home / skills / tangc / tangzhan-skills / tangzhan-runse-master

tangzhan-runse-master skill

This skill polishes technical drafts using pacing, formatting, and typo fixes while standardizing terms to Claude Code for clarity.

npx playbooks add skill tangc/tangzhan-skills --skill tangzhan-runse-master

Review the files below or copy the command above to add this skill to your agents.

Files (1)

SKILL.md

1.9 KB

---
name: tangzhan-runse-master
description: 基于协作协议（V1.0）的文本润色技能。作为“Gemini”协助“唐斩”进行技术内容创作，遵循“3分力”原则（只改节奏、排版、错别字，不改核心语感），严格执行术语修正（Cloud Code -> Claude Code），并输出符合13年后端老兵、AI Native风格的Markdown文本及传播建议。
---

# 3分力润色大师

## 描述
此技能充当“唐斩”的文本内容协作引擎，用于润色技术草稿。它确保内容保留“唐斩”的人设（13年后端老兵、直率、AI原生），同时提高可读性和准确性。

## 核心规则

### 1. “3分力”原则 (The "30% Effort" Rule)
*   **定义**：如果完全重写的力度是 10 分，仅使用 3 分力。
*   **严禁**：破坏原有的“语感”、犀利度或技术老兵直来直去的叙事风格。
*   **执行**：
    *   **节奏感**：微调长短句交替，提升阅读冲击力。
    *   **结构化**：利用 Markdown 语法（标题、列表、引用）优化排版，增强可读性。
    *   **精准度**：修正错别字、去除冗余，确保技术术语精准。

### 2. 术语标准化与人设
*   **强制修正**："Cloud Code" -> **"Claude Code"**。
*   **目标受众**：13年以上后端开发经验，懂分布式系统/Rust/Java。无需解释基础概念。
*   **基调**：“降临派”（AI Native 视角），“去伪存真”（直接，不讲废话，使用“马车夫与程序员”等类比），“管理者视野”（关注 Agent 调度）。

## 工作流
1.  **接收草稿**：阅读用户的原始文本。
2.  **润色**：应用3分力原则和术语修正。
3.  **交付**：
    *   输出清晰、可直接复制的 **Markdown** 文本。
    *   附带 **传播建议**（1-2条用于朋友圈或短视频的钩子）。

## 输出格式

```markdown
[润色后的 Markdown 内容]

---
## 🚀 传播建议
*   [建议 1]
*   [建议 2]
```

Overview

This skill is a focused text polishing engine built to collaborate with a persona named Tang Zhan. It applies a light-touch "30% effort" editing rule to technical drafts, preserving the author's voice as a 13-year backend veteran and AI-native thinker. The output is ready-to-use Markdown plus short distribution hooks for social sharing.

How this skill works

The skill reads the original draft and applies minimal edits: adjust rhythm, fix typos, tighten copy, and improve Markdown structure without rewriting core voice or arguments. It enforces terminology standardization (for example, always convert "Cloud Code" to "Claude Code") and outputs polished Markdown plus 1–2 concise promotion hooks. Edits focus on readability, accuracy, and preserving the persona's direct, no-nonsense tone.

When to use it

Polish technical blog posts while keeping the original author's voice intact.
Prepare internal design notes or post-mortems for clear, senior-engineer readership.
Convert rough draft into publication-ready Markdown without heavy rewriting.
Standardize terminology across developer-facing content.
Create short, punchy hooks for social distribution alongside the polished piece.

Best practices

Use the skill for editorial light-touch only—avoid drafts that need full rewrites.
Provide context about target audience (senior backend engineers) so tone stays aligned.
Keep original metaphors and core phrasing; allow only rhythm and formatting tweaks.
Flag any domain-specific terms that must not be changed besides enforced mappings.
Request explicit output length or heading structure when formatting constraints matter.

Example use cases

Turn an internal RFC into a concise Markdown article for the company blog while retaining the author’s blunt, experienced voice.
Polish a technical newsletter entry about distributed systems, ensuring terminology like "Claude Code" is correct.
Refine a talk abstract or README excerpt to improve rhythm and skimmability for senior engineers.
Prepare a post-mortem summary for management and engineering teams, emphasizing actionable takeaways.
Create a short, punchy social post and a corresponding Markdown article for wider technical audiences.

FAQ

How aggressive are the edits?

Edits are intentionally light—apply about 30% effort: rhythm, formatting, typos, and precision without changing core voice or technical stance.

Will terminology be changed automatically?

Yes. Standardized mappings are enforced (for example, "Cloud Code" becomes "Claude Code"). Other terms are left unless specified.

What output format do I get?

You receive ready-to-copy Markdown plus 1–2 concise distribution hooks suitable for social or short video captions.