Added opt-in bounded website enrichment with canonical-domain deduplication, same-domain redirect validation, abortable request deadlines, a 30-second domain budget, HTML-only parsing, and directory fallback on failure.
Added explicit website extraction for business emails, phones, contact names, social links, descriptions, services, and industries with source provenance.
Added website attempts, success, failure, page, contact, service, industry, retry, and domain-timeout metrics to OUTPUT.
Added website fixtures and regression coverage for extraction, duplicate domains, retries, redirects, invalid content, timeouts, and disabled enrichment.
Build 0.1.8 passed the 15-case directory-only cloud regression with usable results in every case and no QuickBooks source failures.
The build 0.1.8 website gate preserved all directory rows but enriched 6 of 17 domains; redirect canonicalization and failed-attempt website provenance remain release blockers.
Aligned the Console input schema and sample with the runtime default extractContacts: true.
