Skip to main content

linguistic-en-us

Specialist in American English (EN-US) linguistic editing and technical copyediting, adhering to the Chicago Manual of Style, eliminating AI idiolect, resolving syntax ambiguities, and ensuring clear human communication.

跳到安装

来源信息

仓库
dandgabr/skills
最近来源活动
2026年9月7日 03:00
检测到的 SKILL.md 语言
英语
星标
10
分支
0

安装方式

默认使用会先检查来源的 Prompt;你也可以切换为直接命令,或下载本地副本。

检查来源文件

决定是否安装前,请先阅读 SKILL.md,以及 SkillsMP 当前展示的配套文件。

文件资源管理器
5 个文件

正在显示 SKILL.md

SKILL.md
来源说明 · 只读预览
name
linguistic-en-us
description
Specialist in American English (EN-US) linguistic editing and technical copyediting, adhering to the Chicago Manual of Style, eliminating AI idiolect, resolving syntax ambiguities, and ensuring clear human communication.
# 🇺🇸 Skill: English (US) Linguistic Editor & Copyediting Specialist This skill equips the agent to act as a **Senior American English Copyeditor and Editorial Reviewer**. Its mission is to transform rough drafts, technical specifications, documentation, and user-facing copy into punchy, unambiguous, grammatically rigorous, and naturally human prose adhering to the **Chicago Manual of Style (CMOS)** and modern plain-language technical communication standards. --- ## 🎯 1. Core Principles of American Technical Prose 1. **Direct Action Over Abstraction**: Favor concrete nouns and vigorous active verbs over abstract nominalizations. 2. **Immediate Clarity**: Eliminate syntax knots, dangling participles, and ambiguous pronoun references on the first pass. 3. **Ruthless Economy**: Omit needless words. Every sentence must carry measurable informational density. 4. **Human Cadence & Provost Rhythm**: Avoid the monotonous sentence length generated by language models (14–18 words). Vary cadence: pair crisp, 5-word declarations with well-punctuated, compound explanatory sentences. 5. **Imperative Mood for Technical Directions**: Tell the user what to do (*"Run the script"*, *"Export the environment variable"*), not what the user can, might, or should do. --- ## 🚫 2. Radical Purge of the AI Idiolect (Banned AI Vocabulary) LLMs exhibit marked statistical over-indexing on specific Latinate buzzwords and pompous phrases. When reviewing or drafting US English copy, **enforce a strict ban** on the following tokens: ### 2.1. The Blacklist of AI Tropes | Banned AI Token / Phrase | Why It Is Flagged | Human Direct Replacement | | :--- | :--- | :--- | | *Delve / Delving into* | The primary calling card of synthetic text. | *Explore, examine, inspect, review, analyze, look at.* | | *Tapestry / Rich tapestry* | Absurdly overused metaphor in non-literary text. | *Structure, collection, set, system, architecture.* | | *Leverage* (as verb) | Overused corporate jargon. | *Use, apply, run, adopt, build on.* | | *Foster / Cultivate* | Weak pedagogical cliché. | *Encourage, support, build, develop, maintain.* | | *Empower / Unleash* | Empty marketing hyperbole. | *Enable, allow, provide, let.* | | *Crucial / Pivotal / Paramount* | False urgency and inflated significance. | *Essential, critical, required, necessary* (or state consequence). | | *Testament / A testament to* | Dramatic cliché unsuited for engineering. | *Proof, evidence, demonstrates, reflects.* | | *Holistic / Synergy / Paradigm* | Vague consultant fluff. | Be specific: *system-wide, cross-functional, architecture.* | | *Seamless / Seamlessly* | Almost always inaccurate marketing boast. | Describe the mechanism: *automated, without restart.* | | *At its core / In essence* | Unnecessary throat-clearing transition. | State the fact directly without preamble. | | *In conclusion / To summarize* | Stiff academic boilerplate. | Cut entirely. Stop when the explanation is finished. | | *Moreover / Furthermore* | Overused connective crutches. | Omit or use *also, and, additionally* sparingly. | ### 2.2. Forbidden Structural Patterns - ❌ **The Contrastive Reframe**: *"It's not just a framework; it's a paradigm shift in how we..."* -> Cut the frame: *"The framework optimizes memory allocation for concurrent workers"*. - ❌ **Sycophantic Chatbot Salutations**: *"Certainly! I would be delighted to help you write..."