diff --git a/README.md b/README.md
index 8a019e2..bbef532 100644
--- a/README.md
+++ b/README.md
@@ -92,8 +92,8 @@ Include the ClawHub link for your skill (e.g. `https://clawhub.ai/steipete/slack
 
 AI agents are only as good as the web data they can reach. Crawling at scale means dealing with JavaScript-heavy pages, rotating proxies, and anti-bot systems — you can build all that yourself, or use an API that handles it and hands your agent clean, ready-to-use data.
 
-<a href="https://s.getdesign.md/crawl-og">
-<img src="https://cdn.voltagent.dev/awesome-repo/crawlbase/crawlbase-sponsor-banner-560x120.png" alt="Crawlbase" width="560" /><br/>
+<a href="https://crawlbase.com/?utm_source=awesome-openclaw-skills&utm_medium=sponsorship&utm_campaign=voltagent_2026q3&utm_content=readme_banner">
+<picture><source media="(prefers-color-scheme: dark)" srcset="https://cdn.voltagent.dev/awesome-repo/crawlbase-new/crawlbase-sponsor-banner-dark-2760x480%402x.png"><img src="https://cdn.voltagent.dev/awesome-repo/crawlbase-new/crawlbase-sponsor-banner-light-2760x480%402x.png" alt="Crawlbase" width="690" /></picture><br/>
 Crawlbase is web data infrastructure trusted by 70,000+ developers: one API to crawl any URL at scale, with JS rendering, proxy rotation and anti-bot handling. Its MCP server gives agents live web access: crawl, crawl_markdown, crawl_screenshot.
 </a>