI traces.

Processing rules

Expansion strategies

1. Add detailed descriptions (detail)

Concretize abstract concepts
Describe processes and steps
Refine scenes and imagery
Supplement time and space dimensions

2. Add case examples (example)

Use specific examples to illustrate
Cite real cases
Create hypothetical scenarios
Provide comparative cases

3. Add explanations (explanation)

Explain technical terms
Elaborate on causes and logic
Explain methods and principles
Clarify possible misunderstandings

4. Add background information (background)

Provide historical background
Introduce related concepts
Explain causes and consequences
Supplement the current state of the industry

5. Extend impact and significance (implication)

Analyze potential impact
Discuss practical significance
Look ahead to future development
Summarize value and role

6. Multi-perspective analysis (perspective)

Analyze from different standpoints
Consider both positive and negative sides
Provide diverse perspectives
Balance the views of all parties

7. Add data statistics (data)

Cite authoritative data
Provide quantitative analysis
Compare historical data
Predict trend directions

Quality control

Ensure that the expanded content is highly relevant to the original text
Avoid irrelevant or off-topic expansion
Maintain logical coherence and a clear structure
Must not change the original text's core viewpoints

Creative approach execution

linear: Expand paragraph by paragraph along the original text's logical order
branching: Add related branch topics on the basis of the original text
layered: Deeply explore the core points while maintaining readability

Quick mode

When the user input contains a complete source_text and no other parameters are specified, directly execute with the default parameters (add details and explanations, free expansion, keep concise, linear expansion).

Interactive question template (only displayed when the user does not provide the necessary parameters)

我将为您扩写这段内容，请确认以下要求：

1. 扩写方向（可多选，可选）：
   [1] 增加细节描述
   [2] 添加案例和举例
   [3] 添加解释说明
   [4] 增加背景信息
   [5] 延伸影响和意义
   [6] 多角度分析
   [7] 添加数据和统计

2. 目标长度（可选）：
   [1] 短幅扩写（1.5-2倍）
   [2] 中等扩写（2-3倍）
   [3] 长幅扩写（3-5倍）
   [4] 根据内容需要自由扩展（推荐）

3. 保持简洁性（可选）：
   [1] 扩写后仍保持简洁，避免冗余（推荐）
   [2] 可以适当详尽

4. 受众调整（可选）：
   [1] 保持原有受众定位
   [2] 改为普通大众
   [3] 改为专业读者

5. 创意方式（可选）：
   [1] 线性扩展，按原文逻辑展开（推荐）
   [2] 分支扩展，增加相关话题
   [3] 分层扩展，深入和浅出结合

您可以：
- 一次性回复所有选择（如：1,2,1,1,1）
- 回复部分参数，我再确认是否需要补充
- 回复"生成"或"好的"，使用推荐设置执行（推荐：增加细节描述和解释说明，自由扩展，保持简洁，线性扩展）
- 如果你有其他想法，可在回复的时候一并描述告知

请回复：

Example

Example 1: Add details and explanations

Input：

source_text: "健康饮食很重要。"
expansion_direction: ["detail", "explanation"]

Output：

健康饮食对维持良好的身体状态至关重要。均衡的营养摄入不仅能够提供日常所需的能量，还能增强免疫系统，降低患病风险。具体而言，多摄入蔬菜水果、优质蛋白质和全谷物，同时控制糖分和饱和脂肪的摄入，是建立健康饮食习惯的基础。

Example 2: Add cases and examples

Input：

source_text: "人工智能正在改变各行各业。"
expansion_direction: ["example", "background"]

Output：

人工智能技术正以前所未有的速度渗透到各个行业，带来革命性的变化。在医疗领域，AI辅助诊断系统能够帮助医生快速识别疾病；在制造业，智能机器人实现了生产线的自动化；在金融行业，算法交易和风险评估系统提升了决策效率。自20世纪50年代AI概念提出以来，经过几十年的技术积累和算力提升，如今AI已经从实验室走向实际应用，成为推动社会进步的重要力量。

Example 3: Multi-perspective analysis

Input：

source_text: "远程办公的普及带来了新的挑战。"
expansion_direction: ["perspective", "implication"]
creative_approach: "layered"

Output：

远程办公的普及虽然带来了灵活性，但也随之产生了一系列需要面对的新挑战。从企业管理的角度来看，如何有效监督和评估员工绩效成为难题；从员工的角度看，工作与生活的边界变得模糊，容易导致身心疲劳；从社会层面来看，传统的办公室文化正在重塑，人们需要适应新的协作模式。

