How to fix openclaw gateway high CPU usage?

Check GitHub issue #65062 for updates. As a workaround, remove unused extension plugins from ~/.openclaw/extensions/ and manually clean up malformed QMD collections.

What causes QMD collection errors in openclaw?

Malformed per-agent QMD collections are recreated on every gateway restart, generating repeated warnings that drive CPU usage up.

Why are disabled plugins loading in openclaw gateway?

A bug in provider normalization causes Graph, compliance, and skill-resolver plugins to be loaded at runtime even when explicitly disabled in config.

Which openclaw version has the gateway CPU bug?

Reported on v2026.3.12 Gateway with multi-agent setups on Linux with extension plugins in ~/.openclaw/extensions/.

← Back to dashboard

clawsmith.com/signal/openclaw-gateway-high-cpu-qmd-plugin-loading-bugs

⚠ IssueWide OpenLive

OpenClaw Gateway Startup Drives CPU Very High from Malformed QMD Collections and Unrelated Plugin Loading

Q: Why is openclaw gateway using high CPU on startup?

Two separate runtime bugs: malformed per-agent QMD collection recreation causes repeated skip/conflict warnings, and unrelated extension plugins are loaded during provider normalization even when disabled in config.

Two separate runtime bugs cause high CPU on gateway startup: malformed per-agent QMD collection recreation causes repeated skip/conflict warnings, and unrelated extension plugins (Graph, compliance, skill-resolver) are loaded during provider normalization despite being disabled in config.

Product Idea from this Signal

A CI tool that catches OpenClaw release regressions against your config before upgrading

250.9k ▲

OpenClaw shipped 13 point releases in March 2026 alone, and users report broken production setups after each update with no way to test compatibility first. There is no canary system, no community QA pipeline, and the built-in doctor command only checks basics. This tool runs your actual OpenClaw config, skills, and workflows against the latest release in an isolated sandbox, flags breaking changes, and gives you a pass/fail verdict before you hit upgrade.

CLIDEVTOOLTESTINGOPEN-SOURCE

UnderservedView Opportunity →