job-search-de

🇩🇪 job-search-de

Find jobs in Germany, compare them with your CV, and track applications in a browser workbench. Works across professions as a skill for your AI coding agent.

English · Deutsch · 中文 · 日本語 · 한국어

Quick start

You need an AI agent that can load skills and run commands, internet access, and Python 3. The installer below also needs Node.js/npm (npx). This skill requires no additional LLM API key; your agent's usual costs still apply.

1. Install

npx skills add Kevoyuan/job-search-de -g

Select your agent in the installer. If the skill does not appear, restart the agent or open a new session.

Clone into your agent's skill directory. For agents that read ~/.agents/skills/:

git clone https://github.com/Kevoyuan/job-search-de.git ~/.agents/skills/job-search-de

If your agent uses a different directory, change the destination accordingly.

2. Provide your CV

Open a folder for your job search in your agent. Attach your CV or give its file path, such as ./resume.pdf or ./CV.md.

No CV ready? Describe your experience and preferences. You can also request job leads first, without personal match scores.

3. Ask the agent

Copy this and replace the role and locations:

Use job-search-de. My CV is at ./resume.pdf.
Find software engineer jobs in Berlin or remote within Germany.
Prioritize English-speaking roles posted in the last 14 days.
Generate a match report and a job workbench in English.

The agent extracts your background, asks for missing preferences, searches company hiring pages, checks availability and dates, and compares requirements with evidence from your CV.

What you get

• Job shortlist: official application links, locations, and freshness labels. Missing dates are marked as unknown.
• Match report: strengths, gaps, and supporting CV evidence.
• Browser workbench: open the generated job-hunt-workbench.html to filter jobs, switch between table and kanban views, and track applications. Includes five themes and Chinese, English, and German interfaces.

The skill does not submit applications or contact employers without your explicit authorization.

The Analysis view explores collected jobs by location, industry, role, technology and work model. Click charts to filter matching jobs. Missing evidence stays unknown. See analytics data and counting.

Everyday use

Ask in plain language, or use these shortcuts in your agent chat after loading the skill. They are not terminal commands; if your agent reserves slash commands, use the plain-language request instead.

Change your preferences

Tell the agent what to change, for example:

Search Munich too, exclude senior roles, and change the workbench to German.

The agent stores your profile and preferences in .job-search/ inside your job-search folder. You do not need to edit these files manually.

For manual settings and optional company/keyword lists, see the configuration guide.

Keep the skill up to date

The workbench checks for new releases when it is built and again when you open it. If a newer version exists, a banner appears at the top with a copy button for either update path. Open the Usage Guide (the book icon in the header, or press G) to see your installed version, re-check manually, and copy the commands.

In your agent chat:

/update-skill

Or straight from a terminal:

npx skills update job-search-de -g

To check the published version without updating:

python3 scripts/check_update.py

Data and privacy

Profile and configuration files are stored locally. Your AI agent may process CV content through its model provider, depending on its settings. Job discovery needs internet access; optional Notion sync sends selected data to Notion.

Availability checks reflect the time of the search and cannot guarantee that a job stays open.

More details

• Skill workflow and commands
• How matching is scored
• Interactive architecture diagram — download and open in a browser
• MIT License · Third-party chart and font licenses