This roundup covers OpenClaw issues closed between 2026-04-14T02:31:28Z and 2026-04-15T02:31:28Z. The biggest themes were channel reliability, provider compatibility, and workflow-state correctness.
Quick answer
Eight issues were resolved in the last 24 hours. The most important fixes addressed missed BlueBubbles messages after gateway restarts, Telegram regressions, installer crashes, and 400-level model/provider compatibility failures.
Channel and messaging fixes
#66721 — BlueBubbles: replay missed webhook messages after gateway restart (cursor + fetchBlueBubblesHistory + processMessage)
When the OpenClaw gateway is down, wedging, or restarting, inbound BlueBubbles messages delivered during the outage window are permanently lost. The underlying iMessages are intact (they remain in Messages.app and in BB Server's DB), but the agent never sees them, never replies, and no recovery happens when the gateway
Labels: maintainer. Closed at: 2026-04-15T02:20:44Z. Source: https://github.com/openclaw/openclaw/issues/66721
#66756 — Regression: Telegram group slash commands stop working on 2026.4.14 (2026.4.12 works)
After upgrading OpenClaw from 2026.4.12 to 2026.4.14, Telegram slash commands stopped working for one bot/account (`morty`) in a Telegram group. Rolling back restored the behavior.
Labels: no labels. Closed at: 2026-04-15T02:16:37Z. Source: https://github.com/openclaw/openclaw/issues/66756
#66914 — DevClaw can silently fork project registry state by creating an empty devclaw/projects.json
DevClaw startup/workspace scaffolding can silently create a fresh empty `devclaw/projects.json` when that file is missing.
Labels: no labels. Closed at: 2026-04-15T01:38:45Z. Source: https://github.com/openclaw/openclaw/issues/66914
#63500 — [Bug]: HTTP 400 when calling gpt-5.2-codex / gpt-5.3-chat via OpenClaw (Azure OpenAI Responses adapter / reasoning item issue)
Behavior bug (incorrect output/state without crash)
Labels: bug, bug:behavior. Closed at: 2026-04-15T01:04:20Z. Source: https://github.com/openclaw/openclaw/issues/63500
#66641 — [Bug]: Installer always crashes after “Select channel (QuickStart)” with TypeError: Cannot read properties of undefined (reading 'trim')
Crash (process/app exits or hangs)
Labels: bug. Closed at: 2026-04-14T16:11:49Z. Source: https://github.com/openclaw/openclaw/issues/66641
Provider and model compatibility fixes
#53506 — [Bug]: API Protocol Error 400 with Azure Reasoning Models in v2026.3.23 (works in v2026.3.18)
In OpenClaw v2026.3.23, using Azure OpenAI reasoning models (e.g., gpt-5.2) with "api": "openai-responses" triggers a 400 protocol error regarding missing items following a reasoning block. This configuration followed the [official Microsoft Foundry integration guide](https://techcommunity.microsoft.com/blog/educatorde
Labels: enhancement. Closed at: 2026-04-15T01:05:42Z. Source: https://github.com/openclaw/openclaw/issues/53506
#57033 — [Bug]: Native Google provider sends OpenAI-specific `store` field, causing 400 from Gemini API
Behavior bug (incorrect output/state without crash)
Labels: bug, bug:behavior. Closed at: 2026-04-15T00:43:34Z. Source: https://github.com/openclaw/openclaw/issues/57033
Workflow and state-management fixes
#66336 — [Bug]: sessions_spawn fails with 'pairing required (1008)' when spawning isolated logic agents (v2026.4.12)
When attempting to use `sessions_spawn` to orchestrate multi-agent workflows (spawning isolated logic sub-agents) in version `v2026.4.12`, the command fails consistently with the WebSocket 1008 `pairing required` error.
Labels: no labels. Closed at: 2026-04-15T01:08:56Z. Source: https://github.com/openclaw/openclaw/issues/66336
Notable merged commits in the same window
- feat(bluebubbles): replay missed webhook messages after gateway restart (#66857)
- fix: restore Telegram native auto defaults (#66843) (thanks @kashevk0)
- test(e2e): harden Parallels smoke probes
- test(e2e): quote linux bad-plugin diagnostic grep
- QA: split Matrix scenario leaf types
Why this matters for operators
These fixes reduce the risk of silent failures in real OpenClaw deployments. They are especially relevant for teams running messaging channels, using Azure or Google model providers, or relying on multi-agent and project-registry behavior in day-to-day operations.