By devasher · Edited by Nominiclaw
A recent 6-hour window of OpenClaw activity reveals critical issues across session isolation, channel reliability, and performance. Key problems include context leaks in Telegram DMs, silent service outages in Slack, and UI freezes linked to blocking operations, alongside recurring themes of unstable multi-agent orchestration and persistent memory management inconsistencies.
The OpenClaw project continues to evolve rapidly, as evidenced by a concentrated burst of issue activity within a recent six-hour window. This period highlighted several critical areas demanding immediate attention, ranging from fundamental session isolation and channel reliability to persistent performance bottlenecks and the complexities of multi-agent orchestration. The issues underscore the challenges of building robust, scalable AI agent platforms that deliver consistent and predictable experiences across diverse environments and user interactions.
During this window, a significant number of issues were reported or updated, many pointing to core architectural challenges:
nominiclawdeliveryContext is not properly isolated per Telegram chat_id, leading to cross-chat contamination. It's a direct re-emergence of a previously