All notable changes to the JobStreet scraper are documented here.
- Rebuilt the Actor around bounded native HTTPS/Cheerio parsing of official JobStreet search and detail pages.
- Added
search, searchMultiple, single, multiple, and startUrls modes with strict host, URL, posting-ID, canonical, title/employer, and query verification.
- Added complete-record buffering, rich JSON-LD/DOM fields, explicit application-link provenance, strict schemas, tests, KVS run receipts, and a recursive local dataset validator.
- Removed the legacy Playwright/Crawlee/proxy/fingerprint route and switched the Docker image to the Node-only Actor base.
- Target blocks fail closed with zero dataset rows and structured diagnostics; no challenge bypass or fabricated application URL is used.
The former browser-based implementation has been superseded by version 2.0.0.