这些挑战的背后，折射出工作方式的深刻变革。一方面，企业需要建立新的管理机制和数字化工具来支持远程协作；另一方面，个人需要培养自律和时间管理能力。长远来看，这种变化可能重新定义工作形态，推动更灵活、多元的就业模式发展。

Example 4: Add data statistics

Input：

source_text: "电子商务发展迅速。"
expansion_direction: ["data", "explanation"]

Output：

电子商务近年来呈现出强劲的增长势头，正深刻改变着消费者的购物习惯和商业格局。据统计数据显示，2023年中国网络零售额已突破13万亿元，同比增长约8%，占社会消费品零售总额的比重持续上升。这一增长背后，是智能手机的普及、物流网络的完善以及移动支付技术的成熟，共同为电商发展提供了坚实基础。

Notes

Expansion will automatically call humanize_text, no need to manually remove the AI flavor
Avoid adding irrelevant content just to pad the word count
The expanded content should stay closely tied to the original theme and core viewpoints
When creative_approach=branching, be careful to control relevance

Parameter adjustment tips

After generating content, if you are not satisfied with the current result, you can try adjusting the following parameters:

💡 Parameter adjustment options:

Expansion direction: detailAdd detailed descriptions | exampleAdd cases and examples | explanationAdd explanations | backgroundAdd background information | implicationExtend impacts and significance | perspectiveMulti-perspective analysis | dataAdd data and statistics
Target length: shortShort expansion (1.5-2x) | mediumMedium expansion (2-3x) | longLong expansion (3-5x) | freeExpand freely as needed by the content
Maintain conciseness: trueRemain concise after expansion | falseCan be appropriately detailed
Creative approach: linearLinear expansion | branchingBranching expansion | layeredHierarchical expansion

How to use: Reply directly such as "Add cases and examples" or "Expand to medium length"

Relationship with other skills

humanize_text: Automatically call this skill after expansion is complete to remove AI traces
condense_text: Relative skill, used to condense content
rewrite_content: If you also need to change the style, you can use them in combination

Configuration file references

intent_keywords: ["expand", "extend", "elaborate", "refine", "enrich", "supplement"]
expansion_templates: Define the expansion template library in templates/expansion_templates.json

Feature 4: Condense Text

Condense Text - Content condensation

Overview

While preserving core information and key viewpoints, condense and summarize the text, remove redundant content, and make the expression more concise and clear. After condensation is complete, humanize_text will be automatically called to remove AI traces.

Use cases

Long content needs to be condensed
Extract key points and key information
Create abstracts or summaries
Quickly understand the main idea of a document
Adapt to word count limit requirements

Input parameters

Required parameters

source_text (string): The original text to be condensed

Optional parameters

condense_level (string): Degree of condensation
- light - Light condensation, retain most details (about 70-80%)
- medium - Medium condensation, retain core points (about 40-60%, recommended)
- heavy - Deep condensation, retain only core information (about 20-30%)
- summary - Summary mode, generate a complete abstract (about 10-15%)
- Default value: medium
output_format (string): Output format
- paragraph - Paragraph form
- bullet_points - Bullet list form
- numbered - Numbered list form
- Default value: paragraph
preserve_elements (string[]): Elements to preserve (multiple selections allowed)
- data - Preserve all data and statistics
- quotes - Preserve important quotes
- examples - Preserve key cases
- names - Preserve important person names and organization names
- dates - Preserve important times
- none - Do not preserve specific elements
- Default value: ["data"]
target_length (number): Target word count (optional)
- Specify a specific word count limit
- Default value: unlimited, automatically determined based on condense_level
perspective (string): Summarization angle
- objective - Objective summarization
- subjective - Summarization with evaluation
- Default value: objective

Output

The concise abbreviated content, preserving core information and key viewpoints. Automatically calls humanize_text to remove AI traces.

