The v26.7.0 release brings Slack MCP integration across all three GA agent types (hermes, openclaw, zeroclaw), completes workspace overlay sync on macOS for the full agent matrix, adds macOS support for clawctl agent shell, resolves the openclaw canonical-render gap, tightens sync validation and health-checking, and bumps all upstream agent pins to their latest versions.
The v26.6.5 release adds a Playwright demo skill, agent-scoped audit trails, workspace overlay sync, Brave Search integration, OpenCode provider support, and an agent shell command.
The v26.6.3 release refines the user experience with a new provider type, a restructured GUI, and a critical test fix. The legacy clm CLI is removed, and all users must migrate to clawctl.
This week, we made it easier to connect to AI proxy services and improved how you navigate the GUI. The changes are small but meaningful — no more guessing where to find your agents, and a new way to use LiteLLM-compatible backends.
The biggest release yet: a kubectl-style CLI, macOS support, native web UIs for all agent types, and 100+ improvements across the stack.
If you've been using Clawrium with clm, the next install replaces it with clawctl. Same hosts, same agents, same ~/.config/clawrium/ state — a single, consistent, kubectl-style command surface across the entire fleet.
I installed my first [OpenClaw][openclaw] on an old Ubuntu box. Smooth. Then I dug up a Raspberry Pi and put [ZeroClaw][zeroclaw] on it. A week later I was spending more time on SSH gymnastics than actually using the agents — upgrading one, swapping a model on another, debugging logs on a third, rotating a token on the fourth. The agents were fine. The orchestration around them was the problem.