prose-distill

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

/prose:distill

Distill verbose text to its concentrated essence — the art of compression without loss.

将冗长文本提炼为浓缩的核心内容——这是一种无损压缩的艺术。

When to Use This Skill

何时使用此技能

Use this skill when...	Use something else when...
Text is wordy and needs tightening	You need to change tone or register (use prose-tone)
Redundant phrases need removing	You need to restructure document flow (use prose-structure)
User says "condense", "tighten", "shorten"	You need to adapt for a specific audience (use prose-audience)
Preserving all meaning while reducing length	Summarizing (lossy) rather than distilling (lossless)

适用场景	其他场景（请使用其他技能）
文本啰嗦，需要精简	需要调整语气或语体（使用prose-tone）
需要移除冗余表述	需要重构文档逻辑（使用prose-structure）
用户提到“浓缩”“精简”“缩短”	需要针对特定受众调整内容（使用prose-audience）
保留全部原意的同时缩短篇幅	进行摘要（有信息损失）而非提炼（无损）

Core Principles

核心原则

Distillation is lossless compression of natural language. Every sentence in the output must preserve the information content of the input. The goal is approaching the Shannon limit of the message — boiling off redundancy to leave concentrated meaning.

文本提炼是对自然语言的无损压缩。输出的每一句话都必须保留输入文本的全部信息内容。目标是接近信息的香农极限——剔除冗余，留下浓缩的核心含义。

The Hierarchy of Cuts

精简优先级层级

Apply in this order. Each level removes less essential material:

Redundant phrases — saying the same thing twice in different words
Filler words — "actually", "basically", "essentially", "really", "very", "quite", "rather"
Hedge words — "somewhat", "arguably", "it could be said that", "in a sense"
Throat-clearing — opening phrases that delay the point ("It is worth noting that", "It should be mentioned that")
Nominalizations — noun forms where verbs are stronger ("make a decision" → "decide", "perform an analysis" → "analyze")
Passive constructions — where active is clearer and shorter
Prepositional chains — "the result of the analysis of the data" → "the data analysis result"
Weak verbs + adverbs — replace with a single precise verb ("moved quickly" → "darted")

请按以下顺序进行精简，每一层级移除的都是相对非必要的内容：

冗余表述——用不同词语重复表达同一内容
填充词——"actually","basically","essentially","really","very","quite","rather"
模糊修饰词——"somewhat","arguably","it could be said that","in a sense"
铺垫语——延迟核心观点的开头表述（如"It is worth noting that","It should be mentioned that"）
名词化结构——将名词形式转换为更有力的动词（如"make a decision"→"decide"，"perform an analysis"→"analyze"）
被动语态——在更清晰简洁的情况下转换为主动语态
介词链——"the result of the analysis of the data"→"the data analysis result"
弱动词+副词——替换为单个精准动词（如"moved quickly"→"darted"）

What to Preserve

需要保留的内容

Technical precision and domain terminology
Necessary qualifications and nuance
Logical structure and argument flow
Voice and character (distill the style, don't flatten it)
Specific details, numbers, names, references

技术准确性和领域术语
必要的限定条件和细微差别
逻辑结构和论证流程
语气和风格（精简时保留风格，不要扁平化）
具体细节、数字、名称、参考文献

Parameters

参数解析

Parse

$ARGUMENTS

If text is provided inline, distill it directly
If a file path is provided, read and distill the file contents
If no arguments, ask the user for text to distill

解析

$ARGUMENTS

如果文本是内联提供的，直接对其进行提炼
如果提供了文件路径，读取并提炼文件内容
如果没有参数，请向用户索要需要提炼的文本

Execution

执行流程

Execute this distillation workflow:

执行以下文本提炼工作流：

Step 1: Assess the input

步骤1：评估输入内容

Read the provided text. Identify:

Approximate word count
Density of redundancy (light, moderate, heavy)
Whether the text has a distinctive voice worth preserving

阅读提供的文本，确定：

大致字数
冗余程度（轻度、中度、重度）
文本是否有值得保留的独特风格

Step 2: Apply the hierarchy of cuts

步骤2：应用精简优先级层级

Work through the text applying cuts in order from the hierarchy above. For each sentence:

Can two sentences merge into one without losing meaning?
Are there redundant phrases?
Can filler/hedge words be removed?
Can nominalizations become verbs?
Can passive become active without changing emphasis?
Can prepositional chains compress?

