extract-job-posting

The skill handles job posting extraction by using Chrome MCP to navigate platforms like Wanted, Saramin, JobKorea, Remember, and Jumpit, fetching text, evaluating JavaScript, and saving normalized markdown.

Yes, job posting extraction handles image-only ads by applying OCR to capture text that standard page scraping cannot retrieve. This ensures structured data extraction even from image-based job descriptions on supported recruiting platforms.

Yes, you can capture job descriptions from multiple platforms and store them consistently. The skill normalizes fields like company, position, and experience, saving each posting as structured markdown under private/job_postings with a uniform filename.

To avoid duplicate job postings during extraction, the skill checks for existing entries via the JD pipeline before saving. This deduplication process ensures only unique job postings are persisted under the private/job_postings directory.

To automate job description capture from Korean recruiting sites, you need the Chrome MCP environment configured for browser navigation and JavaScript evaluation. The skill then handles platform-specific scraping and OCR to fetch and persist the posting automatically.

Limitations when extracting job data from JobKorea or Jumpit include reliance on Chrome MCP for browser automation and potential challenges with dynamically loaded content or image-only ads, though OCR is applied to handle the latter.