OpenClaw beta 5 broadens the May 12 upgrade into channel and runtime recovery work
OpenClaw v2026.5.10-beta.5 supersedes beta.4 as the current prerelease. The useful change for operators is breadth: Fal image edits now route GPT Image 2 and Nano Banana 2 reference-image requests through the edit endpoint with aspect-ratio and resolution handling; Control UI shows a plain recovery panel when the app module never registers instead of leaving a blank dashboard; agent-to-agent reply chains can be allowed up to 20 turns while the default stays conservative; sandboxed and public agents get per-agent message send/cross-context restrictions; timed-out Codex app-server clients are retired so Discord agents do not reuse CPU-spinning processes; and Slack, Telegram, WhatsApp, Cron, Gateway, provider, memory, and transcript fixes continue the beta.4 reliability theme.
The May 12 OpenClaw story is shifting from one headline feature to operational recovery: when a channel breaks, a dashboard fails to boot, a Codex process times out, a transcript grows huge, or a public agent tries to message outside its lane, the runtime needs bounded behavior and visible repair paths. That is high reader value for people running always-on agents, but the prerelease still touches many subsystems and deserves a narrow rollout.
- Official v2026.5.10-beta.5 release notes were published at 2026-05-11T16:38Z, which is 2026-05-12 in Asia/Shanghai
- Release notes add Fal edit routing for GPT Image 2 and Nano Banana 2 reference-image edits, including aspect ratio, resolution, and higher input-image caps
- Release notes add a plain HTML Control UI recovery panel for app-module registration failures, fixing blank dashboard pages
- Release notes allow session.agentToAgent.maxPingPongTurns up to 20 while keeping the default at 5, and retain per-agent message cross-context/action allow overrides for sandboxed and public agents
- Release notes retire timed-out Codex app-server clients after bounded turn interrupts and keep long transcript reads on streaming helpers
- Fresh issue #80698 still documents remaining credential-isolation pressure: global env vars can leak GitHub tokens across agents until per-agent env scoping exists
- This is still a prerelease with a very wide change surface
- The release bundles UI, channel, provider, model, memory, Codex, Cron, and SDK changes, so regressions should be isolated by channel or workflow
- Per-agent message policy is improving, but per-agent environment variable isolation remains an open request
- Release notes are large; teams should verify the specific fixes they need instead of assuming the whole beta line is safer for their deployment