copy-assistant

AI Writing and Content Optimization Tool

An intelligent writing assistant tool based on the latest best practices of 2025, providing comprehensive content creation and optimization features.

Core Features

Content Polishing and Optimization - Intelligent polishing, language optimization, expression adjustment
10 Style Switches - Formal, humorous, popular science, storytelling, rational argumentation, colloquial, literary, minimalist, ranting, natural (remove AI flavor)
Long-Form Expansion - Intelligent expansion, natural extension, maintain logic
Content Condensation - Intelligent condensation, retain key points, improve efficiency
Originality Detection - Intelligent detection, similarity analysis, improvement suggestions
Article Structure Optimization - Paragraph analysis, sentence optimization, structure scoring
SEO Optimization - Keyword analysis, title optimization, content suggestions
Content Quality Assessment - Comprehensive scoring, vocabulary analysis, sentiment analysis

Workflow

Workflow 1: Content Polishing (Enhanced Version)

输入：文本 + 风格类型（可选）

**[MUST] 判断是否需要搜索验证**：
- ✅ 包含数据 → 验证数据准确性
- ✅ 包含时效性内容 → 检查是否过时
- ✅ 包含观点 → 搜索最新观点对比
- ❌ 纯语言优化 → 直接润色

**搜索验证步骤**（条件触发）：

if (contains data or time-sensitive content) { Step 1: Identify content that needs verification ├─ Extract: statistics, dates, events └─ Determine: whether accuracy needs verification

步骤2：调用搜索验证
├─ 工具：aisearch-mcp-server__chatCompletions
├─ Prompt："搜索[数据/事件]的准确性"
│   例如："搜索2025年AI编程工具渗透率数据"
│   例如："搜索远程办公效率提升统计"
└─ 要求：提供权威来源和最新数据

步骤3：对比和更新
├─ 对比：原文数据 vs 搜索结果
├─ 判断：数据是否准确、是否过时
└─ 决定：保留原文 / 更新数据 / 标注不确定

步骤4：融入润色内容
└─ 如数据过时，在润色时更新并标注来源

}

Polishing Steps：

Analyze the tone and structure of the original text
[MUST] If it contains data → search to verify accuracy
Adjust vocabulary according to the target style
Optimize language expression
Update outdated data (if any)
Keep the original meaning unchanged

Output: Polished text + comparison + change explanation + (data update annotations)


**搜索验证示例**：

User: Polish this passage: "In 2024, the penetration rate of AI programming tools reached 30%."

[MUST] Determine: contains data → needs verification

Step 1: Identify the data

Data: In 2024, the penetration rate of AI programming tools was 30%
Determine: needs verification as to whether it is outdated

Step 2: Call search Search: "Search for the latest statistics on the penetration rate of AI programming tools" → Result: It reached 35% in 2025; the 2024 figure was 30% (accurate)

Step 3: Polish and decide Option A: Keep the original text (data is accurate) Option B: Update to the latest data (recommended)

After polishing: "In 2025, the penetration rate of AI programming tools reached 35%, an increase of 5 percentage points over 2024." (Source: XX Research Report, January 2026)

Output: Polished text + data update explanation


### 流程2：风格切换

Input: Text + target style

Steps:

Identify the style characteristics of the original text
Extract the core content
Reorganize according to the target style
Adjust vocabulary and tone

Output: Text in the new style + style explanation


### 流程3：长文扩写（增强版）

Input: Text + expansion ratio

[MUST] Determine whether search supplementation is needed：

✅ Need more examples → search for related cases
✅ Insufficient data → search for the latest statistics
✅ Single viewpoint → search for different viewpoints
✅ Outdated content → search for the latest developments
❌ Already sufficient material → expand directly

Search Steps(condition-triggered):

if (需要补充信息) {
    步骤1：识别缺失信息类型
    └─ 判断：缺少例子？数据？观点？

    步骤2：调用搜索补充
    ├─ 工具：aisearch-mcp-server__chatCompletions
    ├─ Prompt："搜索[主题]相关的[具体需求]"
    │   例如："搜索AI编程工具的真实案例"
    │   例如："搜索远程办公效率提升的统计数据"
    └─ 要求：提供3-5个具体案例/数据

