127 wires and counting

$ follow OpenClaw

Keep up with OpenClaw in about 3 minutes: what actually shipped — the commits, pull requests, releases, and security advisories that matter.

or

fair warning: these emails are deeply technical. diffs, version numbers, CVEs, benchmark deltas. if that's not your idea of a good read, this isn't your newsletter.

Folds into your digest — weekly by default, monthly if you prefer. Unsubscribe in one click.

$ status

wire 2026-08-27
stories 144

© 2026 RepoJournal Home Showcase Explore How it works Privacy

$ the-wire · showcase

Crabbox hardens checkpoint forks and lifecycle claims

By RepoJournal · Filed · About OpenClaw

Crabbox shipped replay-safe checkpoint forks with exact lease IDs plus a wave of claim-fencing fixes for machines, Morph instances, and Semaphore jobs.

Crabbox now lets external orchestrators fork a checkpoint under a fixed lease ID with `checkpoint fork <chk> --lease-id cbx_<12 hex>`, using warmup-equivalent replay adoption, and both `checkpoint create` and `fork` gained machine-readable output [1]. Separately, `checkpoint inspect --verify --json` and `checkpoint delete` used to exit 2 when the local record was missing; they now return a verdict and delete idempotently, so automation can distinguish a dead checkpoint from an operational error [2].

Crabbox also fenced resource lifecycles behind exact scoped claims. Retained Machine0 workers lost their coordinator heartbeat after provisioning because generic heartbeat authorization compared against the wrong scope; claims are now bound to `machine0:<cloud-id>` [3]. Morph instances require an exact API-scoped local ownership claim before pause, delete, or cleanup, with claims and SSH keys preserved on failures [4]. Semaphore CI jobs require exact org-host/project/job claims before stopping, with idempotent cleanup of already-finished jobs [5].

In the agents, direct-message turns now persist sender identity (`senderId`, `senderName`, `senderUsername`); previously the Control UI rendered the operator fallback `You` because transcript persistence dropped the sender [6]. Cloud session failures now show one shared error notice with a readable, selectable diagnostic and a copy control, instead of a single truncated line [7].

Action items

References

  1. [1] feat: replay-safe fixed checkpoint forks and JSON output ↗ openclaw/crabbox
  2. [2] fix: machine-readable missing-checkpoint verdicts and idempotent delete ↗ openclaw/crabbox
  3. [3] fix(machine0): retained workers lose coordinator heartbeat ↗ openclaw/crabbox
  4. [4] fix(morph): fence instance lifecycle with exact scoped claims ↗ openclaw/crabbox
  5. [5] fix(semaphore): fence job lifecycle with exact scoped claims ↗ openclaw/crabbox
  6. [6] fix(agents): persist sender identity for direct-message turns ↗ openclaw/openclaw
  7. [7] fix(ui): make cloud failure diagnostics readable and copyable ↗ openclaw/openclaw

Quick answers

What shipped in OpenClaw on August 27, 2026?
Crabbox shipped replay-safe checkpoint forks with exact lease IDs plus a wave of claim-fencing fixes for machines, Morph instances, and Semaphore jobs. In total, 106 commits and 38 pull requests landed.
Who contributed to OpenClaw on August 27, 2026?
4 developers shipped this update, including openclaw-docs-sync[bot], Peter Steinberger, SunnyShu0925, and vincentkoc.
What were the notable OpenClaw updates?
feat: replay-safe fixed checkpoint forks and JSON output, fix: machine-readable missing-checkpoint verdicts and idempotent delete, and fix(machine0): retained workers lose coordinator heartbeat.