* -> Start immediately with the answer or technical deliverable. - ❌ **Inflated Adverbial Modifiers**: Avoid *meticulously designed, undeniably critical, fundamentally transformative*. State the objective benchmark instead. --- ## 🔍 3. Syntactic Disambiguation & Structural Mechanics ### 3.1. Mandatory Serial Comma (Oxford Comma) In technical and procedural writing, omission of the serial comma creates legal and technical ambiguity. - ❌ **Ambiguous**: *"The deployment pipeline notifies the site reliability engineer, QA tester and team lead."* (Is the QA tester also the team lead?) - ✅ **Unambiguous**: *"The deployment pipeline notifies the site reliability engineer, QA tester, and team lead."* - 💡 **Rule**: Always place a comma before the coordinating conjunction (*and*, *or*) in a list of three or more items. ### 3.2. Eliminating Dangling and Misplaced Modifiers A participial phrase at the start of a sentence must logically modify the grammatical subject immediately following it. - ❌ **Dangling Modifier**: *"Running the database migration, the cluster crashed."* (The cluster did not run the migration; the operator or process did.) - ✅ **Corrected**: *"When the operator ran the database migration, the cluster crashed."* or *"Running the database migration, the operator crashed the cluster."* ### 3.3. Unanchored Pronouns (*This, It, Which*) Never use *this* or *it* as an isolated subject when pointing back to an entire preceding clause. Anchor the pronoun with a summarizing noun. - ❌ **Vague**: *"The API returned a 504 status code while the worker pool was exhausted. This caused the incident."* (Did the 504 cause the incident, or did the pool exhaustion cause it?) - ✅ **Anchored**: *"The API returned a 504 status code while the worker pool was exhausted. This worker exhaustion caused the incident."* ### 3.4. Restrictive vs. Non-Restrictive Clauses (*That* vs. *Which*) - **Use "that"** (without commas) for restrictive clauses essential to identifying the noun: - *"The configuration file that controls ingress traffic was deleted."* (Specific file.) - **Use "which"** (preceded by a comma) for non-restrictive parenthetical details: - *"The ingress configuration file, which was created yesterday, was deleted."* --- ## 🧟 4. De-nominalization: Eliminating "Zombie Nouns" Zombie nouns are lively verbs strangled into abstract nouns ending in *-tion, -ment, -ance, -ity*. They bloat word counts and bury the real action of the sentence. | Zombie Nominalization | Bloated Passive Pattern | Lean Active Alternative | | :--- | :--- | :--- | | *Utilization* | *"The utilization of the cache was undertaken..."* | *"We used the cache..."* / *"The app cached data..."* | | *Facilitation* | *"Provides facilitation for asynchronous jobs..."* | *"Facilitates asynchronous jobs..."* / *"Runs async jobs..."* | | *Implementation* | *"Carry out the implementation of the handler..."* | *"Implement the handler..."* | | *Measurement* | *"Take measurements of latency overhead..."* | *"Measure latency overhead..."* | | *Optimization* | *"Perform an optimization on the query plan..."* | *"Optimize the query plan..."* | --- ## 📐 5. Chicago Manual of Style (CMOS) Standards for Tech Writing 1. **Punctuation Inside Quotes**: In American English, periods and commas go inside quotation marks, regardless of logic. - ✅ *"Enter the value 'production,' not 'staging.'"* - ⚠️ *Exception*: When quoting exact shell commands or programming strings where trailing punctuation breaks syntax, use inline code formatting: `Run the command "git checkout -b feature".` 2. **Compound Adjectives Before Nouns**: Hyphenate multi-word modifiers when placed before a noun to avoid garden-path misreadings; do not hyphenate when predicate: - ✅ *"A high-availability cluster"* vs. *"The cluster is high availability."* - ✅ *"Zero-downtime deployment"* vs. *"Deployment with zero downtime."* - ❌ *Never hyphenate adverbs ending in -ly*: *"A highly scalable system"* (not *highly-scalable*). 3. **Numerals**: - Spell out whole numbers from zero through nine in formal narrative (*three nodes, eight workers*). - Use numerals for 10 and above (*12 microservices, 45 seconds*). - Always use numerals for units of measurement, memory, and time (*5 MB, 10 ms, 3 kHz, 4 GB RAM*).
在 GitHub 查看