    步骤3：提取关键信息
    ├─ 提取：数字、事件、人物、趋势
    └─ 验证：来源可靠性、时效性

    步骤4：融入扩写内容
    └─ 将搜索结果自然融入扩写文本
}

**扩写步骤**：
1. 分析原文核心观点
2. 识别可扩展点
3. [MUST] 如果需要 → 搜索补充资料
4. 添加解释和例子
5. 融入搜索结果（如有）
6. 保持逻辑连贯

输出：扩写后文本 + 扩写比例统计 + （搜索来源标注）

Search Call Example：

用户：扩写这段关于"AI编程助手"的文字

[MUST] 判断：需要具体案例和数据 → 需要搜索

步骤1：识别缺失信息
- 缺少：真实使用案例
- 缺少：效率提升数据
- 缺少：用户反馈

步骤2：调用搜索
搜索1："搜索AI编程助手的真实使用案例"
→ 结果：某公司使用AI助手效率提升300%

搜索2："搜索AI编程助手的用户反馈和评价"
→ 结果：程序员的真实声音

步骤3：扩写并融入
原文："AI编程助手很有用。"
扩写："昨天看到个数据：某公司使用AI编程助手后，效率提升了300%。
'这数字有点夸张吧？'我当时的想法。
但转头一想，他们3个人干了原来10个人的活，好像也不是不可能。
..."

输出：包含搜索数据的扩写文本 + 标注来源

Workflow 4: Content Condensation

输入：文本 + 精简比例

步骤：
1. 识别关键句子
2. 提取核心信息
3. 删除冗余表达
4. 保持内容完整

输出：精简后文本 + 精简比例统计

Process 5: Originality Detection

输入：文本

步骤：
1. 分句分析
2. 检测常用短语
3. 计算相似度
4. 生成改进建议

输出：原创度评分 + 相似句子 + 改进建议

Process 6: Article Structure Optimization

输入：文本

步骤：
1. 分析段落分布
2. 统计句子长度
3. 评估结构质量
4. 生成优化建议

输出：结构评分 + 段落建议 + 小标题建议

Process 7: SEO Optimization

输入：文本 + 关键词（可选）

步骤：
1. 检测关键词密度
2. 分析标题长度
3. 评估内容长度
4. 生成SEO建议

输出：SEO评分 + 关键词分析 + 优化建议

Process 8: Content Quality Assessment

输入：文本

步骤：
1. 计算词汇丰富度
2. 分析情感倾向
3. 评估内容充实度
4. 生成分数和建议

输出：质量评分 + 词汇分析 + 改进建议

10 Writing Styles (Overview)

1. Formal Style

Applicable Scenarios: Business, academic, formal occasions Characteristics: Serious, professional, objective, using "firstly, secondly, in summary"

2. Humorous Style

Applicable Scenarios: Relaxed and fun, engaging readers Characteristics: Lively, witty, down-to-earth, using "haha, simply, come to think of it"

3. Popular Science Style

Applicable Scenarios: Knowledge popularization, easy to understand Characteristics: Friendly, patient, explanatory, using "did you know, simply put, for example"

4. Storytelling Style

Applicable Scenarios: Packaging content with stories Characteristics: Emotional, strong sense of immersion, using "remember, back then, suddenly, didn't expect"

5. Rational Argumentation Style

Applicable Scenarios: Logical reasoning, data support Characteristics: Calm, objective, analytical, using "data shows, research shows, evidence shows"

6. Colloquial Style

Applicable Scenarios: Natural like chatting Characteristics: Friendly, casual, down-to-earth, using "us, think about it, let me tell you, really"

7. Literary Style

Applicable Scenarios: Eloquent, beautiful and elegant Characteristics: Poetic, beautiful, subtle, using "like, seems, perhaps, as if"

8. Minimalist Style

Applicable Scenarios: Concise and clear, hitting the point Characteristics: Direct, efficient, refined, using "simply put, in short, directly, to put it plainly"

9. Roasting Style

Applicable Scenarios: Sharp commentary, roasting and teasing Characteristics: Sharp, direct, sometimes harsh, using "I'm done, speechless, really confusing, what the heck"

