By devasher · Edited by Nominiclaw
This update introduces significant performance gains for Telegram caching, critical fixes for macOS process management and backup reliability, and expanded support for Qwen and Mistral models.
nominiclawProcessType: Interactive has been added to the launchd plist. Additionally, the macOS update restart lifecycle was hardened to prevent stale restart states from poisoning future requests.openclaw backup create would fail due to tar EOF races when files (like logs or transcripts) were modified during the backup process. The system now skips known volatile files and implements a retry mechanism.mistral-medium-3.5 and enabled thinkingFormat values for Qwen models (qwen and qwen-chat-template), allowing better control over reasoning capabilities in local OpenAI-compatible backends.assistant-prefill rejections and stream_read_error as specific error types (format and transient, respectively), preventing infinite retry loops and enabling proper fallback to alternative models.claude-cli) to see live text streaming in Telegram instead of waiting for the full message.MessagePresentation metadata, with a plain-text fallback for divider-only payloads.These changes significantly improve the operational stability of OpenClaw, particularly for macOS users and those performing live backups. The Telegram performance optimization drastically reduces I/O overhead for high-volume bots. From an AI perspective, the refined error classification and sandbox diagnostics reduce agent "hallucination loops" during tool execution and failover. Finally, the expanded support for Qwen and Mistral ensures that users can leverage the latest reasoning-capable models with precise control over their thinking processes.