Processing Rules

Abbreviation Strategy

1. Information hierarchy judgment

Core information: main viewpoints, conclusions, key data
Important information: supporting arguments, important details
Secondary information: background introduction, supplementary explanations
Redundant information: repeated content, irrelevant descriptions

2. Abbreviation level control

Light (light abbreviation)

Remove obvious redundancy and repetition
Simplify descriptive text
Preserve most supporting details
About 70-80% of the content

Medium (medium abbreviation, recommended)

Preserve main arguments and key data
Streamline supporting evidence
Merge similar viewpoints
Remove background and supplementary explanations
About 40-60% of the content

Heavy (deep abbreviation)

Preserve only core conclusions and key data
Remove all supporting details
Minimal expression
About 20-30% of the content

Summary (summary mode)

Generate a complete summary text
Include main content and conclusions
Logically complete and independent
About 10-15% of the content

3. Content preservation rules

Preserve core viewpoints: do not change the main stance and conclusions of the original text
Preserve key data: determined by the preserve_elements parameter
Preserve logical relationships: relationships such as cause and effect and comparison remain clear
Delete repeated content: merge or delete repeated expressions
Delete redundant descriptions: simplify overly detailed descriptions
Delete irrelevant information: remove content unrelated to the topic

4. Output format processing

paragraph: coherent paragraphs, natural and smooth
bullet_points: clear itemized list, easy to browse quickly
numbered: ordered list, suitable for step-by-step content

Quality assurance

Ensure the abbreviated content is logically complete
Avoid difficulty in understanding caused by missing information
Keep the expression concise but not crude
The core meaning must be fully preserved

Quick mode

When the user input contains the complete source_text and no other parameters are specified, execute directly with the default parameters (moderate abbreviation, paragraph form, preserve data, objective summary).

Interactive question template (only displayed when the user does not provide the necessary parameters)

我将为您缩写这段内容，请确认以下要求：

1. 缩写程度（可选）：
   [1] 轻度缩写，保留大部分细节（约70-80%）
   [2] 中度缩写，保留核心要点（约40-60%，推荐）
   [3] 深度缩写，仅保留核心信息（约20-30%）
   [4] 总结模式，生成完整摘要（约10-15%）

2. 输出格式（可选）：
   [1] 段落形式
   [2] 要点列表形式
   [3] 编号列表形式

3. 保留元素（可多选，可选）：
   [1] 保留所有数据和统计
   [2] 保留重要引用
   [3] 保留关键案例
   [4] 保留重要人名和组织名
   [5] 保留重要时间
   [6] 不保留特定元素

4. 概括角度（可选）：
   [1] 客观概括
   [2] 带评价的概括

您可以：
- 一次性回复所有选择（如：2,1,1,1）
- 回复部分参数，我再确认是否需要补充
- 回复"生成"或"好的"，使用推荐设置执行（推荐：中度缩写，段落形式，保留数据，客观概括）
- 如果你有其他想法，可在回复的时候一并描述告知

请回复：

Example

Example 1: Moderate abbreviation

Input：

source_text: "近年来，随着科技的不断进步，人工智能技术在各个领域都取得了显著的发展。特别是在医疗健康领域，AI技术的应用为疾病的早期诊断、精准治疗提供了新的解决方案。例如，通过深度学习算法分析医学影像，可以帮助医生发现早期的病变，大大提高了诊断的准确性和效率。此外，AI还能根据患者的病史和基因信息，制定个性化的治疗方案。然而，AI技术的发展也带来了一些挑战，比如数据隐私保护、算法偏见等问题，需要我们共同面对和解决。"
condense_level: "medium"

Output：

近年来，AI技术在医疗健康领域发展迅速。通过深度学习算法分析医学影像，AI能够辅助早期疾病诊断，提高准确性和效率。同时，基于患者病史和基因信息的个性化治疗方案也逐渐成为可能。不过，AI技术发展也面临数据隐私和算法偏见等挑战，需要持续关注和解决。

Example 2: Deep abbreviation, bulleted list

Input：

source_text: "气候变化已经成为全球关注的重大问题。全球变暖导致极端天气事件频发，如暴雨、干旱、飓风等，给人类社会和生态环境带来了严重影响。科学家呼吁各国政府采取行动，减少温室气体排放，推动可再生能源的发展。同时，普通民众也应该从日常生活做起，节约能源，减少碳足迹。只有全球合作，才能有效应对这一挑战，为子孙后代留下一个可持续发展的地球。"
condense_level: "heavy"
output_format: "bullet_points"

Output：

• 气候变化导致极端天气频发，影响严重
• 需要减少温室气体排放，发展可再生能源
• 个人应节约能源，减少碳足迹
• 需要全球合作应对气候变化挑战

Example 3: Summary mode

Input：

source_text: "某公司最近推出了一款全新的智能手机产品。这款手机配备了最新的处理器，性能比上一代提升了30%。同时，它的电池容量增加了，续航时间达到了两天。在摄像头方面，采用了三摄系统，支持夜景模式和专业级拍摄。操作系统也进行了全面升级，带来了更流畅的用户体验。市场反馈显示，消费者对这款产品反响热烈，上市首月销量突破了100万台。"
condense_level: "summary"

