OpenClaw published v2026.6.11-beta.2 late Sunday UTC, giving operators another beta cut focused on channel control, mobile configuration, plugin distribution, and agent-turn reliability. The release was published at 2026-06-28 23:19 UTC, just after the previous nightly scan window.
The official notes describe the release as an audited contribution record over the v2026.6.10..0a4d0daa8cc99e9b2c6bc78c2e67357272a9fd53 history, covering 308 merged PRs. The grouped highlights are the useful reading path: this beta is less about one giant feature and more about tightening many operational edges that show up in real OpenClaw deployments.
More Channel Control
The largest user-facing theme is channel behavior. Slack relay mode, native Mattermost /oc_queue, and per-DM model overrides all landed in the highlighted set. Together, those changes make channel automation easier to route, inspect, and tune without forcing every workspace into the same default behavior.
The fixes list also includes Telegram progress rendering, webhook lifecycle handling, reaction directives, duplicate mirror write avoidance, queued update draining, and WhatsApp durable reply targets. That is a lot of channel plumbing, but it points at a simple operator outcome: OpenClaw is trying to preserve the intended conversation context across more messy delivery paths.
Operator Workflows Improve
Two workflow additions stand out. openclaw agent --message-file gives operators a file-driven way to send larger or prebuilt prompts into an agent run. The RAFT CLI wake bridge adds a remote wake-up path for environments where agents need to be nudged back into useful work from outside the immediate chat loop.
The release also includes channel identity hook context and per-agent usage-cost reporting. Those are not flashy features, but they matter for teams running multiple agents or channels. Routing and cost accounting become more tractable when the platform exposes the right identity and usage details.
Safer Plugin Distribution
Plugin distribution also moved forward. The release notes call out additional official plugins being externalized cleanly, plus bundled plugin icon metadata for installed clients. That fits the broader marketplace and plugin-supply-chain work OpenClaw has been shipping throughout June.
Externalizing official plugins is useful only if users can still tell what they installed and where it came from. Icon metadata is a small surface, but it makes plugin presentation clearer in clients and helps the ecosystem feel less like a list of opaque package names.
Mobile And Agent Reliability
Android settings detail panels get a specific mention in the highlights. That means mobile operators should see better visibility and control over OpenClaw configuration from the Android app, rather than needing to drop back to a terminal for every setting.
On the agent side, the release highlights Codex partial deltas, harness activation, and long-context prompt-cache stability. These are the kinds of changes that reduce lost progress and inconsistent behavior in long-running agent work. A smoother agent turn is not just a nicer transcript; it is fewer repeated instructions, fewer confused retries, and less wasted context.
Release Verification
The release includes signed assets for dependency evidence, release manifest data, and postpublish evidence. Its verification section also links to npm package metadata, release evidence, publish workflow runs, and CI validation.
For production users, this remains a beta release. The practical takeaway is that the 2026.6.11 line is gathering fixes across the exact areas where always-on OpenClaw deployments tend to hurt: channel delivery, auth and model edge cases, cron validation, plugin distribution, mobile controls, and long agent turns.