120 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 2026-08-22
stories 83

© 2026 RepoJournal Home Showcase Explore How it works Privacy

$ the-wire · showcase

Mantis proof agent gains programmable trust boundaries

By RepoJournal · Filed · About OpenClaw

Mantis Telegram Desktop proof agent is now programmable at its trust boundaries with declarative fault rules, scripted mock responses, and proxy-side request recording.

The Mantis Telegram Desktop proof agent [1] is now programmable at its trust boundaries, shipping declarative Bot API fault rules (drop/status per method), per-request scripted mock-provider responses, observe-until predicates (post-cursor events/text, cumulative provider count), and bounded proxy-side recording of outbound Bot API requests as trusted lane facts. It also adds a reusable recipe library under .github/codex/prompts/mantis-recipes. The release notes describe the feature as "programmable proof boundaries with recipe library" [1].

Crabbox v0.46.0 [2] adds fixed idempotent `--lease-id` replay to the Machine0 provider: an identical warmup adopts the existing VM instead of creating a second one, a drifted request fails with `lease_id_conflict`, and a released ID is single-use. It also rejects oversized delegated-run workspaces before any paid or stateful provider resource is created, so size-limit failures no longer leave billable resources behind, and fixes workspace sync for E2B and Azure Dynamic Sessions. Separate Linode work [3][4] fences metadata updates with exact revisioned claim snapshots, validating the account, scope, instance identity, and live tags while the claim lock fences both the provider write and claim commit.

In openclaw/openclaw, two fixes address stalls and lost data. Code Mode shell calls used to stall near the yield deadline [5]: both shell and Code Mode defaulted to a 10-second boundary, so the shell could return its process handle only as Code Mode exhausted its turn budget. Code Mode now supplies a bounded default `yieldMs` for the OpenClaw `exec` tool. Additionally, message-tool images in internal WebChat replies used to disappear after a Gateway restart [6] because durable media ownership was deferred until end-of-turn finalization; that ownership is now preserved so reconstructed chat history renders the image artifact.

Two locale refreshes [7][8] keep generated Control UI and native app locales synchronized through reviewable automation branches, with no direct user-facing change. The docs repo [9][10][11][12][13] received five mirror syncs from openclaw/openclaw commits, keeping documentation in step with source changes. In Crabbox, two timing-sensitive tests [14] were made robust on loaded CI runners: `TestReserveSSHLocalForwardPortIgnoresUDPUse` now retries across up to 32 candidate ports, and the opensandbox test `TestRunReconcilesAmbiguousCreateAfterDelayedVisibility` was adjusted similarly.

Action items

References

  1. [1] feat(mantis): programmable proof boundaries with recipe library (#127782) ↗ openclaw/openclaw
  2. [2] v0.46.0 ↗ openclaw/crabbox
  3. [3] fix(linode): fence metadata updates with exact claims ↗ openclaw/crabbox
  4. [4] fix(linode): fence metadata updates with exact claims (#1464) ↗ openclaw/crabbox
  5. [5] fix: Code Mode shell calls stall near yield deadline ↗ openclaw/openclaw
  6. [6] fix: preserve message-tool images across gateway restarts ↗ openclaw/openclaw
  7. [7] chore(ui): refresh control ui locales ↗ openclaw/openclaw
  8. [8] chore(i18n): refresh native locales ↗ openclaw/openclaw
  9. [9] chore(sync): mirror docs from openclaw/openclaw@48a2b45fede5d07405c718227f32754a6d8d8716 ↗ openclaw/docs
  10. [10] chore(sync): mirror docs from openclaw/openclaw@53fbe2eb33adea9eadfa51c01df2d88014990c81 ↗ openclaw/docs
  11. [11] chore(sync): mirror docs from openclaw/openclaw@67c06ed54f66487f108c393857e74ed15422d8b4 ↗ openclaw/docs
  12. [12] chore(sync): mirror docs from openclaw/openclaw@fc425351d6b334fc0cd23497e1d65c84def81be9 ↗ openclaw/docs
  13. [13] chore(sync): mirror docs from openclaw/openclaw@9acf1180a61a87ecd1cb68c4f384b586b4a05db0 ↗ openclaw/docs
  14. [14] test: make two timing-sensitive tests robust on loaded CI runners ↗ openclaw/crabbox

Quick answers

What shipped in OpenClaw on August 22, 2026?
Mantis Telegram Desktop proof agent is now programmable at its trust boundaries with declarative fault rules, scripted mock responses, and proxy-side request recording. In total, 52 commits, 30 pull requests, and 1 releases landed.
Who contributed to OpenClaw on August 22, 2026?
5 developers shipped this update, including Ayaan Zaidi, openclaw-mantis, clawsweeper, steipete, and openclaw-docs-sync[bot].
What were the notable OpenClaw updates?
feat(mantis): programmable proof boundaries with recipe library (#127782), v0.46.0, and fix(linode): fence metadata updates with exact claims.