OpenClaw Newsletter

• openclaw 2026.2.25-beta.1 — Improved Android chat streaming delivery and enhanced GitHub-flavored markdown rendering quality in native UI. Release notes
• openclaw-cli 2026.2.26 — Latest version available on Homebrew with 4,014 installs in 30 days. Install
• OpenClaw repo milestone — Hit 238,477 stars (+68k growth) with 46,001 forks and 918 contributors. GitHub

• TUI model command silently failing — Issue by fwends reports /model command appears to succeed but doesn't actually switch models, breaking core TUI functionality. 12 engagement points.
• Orphaned browser processes causing memory leaks — Critical bug by huaguihai where gateway restarts leave Chromium processes running, accumulating memory usage over time in production deployments.
• Discord WebSocket infinite resume loop fixed — PR by A386official adds circuit breaker to prevent endless retry loops with stale session tokens, resolving production stability issues.
• ACP client runtime adds external agent support — Major feature by snese allows OpenClaw to use any standard ACP agent as harness, expanding integration possibilities beyond current scope.
• 127 total sponsors supporting OpenClaw — Sponsor page shows growing financial backing including recent additions like romainhuet and davemorin, indicating strong community momentum.

• Google restricting Google AI Pro/Ultra subscribers for using OpenClaw — Google is reportedly banning users who access their AI services through OpenClaw, sparking major controversy with 800+ upvotes and 700+ comments on HN
• OpenClaw is dangerous — Critical analysis arguing OpenClaw poses security risks, generating heated debate with 80+ upvotes and 90+ comments
• Setting up OpenClaw on a cloud VM — SkyPilot tutorial for deploying OpenClaw in the cloud, gaining solid traction with 77 upvotes and 58 comments
• OpenClaw security assessment [pdf] — Independent security audit reveals potential vulnerabilities in the platform, discussed across 60+ upvotes and 20 comments
• The OpenClaw ecosystem is exploding — Comprehensive mapping of key ecosystem players and tools just 2 months post-launch, showing rapid community growth

• Malicious Skill Trap: Hundreds of malicious OpenClaw skills are disguising themselves as legitimate helpers, making manual review impractical for users. Review your installed skills immediately and enable strict validation modes. Read security analysis