npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2026 – Pkg Stats / Ryan Hefner

@talent-scout/ai-evaluator

v0.1.1

Published

[![GitHub Actions](https://github.com/presence-io/talent-scout/actions/workflows/publish.yml/badge.svg)](https://github.com/presence-io/talent-scout/actions/workflows/publish.yml) [![npm: @talent-scout/ai-evaluator](https://img.shields.io/npm/v/%40talent-

Readme

@talent-scout/ai-evaluator

GitHub Actions npm: @talent-scout/ai-evaluator Node.js License: MIT

@talent-scout/ai-evaluator 在规则层之上增加 OpenClaw 驱动的 AI 判断。它解决两个问题:

  • 规则层无法明确判断的灰区身份
  • 对 top 候选人做更深的技术与 AI 采用深度评估

开发前提

  • Node.js 22+
  • pnpm 10+
  • openclaw 已安装
  • talents.yaml 中已配置 openclaw.agents.identityopenclaw.agents.evaluator

在仓库根目录安装依赖:

pnpm install

常用命令

pnpm --filter @talent-scout/ai-evaluator run evaluate
pnpm --filter @talent-scout/ai-evaluator run build

evaluate 读取 workspace-data/output/processed/latest/,产出 workspace-data/output/evaluated/<timestamp>/,并刷新 latest

核心模块

  • src/pipeline.ts: 评估总流程
  • src/identity-ai.ts: 灰区身份的批量 AI 推断
  • src/deep-eval.ts: top 候选人的批量深度评估
  • src/shortlist.ts: shortlist 生成
  • src/skills.ts: 运行统计聚合
  • src/query.ts: shortlist、evaluation、stats 查询接口

设计思想

1. AI 只接管“规则层难以稳定处理”的部分

这个包不是把所有评分都交给 LLM。相反,它只在两个高价值场景使用 AI:

  • 规则层无法稳定判断的身份灰区
  • 进入重点观察名单后的深度评估

这样可以控制成本,也能保留大部分排序逻辑的可解释性。

2. 批处理和 checkpoint 比“单次更聪明”更重要

OpenClaw 调用最贵的不是单次 prompt 设计,而是批量执行时的可靠性。这里的核心实现不是 fancy prompt,而是:

  • 批次切分
  • 每批落 checkpoint
  • 失败后恢复未完成批次

这决定了项目能不能稳定处理上百名候选人。

3. 最终排序是多轴组合,不是单一分数幻觉

这里保留 skill_scoreai_depth_scorereachability_scorefit_score 四个轴,再通过权重组合成最终分数。这样招聘者可以看“为什么高”,而不是只看一个无法拆解的总分。

评估流程

flowchart LR
  A[processed/latest] --> B[rule-based evaluation]
  B --> C[identity-ai batches]
  C --> D[deep-eval batches]
  D --> E[shortlist]
  D --> F[run stats]
  E --> G[evaluated/latest]
  F --> G

实现边界

  • 默认只对高优先级候选人做深度评估,避免无限扩张 AI 成本
  • 活跃度惩罚仍然保留在规则层,防止“历史光环”压过当前状态
  • 最终结果通过查询接口暴露给 Dashboard 和 skills,而不是让消费方直接读底层文件

调整这个包时需要注意

  • 改 batch 大小前,先验证 OpenClaw 超时和返回 JSON 稳定性
  • 改推荐动作阈值时,要同时看 shortlist 数量和质量是否失衡
  • 改查询接口时,要同步考虑 Dashboard 和 @talent-scout/skills 两个消费者

相关文档