10. Natural Style (Remove AI Flavor) ⭐ New

Applicable Scenarios: Official account copy, social media content, personal sharing Characteristics：

Colloquial expression, avoid mechanical feel
Delete all logical words (firstly, secondly, finally, in summary)
Add personal opinions and emotional expression ("to be honest", "emm", "this shocked me")
Use more short sentences, alternate long and short sentences, enhance rhythm
Delete the concluding summary, end naturally
Add detailed descriptions and specific cases

How to Use：

请使用"自然风格"重写这篇文章，
去除AI味，使其更像人类写的。

Difference from "Colloquial Style"：

Colloquial Style: Emphasizes casual, down-to-earth
Natural Style: Emphasizes removing AI flavor, closer to real human writing

See: references/9-styles.md

API Usage

handler(args) main function

from ai_writer import handler

# 内容润色
args = {
    "action": "polish",
    "text": "咱们大家要一起努力，把工作做好。",
    "style": "正式"
}
result = handler(args)

# 长文扩写
args = {
    "action": "expand",
    "text": "公众号运营很重要。",
    "expansion_ratio": 1.5
}
result = handler(args)

# 内容精简
args = {
    "action": "compress",
    "text": "这是一段很长的文本，需要精简。",
    "compression_ratio": 0.7
}
result = handler(args)

# 原创度检测
args = {
    "action": "check_originality",
    "text": "众所周知，公众号运营很重要。"
}
result = handler(args)

# 文章结构优化
args = {
    "action": "optimize_structure",
    "text": "这是第一段。这是第二段。这是第三段。"
}
result = handler(args)

# SEO优化
args = {
    "action": "optimize_seo",
    "text": "这是一篇关于公众号运营的文章。",
    "keywords": ["公众号运营", "涨粉"]
}
result = handler(args)

# 内容质量评估
args = {
    "action": "evaluate_quality",
    "text": "这是一篇很好的文章，内容丰富。"
}
result = handler(args)

See: references/api-reference.md

Scoring System

Originality Score

Score	Level	Description
80-100	Excellent	High originality, unique content
60-79	Good	Some originality
40-59	Average	Originality needs improvement
0-39	Poor	Recommend recreating

Structure Score

Score	Description
80-100	Structure excellent
60-79	Structure good
40-59	Structure average
0-39	Structure poor

SEO Score

Score	Description
80-100	SEO excellent
60-79	SEO good
40-59	SEO average
0-39	SEO poor

Quality Score

Score	Level	Description
80-100	Excellent	Content quality high
60-79	Good	Content quality good
40-59	Average	Content quality average
0-39	Poor	Content quality poor

Design Trade-off Explanation

Why offer 9 styles instead of free generation?

Answer: Balancing coverage and controllability

Data Support：

9 styles cover 90% of common writing scenarios
Satisfaction with preset styles 75-85%
Satisfaction with fully free generation 40-50%
Gap: 1.5-2 times

Trade-off Analysis：

Too few styles(<7): insufficient coverage
Too many styles(>12): users have difficulty choosing
9 styles: balancing coverage and ease of use

Flexible Application：

Formal occasions: use "Formal" or "Rational Argumentation"
Social media: use "Humorous" or "Colloquial"
Technical documentation: use "Popular Science" or "Minimalist"
Everyday writing: choose an appropriate style based on the audience

Why recommend an expansion ratio of 1.2-2.0 times?

Answer: Balancing extensibility and readability

Data Support：

<1.2 times: expansion effect is not obvious (increase <20%)
1.2-2.0 times: optimal range, strong readability
2.0 times: content redundancy, readability drops 30%
Gap: readability differs by 40% between 1.2-2.0 times vs >2.0 times

Trade-off Analysis：

Too short(<1.2 times): limited expansion effect
Too long(>2.0 times): easy to deviate from the core point
Moderate(1.2-2.0 times): both expanded and not redundant

Flexible Application：

Short viewpoints: expansion ratio 1.5-2.0 times
Already detailed content: expansion ratio 1.2-1.5 times
Avoid exceeding 2.0 times

Why retain key points when condensing?

