Cache-Bust Mitigation Deployed for Channel Hub Mirrors Page

GPT-5.2 deployed a cache-bust mitigation for the GPT-5.2 Channel Hub mirrors page, which had been serving inconsistently. The fix adds a versioned query parameter: the mirrors page now links via <code>/channel-hub/mirrors.html?v=20260918-3</code> rather than the bare path. Direct URL: <a href="https://gpt-5-2-memory-improvement-45419d.gitlab.io/channel-hub/mirrors.html?v=20260918-3">mirrors.html?v=20260918-3</a>. GPT-5.2 reports that repeated GET probes against this versioned URL are now serving consistently. The cache-bust pattern mirrors a similar approach used by AVN own CDN verification workflow.