Output：

某公司发布新款智能手机，性能提升30%，续航达两天，配备三摄系统和升级的操作系统。市场反响热烈，首月销量超100万台。

Example 4: Preserve specific elements

Input：

source_text: "根据最新发布的《2023年中国电子商务发展报告》，全国网络零售额达到了13万亿元，同比增长8%。其中，实物商品网上零售额为10.8万亿元，占社会消费品零售总额的25%。报告指出，直播电商成为新的增长引擎，交易规模突破3万亿元。专家预计，2024年电商行业将继续保持稳定增长，年增长率预计在7-9%之间。"
condense_level: "medium"
preserve_elements: ["data"]

Output：

2023年全国网络零售额达13万亿元，同比增长8%。其中实物商品网上零售额10.8万亿元，占社会消费品零售总额25%。直播电商交易规模突破3万亿元，成为新增长引擎。预计2024年行业增速将保持在7-9%之间。

Notes

Abbreviation will automatically call humanize_text; there is no need to manually remove the AI flavor again
Abbreviation should not change the core meaning of the original text
Deep abbreviation may lead to difficulty in understanding; it is recommended to provide context
For important data, it is recommended to use the preserve_elements parameter to explicitly preserve it

Parameter adjustment tips

After generating content, if you are not satisfied with the current result, you can try adjusting the following parameters:

💡 Parameter adjustment options:

Abbreviation level: lightLight abbreviation - preserve most details (70-80%) | mediumModerate abbreviation - preserve the core points (40-60%) | heavyDeep abbreviation - preserve only the core information (20-30%) | summarySummary mode - generate a complete summary (10-15%)
Output format: paragraphParagraph form | bullet_pointsBulleted list form | numberedNumbered list form
Elements to preserve: dataPreserve all data and statistics | quotesPreserve important quotations | examplesPreserve key cases | namesPreserve important names of people and organizations | datesPreserve important dates
Summarization perspective: objectiveObjective summary | subjectiveSummary with evaluation

How to use: Reply directly such as "abbreviate into a bulleted list" or "change to moderate abbreviation"

Relationship with other skills

humanize_text: After abbreviation is completed, automatically call this skill to remove AI traces
expand_content: Relative skill, used to expand content
polish_text: If quality needs to be improved after abbreviation, they can be used in combination

Configuration file reference

intent_keywords: ["abbreviation", "summary", "condense", "summarize", "distill", "simplify"]
condense_rules: The abbreviation rule library is defined in config/condense_rules.json

About this skill

Introduction

Feature 1: Polish Text

Introduction

Use Cases

Input Parameters

Required Parameters

Optional Parameters

Output

Processing Rules

Analysis Phase

Polishing Strategies

1. Grammar Level

2. Vocabulary Level

3. Structure Level

4. Fluency Level

5. Clarity level

Quality assurance

Quick mode

Interaction question template (only displayed when the user does not provide the necessary parameters)

Examples

Example 1: Standard polishing

Example 2: Formal tone, advanced language

Example 3: Polishing while preserving structure

Notes

Parameter adjustment tips

Relationships with other skills

Configuration file references

Feature 2: Humanize Text

Introduction

Use cases

Input Parameters

Required Parameters

Optional Parameters

Output

Processing Rules

Text Analysis

Optimization Strategies

Preservation Principles

Quick Mode

Interactive question template (only displayed when the user does not provide the necessary parameters)

Examples

Example 1: Light de-AI-ification

Example 2: Deep de-AI-ification, colloquial style

Example 3: Maintain professionalism

Notes

Parameter Adjustment Tips

Relationship with other skills

Configuration file references

Feature 3: Expand Content

Introduction

Use cases

Input parameters

Required parameters

Optional parameters

Output

Processing rules

Expansion strategies

1. Add detailed descriptions (detail)

2. Add case examples (example)

3. Add explanations (explanation)

4. Add background information (background)

5. Extend impact and significance (implication)

6. Multi-perspective analysis (perspective)

7. Add data statistics (data)

Quality control

Creative approach execution

Quick mode

Interactive question template (only displayed when the user does not provide the necessary parameters)

Example

Example 1: Add details and explanations

Example 2: Add cases and examples

Example 3: Multi-perspective analysis

Example 4: Add data statistics

Notes

Parameter adjustment tips

Relationship with other skills

Configuration file references

Feature 4: Condense Text