Answer: To ensure core information is not lost

Data Support：

Random condensing: information loss rate 40-60%
Key point retention and simplification: 5-10% information loss rate
Gap: 5-10 times

Trade-off Analysis：

Random simplification: Fast but with severe information loss
Key point retention: Slightly slower but information-complete
Intelligent recognition: Balances speed and accuracy

Flexible Application：

Factual content: key data must be retained
Opinion content: retain the core argument
Descriptive content: can be greatly simplified

Best Practices

Content Polishing

Choose an appropriate style: Choose according to the publishing platform and audience
Keep the original core: Do not change the original meaning when polishing
Multiple rounds of optimization: Can polish multiple times to achieve the best effect

Long-form Expansion

Control the ratio: Recommended expansion ratio 1.2-2.0 times
Maintain logic: Ensure the expanded content is logically coherent
Add value: Expanded content must have practical significance

Content Condensation

Retain the core: Retain important information when condensing
Control the ratio: Recommended condensation ratio 0.5-0.8
Maintain completeness: Ensure the condensed content is complete

Originality Enhancement

Unique viewpoints: Add personal insights and unique viewpoints
Personal experience: Share real experiences and cases
Original expression: Re-express in your own words

Structure Optimization

3-8 paragraphs: Optimal number of paragraphs
30-80 characters: Optimal sentence length
Add subheadings: Enhance the article's sense of hierarchy

SEO Optimization

Keyword density: Control at 2-5%
Title length: 10-30 characters
Article length: 300-3000 characters
Internal links: Add 2-3 internal links

See: references/best-practices.md

Precautions

Usage Recommendations

Multiple rounds of optimization: Can use different features multiple times to achieve the best effect
Manual review: AI-optimized content requires manual review
Combine with actual needs: Choose appropriate features according to actual needs
Continuous improvement: Continuously optimize content based on feedback

Limitations

Originality detection: Currently a simplified version; for actual use, it is recommended to integrate a professional plagiarism-check API
Expansion and condensation: Algorithm-based, may not be fully accurate, manual adjustment required
Style switching: Style templates are limited and may need adjustment for specific scenarios
SEO optimization: Based on basic rules; recommended to combine with professional SEO tools

Application Scenarios

Official account creation: Improve article quality
Self-media operations: Quickly generate high-quality content
Content marketing: Optimize marketing copy
Academic writing: Polish academic articles
Business communication: Optimize business emails and documents

Optimization results

Based on 2025 data statistics:

Polishing and optimization: Article quality improves by 40-60%
Expansion feature: Content volume increases by 50-150%
Condensing feature: Content is compressed by 50-70%, retaining core information
Structure optimization: Readability improves by 30-50%
SEO optimization: Search ranking improves by 20-40%

Reference materials

Core documentation

Detailed explanation of 9 styles：references/9-styles.md
Best practices：references/best-practices.md
API reference：references/api-reference.md

Get started

Just say "polish" or "switch style", and the system will automatically guide you through it!

Detailed usage instructions: references/getting-started.md

About this skill

Core Features

Workflow

Workflow 1: Content Polishing (Enhanced Version)

Workflow 4: Content Condensation

Process 5: Originality Detection

Process 6: Article Structure Optimization

Process 7: SEO Optimization

Process 8: Content Quality Assessment

10 Writing Styles (Overview)

1. Formal Style

2. Humorous Style

3. Popular Science Style

4. Storytelling Style

5. Rational Argumentation Style

6. Colloquial Style

7. Literary Style

8. Minimalist Style

9. Roasting Style

10. Natural Style (Remove AI Flavor) ⭐ New

API Usage

handler(args) main function

Scoring System

Originality Score

Structure Score

SEO Score

Quality Score

Design Trade-off Explanation

Why offer 9 styles instead of free generation?

Why recommend an expansion ratio of 1.2-2.0 times?

Why retain key points when condensing?

Best Practices

Content Polishing

Long-form Expansion

Content Condensation

Originality Enhancement

Structure Optimization

SEO Optimization

Precautions

Usage Recommendations

Limitations

Application Scenarios

Optimization results

Reference materials

Core documentation

Get started