agentleFS
JakeyhbJakeyhb/football-content-ops-skill/skills/jev-judge/SKILL.md

jev-judge

Skill1 starsChanged 14 days agoMITView on GitHub

Judges a note before distribution: what kind it is, which pool it sits in, and whether Douyin is allowed. Use when the user says 过闸, 能不能发抖音, 这篇发哪, 判断一下, or before dy-note-adapt.

  • Reads credentials

2 other projects use a version of this.

Adopted by

Most recently taken up by

The file

Files
SKILL.md

jev-judge — 发布前的判断层

写正文、改标题、出图仍然走原来的 skill。这一步只回答:这篇是什么、落在哪个池、抖音过不过闸。

Jev 是判断模型,不写文案。它一次返回形态、池子、冲突钩和概率。发不发、转不转,由 scripts/jev-judge.py 对照 dy-publish 和小红书复盘闸决定。把握不够就停下来人工看,不要让写文章的模型临场猜。

什么时候跑

  • 准备 dy-note-adapt 之前
  • 用户问这篇发小红书还是也发抖音
  • 选题是今晨汇总、亚冠、国安、海港、利兹、意甲、普通中超

小红书正文、对撞板、去 AI 味,不要改用这一步。

命令

python3 scripts/jev-judge.py --topic-dir "doc/<主题>"
退出码做什么
0小红书照发。可以转抖音,封面另做夜场比分卡,配乐,间隔 ≥4 小时
2只发小红书。不要写 douyin-*.txt,不要入库抖音表
3先人工看。不要自动转抖音
4这条不出
1API 失败。停,不要改跑关键词假装过了闸

关键词兜底(输出第一行是 rules):

python3 scripts/jev-judge.py --rules --topic-dir "doc/<主题>"

配给 Claude

稳定约定在 .cursor/rules/jev-claude.mdc。Claude 仍是主模型。Jev 不进模型列表,不打 /chat/completions。

export JEV_API_KEY=...   # https://www.jevai.org/agent/keys

判断请求 POST https://www.jevai.org。原生接口 /api/v1/decisions 的 model 用 typesafe-ai/jev。要挑模型走 /api/v1/decisions/model-route,候选必须是这台环境真能调用的模型。

钥匙(本仓库脚本)

jev-judge.py 有一把就走模型,没有就走关键词。模型名默认钉在 jev-1.13.0,调过阈值不要改成 jev-latest。

export TYPESAFE_API_KEY=...
# 或者只配 OpenRouter,脚本会打到 /api/v1/systemone
export OPENROUTER_API_KEY=...

闸(脚本里,和规则文件同一套)

可转抖音:拜仁/凯恩、曼城、曼联、阿森纳、欧冠夜场的赛前前瞻;或这些池的单场冲突复盘(逆转、屠杀、绝杀、大比分)。中超只有冠军悬念。

只留小红书:亚冠、海港、国安、德甲其他队、意甲散场、利兹、逐帧、人物数据、比分汇总。

比分汇总可以发,不做当天主菜。封面钩子仍是数字或冲突;标题写成「A vs B」会在结果里标出来。

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.