score-optimizer

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Score Optimizer

这个 skill 只负责一件事：优化文章评分系统本身。

它适合做的是：

调整文章爆款评分 Prompt
批量给样本文集打分
对照标注数据评估评分准确度
研究评分维度、权重、few-shot 示例是否更合理

它不负责：

直接改文章正文
平台改写
发布包装

This skill only does one thing: optimize the article scoring system itself.

It is suitable for:

Adjusting viral article scoring prompts
Batch scoring sample article collections
Evaluating scoring accuracy against labeled data
Researching whether scoring dimensions, weights, and few-shot examples are more reasonable

It is NOT responsible for:

Directly modifying article content
Platform rewriting
Publishing and packaging

Working Scope

适用场景：

用户说“优化文章评分 prompt”
用户说“让这个爆款评分系统更准”
用户说“跑一轮样本文集评分实验”
用户说“校准 viral score / composite score 的打分规则”

不适用场景：

用户说“直接把这篇文章改高分”
用户说“按这个评分器去优化文章内容”

后者应改用

article-optimizer

。

Applicable scenarios:

When the user says "optimize the article scoring prompt"
When the user says "make this viral scoring system more accurate"
When the user says "run a round of sample article collection scoring experiments"
When the user says "calibrate the scoring rules for viral score / composite score"

Inapplicable scenarios:

When the user says "directly modify this article to get a higher score"
When the user says "optimize article content according to this scorer"

The latter should use

article-optimizer

instead.

Required Reads

开始前先读：

references/score_program.md
references/score_prompt.md

需要看实现时再读：

scripts/run_scoring.py
scripts/evaluate.py
assets/articles/labels.example.json —— 标注 schema 示例

样本文集位置（用户自己提供，本 skill 不带样本）：

```
assets/articles/samples/
```
—— 把你自己标注过的 .md 文章放进来,然后
```
cp labels.example.json labels.json
```
填上你的评分。详见 assets/articles/samples/README.md

Read these before starting:

references/score_program.md
references/score_prompt.md

Read these when needing to check implementations:

scripts/run_scoring.py
scripts/evaluate.py
assets/articles/labels.example.json —— Label schema example

Sample article collection location (provided by the user, this skill does not include samples):

```
assets/articles/samples/
```
—— Place your own labeled .md articles here, then
```
cp labels.example.json labels.json
```
and fill in your scores. See details in assets/articles/samples/README.md

Workflow

先确认当前任务是在优化评分器，不是在改文章。
读取
```
references/score_program.md
```
，按它的实验约束执行。
默认只改
```
references/score_prompt.md
```
，不要动评估脚本。
运行批量评分脚本，再运行评估脚本。

比较

composite_score

、

mae

、

spearman

、

classification

。

记录结果，决定保留还是回滚。

First confirm that the current task is optimizing the scorer, not modifying articles.
Read
```
references/score_program.md
```
and execute according to its experimental constraints.
By default, only modify
```
references/score_prompt.md
```
; do not touch the evaluation scripts.
Run the batch scoring script, then run the evaluation script.

Compare

composite_score

mae

spearman

classification

Record the results and decide whether to keep or roll back.

Commands

在 skill 根目录执行：

bash

python3 scripts/run_scoring.py > scoring.log 2>&1
python3 scripts/evaluate.py > eval.log 2>&1
grep "^composite_score:\|^mae:\|^spearman:\|^classification:" eval.log

如果需要图表：

bash

python3 scripts/generate_score_chart.py

Execute in the skill root directory:

bash

python3 scripts/run_scoring.py > scoring.log 2>&1
python3 scripts/evaluate.py > eval.log 2>&1
grep "^composite_score:\|^mae:\|^spearman:\|^classification:" eval.log

If charts are needed:

bash

python3 scripts/generate_score_chart.py

Guardrails

默认不要改
```
scripts/evaluate.py
```
默认不要改
```
scripts/run_scoring.py
```
默认不要安装新依赖
如果只是想提高某一篇文章分数，不要用这个 skill

Do not modify
```
scripts/evaluate.py
```
by default
Do not modify
```
scripts/run_scoring.py
```
by default
Do not install new dependencies by default
Do not use this skill if you only want to increase the score of a single article