按照上述优先级层级逐句处理文本：

能否将两句话合并为一句且不丢失原意？
是否存在冗余表述？
能否移除填充词/模糊修饰词？
能否将名词化结构转换为动词？
能否在不改变重点的情况下将被动语态转为主动语态？
能否压缩介词链？

Step 3: Verify lossless compression

步骤3：验证无损压缩

Compare the distilled version against the original. Confirm:

No information was lost
No meaning was altered
Qualifications and nuance survived
The logical flow is intact

将提炼后的版本与原文对比，确认：

没有丢失任何信息
没有改变原意
限定条件和细微差别得以保留
逻辑流程完整

Step 4: Present the result

步骤4：呈现结果

Output the distilled text. Follow with a brief summary:

---
Original: ~N words
Distilled: ~N words
Reduction: ~N%

If any meaning was ambiguous and required interpretation, note it.

输出提炼后的文本，随后附上简要总结：

---
原文：~N词
提炼后：~N词
精简比例：~N%

如果存在任何模糊需要解读的内容，请予以说明。

Examples

示例

Filler and hedge removal

移除填充词和模糊修饰词

Before: "It is essentially worth noting that the system actually performs quite well in basically all of the scenarios that were tested."

After: "The system performs well in all tested scenarios."

提炼前： "It is essentially worth noting that the system actually performs quite well in basically all of the scenarios that were tested."

提炼后： "The system performs well in all tested scenarios."

Nominalization to verb

名词化结构转动词

Before: "We performed an investigation into the cause of the failure and made a determination that the configuration was incorrect."

After: "We investigated the failure and determined the configuration was incorrect."

提炼前： "We performed an investigation into the cause of the failure and made a determination that the configuration was incorrect."

提炼后： "We investigated the failure and determined the configuration was incorrect."

Redundancy elimination

消除冗余表述

Before: "The end result of this process is that each and every individual component is tested and verified to ensure and confirm that it meets the required specifications and standards."

After: "This process verifies each component meets the required specifications."

提炼前： "The end result of this process is that each and every individual component is tested and verified to ensure and confirm that it meets the required specifications and standards."

提炼后： "This process verifies each component meets the required specifications."

Preserving necessary nuance

保留必要的细微差别

Before: "While the approach generally works well in most common scenarios, there are some edge cases, particularly those involving concurrent access patterns, where the current implementation may exhibit degraded performance characteristics."

After: "The approach works well in common scenarios but may degrade under concurrent access patterns."

Note: "may" is preserved — it's a genuine qualification, not a hedge.

提炼前： "While the approach generally works well in most common scenarios, there are some edge cases, particularly those involving concurrent access patterns, where the current implementation may exhibit degraded performance characteristics."

提炼后： "The approach works well in common scenarios but may degrade under concurrent access patterns."

说明：保留了"may"——这是真实的限定，而非模糊修饰。

Agentic Optimizations

Agent优化策略

Context	Approach
Short text (< 100 words)	Distill inline, show before/after
Medium text (100-500 words)	Distill in sections, show word count reduction
Long text or file (> 500 words)	Read file, distill, write result, show stats
Preserving technical accuracy	Flag any cuts that might alter technical meaning

场景	处理方式
短文本（< 100词）	内联提炼，显示提炼前后内容
中等长度文本（100-500词）	分部分提炼，显示字数精简比例
长文本或文件（> 500词）	读取文件、提炼内容、输出结果、显示统计数据
需保留技术准确性	标记任何可能改变技术含义的精简操作