Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add yanliudesign/offer-toolkit-skill --skill job-search-skillgit clone --depth 1 https://github.com/yanliudesign/offer-toolkit-skillWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/yanliudesign/offer-toolkit-skill/job-search-skill)<a href="https://agentmods.dev/skills/yanliudesign/offer-toolkit-skill/job-search-skill"><img src="https://agentmods.dev/badge/skills/yanliudesign/offer-toolkit-skill/job-search-skill/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/yanliudesign/offer-toolkit-skill/job-search-skill"><img src="https://agentmods.dev/badge/skills/yanliudesign/offer-toolkit-skill/job-search-skill.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00136 | $0.01966 |
| Opus 5 | $0.00068 | $0.00983 |
| Sonnet 5 | $0.00027 | $0.00393 |
| Haiku 4.5 | $0.00014 | $0.00197 |
Grade A, and why
linkedin-job-search-skill scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 178 lines — stays where its author put it; the contents beside it link to each section on GitHub.
LinkedIn Job Search Skill
把「一份我喜欢的 JD + 我的简历」转成一份有证据、可行动的 LinkedIn 岗位 shortlist。
这个 skill 是 Offer Toolkit 的第 0 步:
种子 JD + 简历 → 搜索画像 → LinkedIn 候选池 → 匹配排序 → shortlist
↓
Job Description Skill 深度解码
边界
- 只做岗位搜索、读取公开职位信息、去重、匹配排序和推荐。
- 永远不自动点击 Apply,不填写申请表,不发送消息,不代表用户投递。
- 永远不索取或处理 LinkedIn 密码、验证码、Cookie 或 session token。
- 遇到登录墙、验证码、频率限制或 robots 限制就停止该路径,不尝试绕过;改用公开搜索结果或输出可点击的 LinkedIn 查询链接。
- 岗位是否仍开放只能按搜索当时页面判断;报告必须写明搜索时间。
输入流程
Step 1 · 收集种子 JD
先要一份用户真正感兴趣的 JD 链接或全文。它不是唯一目标,而是用来提取岗位语义:title、level、scope、domain、核心能力和排除项。
若用户没有种子 JD,接受 2-3 个目标 title + 一句方向描述作为降级输入,并标注「无种子 JD,搜索画像置信度较低」。
Step 2 · 收集简历
接受 PDF、Word、纯文本或已结构化简历。只从用户提供的事实提取:
- 当前/最近 title 与大致 level
- 年限和最近 3 年的核心 scope
- 3-6 个有直接证据的能力
- 行业、产品阶段、客户类型和团队类型
- 地点、语言、签证等明确约束
不得把种子 JD 的要求写回用户画像,除非简历里有直接证据。
Step 3 · 补齐硬筛选条件
只追问会显著改变结果集的缺失条件,并且一次只问一个:
- 工作地点,以及是否接受 remote / hybrid / relocation
- 目标 level 或可接受的上下浮动
- 时间范围:24 小时 / 7 天 / 30 天
- 必须排除的公司、行业、合同类型或签证条件
默认值:最近 7 天、full-time、目标 level 上下浮动一级。默认值必须在最终报告显式列出,不能静默假设。
搜索执行
Step 4 · 生成搜索画像
先在内存中形成这份结构:
target_titles: [] # 2-5 个,不堆同义词
level: ""
locations: []
workplace: [] # remote / hybrid / on-site
date_posted: "7d"
employment_types: []
core_capabilities: [] # 3-6 个,必须有简历证据
domains: [] # 0-3 个
company_preferences: []
exclusions: []
seed_signals: [] # 从种子 JD 提取,但不等同于简历能力
回显一段不超过 8 行的搜索画像供用户快速纠错。用户没有纠正就继续,不要求第二次确认。
Step 5 · 构造查询矩阵
生成 6-10 组互补查询,而不是把所有词塞进一个 query:
- Exact title:目标 title + 地点
- Adjacent title:相邻 title + 地点
- Capability-led:title + 1 个核心能力
- Domain-led:title + 目标 domain
- Scope-led:title + 0→1 / platform / enterprise / growth 等 scope 信号
每组查询只放 1-2 个判别词。过长查询会漏掉用词不同但实际匹配的岗位。
LinkedIn 查询 URL 使用公开 Jobs Search 参数;优先设置 keywords、location、date posted、workplace 和 employment type。记录每组实际执行的 query 与 URL。
Step 6 · 采集候选池
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 12d ago First seen · 178 lines · 136 tokens per session scan A 6afd8d8ad250
linkedin-job-search-skill is a skill published in the GitHub repository yanliudesign/offer-toolkit-skill (425 stars, last pushed yesterday), licensed MIT. It adds 136 tokens to every session and 1,966 once invoked, about $0.0007 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
exam-study-guide
A study-guide builder for a completed exam chapter that has not yet passed its required checks. It creates a structured teaching checklist and, in full mode, a self-contained HTML and printable PDF guide with readable formulas, visible images, explanations, examples, and answers.
exam-ingest
A workspace setup tool for exam preparation. It reads PDFs, Word documents, presentations, spreadsheets, images, text files, and Markdown, then creates chapter notes, a question bank, and tracked preparation data.
exam-tutor
A teaching workflow for learning one chapter from a study wiki at a time, using everyday comparisons and step-by-step explanations of formulas and important questions. TDD is not involved here; the material concerns guided study and exam preparation.
exam-audit
A read-only health check for an exam-preparation workspace, meaning the folder that stores study materials, questions, notes, plans, and progress. It reports missing, inconsistent, or incomplete parts without changing them.
exam-cram-coach
A structured last-minute exam study workflow that turns course materials or a syllabus into chapter notes, practice questions, grading, and progress tracking.
exam-review
Clear recorded mistakes/confusions before the exam. Replay only existing records; teach no new chapter and invent no question.