home / skills / zephyrwang6 / myskill / topic-reviewer

topic-reviewer skill

safe

This skill reviews topics against publication standards, returns pass or fail with actionable improvement suggestions to boost relevance and quality.

npx playbooks add skill zephyrwang6/myskill --skill topic-reviewer

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

Files (1)

SKILL.md

2.3 KB

---
name: topic-reviewer
description: 选题审核工具。检查选题是否符合发布标准，给出通过/不通过判断及修改建议。当用户说"审核选题"、"这个选题行不行"、"帮我看看这个能不能写"时触发。
---

# Topic Reviewer - 选题审核

## 审核流程

1. **接收选题**（来自topic-generator或用户输入）
2. **逐项检查**：按审核清单打分
3. **输出结论**：通过/不通过 + 理由

## 审核清单

### 必过项（一票否决）

| 检查项 | 不通过标准 |
|--------|------------|
| 热度 | 讨论量<100或无增长趋势 |
| 相关性 | 与AI/编程/效率无关 |
| 时效性 | 超过72小时的旧闻 |
| 可写性 | 没有可操作的内容输出 |

### 评分项（满分100）

| 维度 | 权重 | 评分标准 |
|------|------|----------|
| 独特角度 | 30% | 有无别人没写过的切入点 |
| 用户价值 | 25% | 读者能获得什么 |
| 国内适配 | 20% | 国内用户是否关注 |
| 标题吸引力 | 15% | 是否想点进去看 |
| 写作难度 | 10% | 能否快速高质量完成 |

**通过线：70分**

## 不通过的常见原因

1. **热度不够高**
   - 表现：源头讨论量低，无传播趋势
   - 建议：等热度起来再写，或换热点

2. **没有独特角度**
   - 表现：已有大量类似文章
   - 建议：找差异化切入点，或放弃

3. **不符合国内用户关注**
   - 表现：纯海外事件，无本土关联
   - 建议：找国内用户共鸣点，或翻译+本土化

4. **纯蹭热点无干货**
   - 表现：只是转述新闻，无增量信息
   - 建议：加入实操/测评/深度分析

5. **写作成本过高**
   - 表现：需要大量调研或技术门槛高
   - 建议：简化范围或分拆系列

## 输出格式

```markdown
## 选题审核结果

**选题**：[标题]

**结论**：✅ 通过 / ❌ 不通过

**评分**：XX/100

| 维度 | 得分 | 说明 |
|------|------|------|
| 独特角度 | /30 | |
| 用户价值 | /25 | |
| 国内适配 | /20 | |
| 标题吸引力 | /15 | |
| 写作难度 | /10 | |

**审核意见**
- 优点：...
- 问题：...
- 建议：...

**修改方向**（如不通过）
1. ...
2. ...
```

## 快速判断

没时间细评时，问自己：
1. 今天不写会后悔吗？
2. 写完能获得什么反馈？
3. 一句话能说清为什么值得写吗？

3个都是"是" = 直接写

Overview

This skill is a topic-reviewer that audits article or post ideas and judges whether they meet publishing standards. It returns a pass/fail verdict, a numeric score out of 100, and actionable revision suggestions to improve chances of publication. It triggers when users ask to audit a topic or check if an idea is worth writing.

How this skill works

It evaluates incoming topics against a two-tier checklist: hard fail conditions (one-vote veto items) and weighted scoring dimensions that sum to 100. The tool flags immediate rejects for low heat, irrelevance, staleness, or lack of actionable content, then scores unique angle, user value, domestic fit, headline pull, and writing difficulty. The output includes a clear verdict, per-dimension scores, strengths, issues, and concrete modification directions.

When to use it

Before committing time to write an article, thread, or tutorial.
When deciding whether a hot topic is timely and worth publishing now.
To get practical revision suggestions for weak or borderline ideas.
When translating or adapting overseas topics for domestic audiences.
To triage multiple topic ideas and prioritize what to write next.

Best practices

Provide source links or metrics of discussion volume so heat can be assessed accurately.
State the intended audience and publish channel (blog, short post, newsletter).
If rejected, try one concrete change (angle, localization, or added how-to) and re-run the audit.
Use the quick-judgment checklist (regret, feedback, one-sentence value) for fast decisions.
Break high-cost topics into smaller, actionable pieces to lower writing difficulty.

Example use cases

A writer proposes a topic about a niche overseas AI policy—get advice on domestic relevance and localization.
A content lead needs to prioritize five candidate headlines for this week's newsletter.
A creator wants to know if a trending product leak is worth immediate coverage.
An author has a draft angle but needs suggestions to increase uniqueness and reader value.
A team wants to decide whether to start a large investigative piece or postpone until heat rises.

FAQ

What immediate checks cause an automatic reject?

Automatic rejects occur when discussion volume is below threshold or shows no growth, the topic is irrelevant to the target domains, the information is older than 72 hours, or there is no actionable output to write about.

What score is required to pass?

The pass line is 70 out of 100 after scoring unique angle, user value, domestic fit, headline pull, and writing difficulty.