api-reference

Shipped Job Portal Skills in the AI Job Search Framework: Complete Reference

August 29, 2026 MadsLorentzen/ai-job-search ↗

The AI Job Search framework ships six production-ready job portal skills—LinkedIn, JobDanmark, Jobnet, Jobindex, Jobbank, and Freehire—each providing a standardized SKILL.md descriptor and TypeScript CLI implementation for automated job querying.

The MadsLorentzen/ai-job-search repository provides a modular architecture for automating employment searches through reusable job portal skills. These skills reside in the hidden .agents/skills/ directory and encapsulate platform-specific logic for querying, parsing, and returning structured job listings from major recruitment sites.

Available Job Portal Skills

As of the master branch, the framework distributes six ready-to-use portal skills:

Skill Architecture and File Structure

Each job portal skill follows a consistent three-part pattern that ensures interoperability across the framework.

SKILL.md Descriptors

Every skill contains a SKILL.md markdown file that defines the portal's search API, required authentication parameters, and output schema. These descriptors serve as both documentation and machine-readable specifications for the AI orchestration layer.

CLI Implementation

The command-line interface for each skill is implemented as a standalone Node/TypeScript package. The entry point resides at cli/src/cli.ts within each skill directory (e.g., linkedin-search/cli/src/cli.ts). These packages compile to executable binaries invocable via npx, accepting standardized flags for queries, limits, and output formats.

Test Coverage

Comprehensive Jest test suites live in cli/tests/ directories (e.g., linkedin-search/cli/tests/search.test.ts). These ensure robust parsing of HTML, RSS feeds, and JSON-LD structured data across different job portal response formats.

Using Job Portal Skills

The framework supports both direct CLI invocation and programmatic integration.

Command-Line Interface

Invoke any skill directly using npx without installing the package globally:


# Fetch the first 5 LinkedIn job results for "data scientist"

npx ai-job-search linkedin-search \
  --query "data scientist" \
  --limit 5 \
  --output json

Programmatic Integration

Import the skill runner to execute searches within Node.js applications:

import { runSkill } from '@ai-job-search/skill-runner';

// Invoke the Jobbank skill
const results = await runSkill('jobbank-search', {
  query: 'backend engineer',
  limit: 10,
  format: 'json',
});

console.log(results);

Workflow Orchestration

Integrate skills into higher-level automation workflows using YAML configuration:


# .claude/skills/custom-job-search/SKILL.md

steps:
  - name: LinkedIn
    skill: linkedin-search
    args:
      query: "{{candidate.title}}"
      limit: 20
  - name: Jobbank
    skill: jobbank-search
    args:
      query: "{{candidate.title}}"
      limit: 20

Extending the Framework

New job portal skills can be added by replicating the existing directory structure under .agents/skills/. Each new skill requires a SKILL.md descriptor, a TypeScript CLI implementation following the cli/src/cli.ts convention, and corresponding Jest tests in cli/tests/ to maintain framework consistency.

Summary

  • Six job portal skills ship with the AI Job Search framework: LinkedIn, JobDanmark, Jobnet, Jobindex, Jobbank, and Freehire
  • All skills reside in the .agents/skills/ directory with standardized file layouts
  • Each skill provides a SKILL.md descriptor, TypeScript CLI implementation, and Jest test suite
  • Skills support both npx command-line execution and programmatic import via @ai-job-search/skill-runner
  • CLI entry points are located at cli/src/cli.ts within each skill subdirectory
  • Test files are stored in cli/tests/ directories to ensure parsing reliability

Frequently Asked Questions

What job portals are supported by the AI Job Search framework?

The framework currently supports six major platforms: LinkedIn, JobDanmark, Jobnet, Jobindex, Jobbank, and Freehire. Each is implemented as a distinct skill module with dedicated parsing logic for that portal's specific HTML structure or API response format.

How do I run a job portal skill from the command line?

Use the npx ai-job-search [skill-name] command followed by platform-specific flags. For example, npx ai-job-search linkedin-search --query "python developer" --limit 10 executes the LinkedIn skill. The CLI implementations in cli/src/cli.ts handle argument parsing and result formatting.

Can I use these job portal skills in my own Node.js application?

Yes. Import the runSkill function from @ai-job-search/skill-runner and pass the skill name (e.g., 'jobbank-search') along with a configuration object containing query, limit, and format parameters. This returns a promise that resolves with structured job listing data.

Where are the skill descriptors and CLI implementations located?

All job portal skills are stored in the .agents/skills/ directory at the repository root. Each skill subdirectory contains a SKILL.md file describing the portal's capabilities, a cli/ folder with package.json and src/cli.ts containing the TypeScript implementation, and a cli/tests/ directory with Jest test suites.

Have a question about this repo?

These articles cover the highlights, but your codebase questions are specific. Give your agent direct access to the source. Share this with your agent to get started:

Share the following with your agent to get started:
curl -s "https://instagit.com/install.md"
Works with
Claude Codex Cursor VS Code OpenClaw Any MCP Client

Maintain an open-source project? Get it listed too →