OpenClaw Issue Digest: Critical Regressions in Codex and Discord Integration
Open Issues
Recent activity in the OpenClaw repository highlights several critical stability issues, particularly affecting the Codex embedded runner and various channel integrations.
High-Severity Regressions
- Codex Embedded Tool Timeouts (#82129): A significant regression in version 2026.5.12 causes Codex embedded agent turns to time out specifically when using tool or status checks, while text-only runs continue to succeed. This effectively blocks diagnostic and tool-based workflows.
- Discord Delivery Suppression (#82134): Users are reporting that answers generated for Discord server channels are blocked by a
Delivery suppressed by sourceReplyDeliveryMode: message_tool_onlyerror. While the WebUI shows the answer, it never reaches Discord, rendering the bot unresponsive in guild channels. - WebChat Node Exec Failures (#82132): A regression introduced in 2026.5.5 causes node execution approvals from WebChat to fail with
APPROVAL_CLIENT_MISMATCH. This is a high-severity issue as it completely blocks node commands from the WebUI.
Data Loss and Stability
- Cron Job Data Loss (#60799): A critical regression in v2026.4.2 causes the gateway to silently drop
jobs.jsonfiles in the plain-array format. The first subsequentcron addoperation clobbers all existing jobs, leading to irreversible data loss for users upgrading from 2026.3.x. - Session Lane Stalls (#59878): Session lanes are becoming permanently stuck in a
runningstate even after the LLM run has died. This blocks all subsequent messages in that session, requiring manual deletion of session registry entries or a/newcommand to recover.
Integration and Provider Bugs
- Xiaomi MiMo Reasoning Issues (#81419): Multi-turn tool calls with Xiaomi MiMo models fail with a 400 error because OpenClaw fails to pass the required
reasoning_contentfield on assistant messages. - Google Chat Space Failures (#58514): While DMs work, messages in Google Chat Spaces/Groups are silently ignored, blocking team collaboration.
- iMessage LaunchAgent Hangs (#81940): On macOS, the iMessage provider hangs during the
imsg rpcreadiness check when started via LaunchAgent, though it works in interactive sessions.
Key Themes
1. Persistence and State Corruption
There is a recurring theme of state diverging between the session registry and the actual transcripts. Issues like #60542 and #59878 suggest that the gateway's method of persisting session states is prone to becoming "wedged," where the registry thinks a session is terminal or running while the transcript suggests otherwise.
2. Tool-Call and Execution Fragility
Several issues point to a breakdown in the tool-execution lifecycle. From the APPROVAL_CLIENT_MISMATCH in WebChat (#82132) to the Codex tool timeouts (#82129), the bridge between the agent's intent to use a tool and the actual execution on the node/provider is currently unstable.
3. Context and Token Management
Users are requesting more granular control over context injection. A common theme is the waste of tokens caused by injecting full bootstrap files (AGENTS.md, etc.) into every turn. Proposals for selective injection (#61001) and query-conditioned compression (#60509) highlight a need for a more efficient context-loading strategy.
Action Required
Immediate Attention Required
- #82129 (Codex Tool Timeouts): This is a primary blocker for users relying on the Codex embedded harness for technical tasks.
- #82132 (WebChat Approval Mismatch): High-severity regression blocking all node exec from the primary UI.
- #60799 (Cron Data Loss): Immediate fix needed for the cron loader to prevent further irreversible data loss during upgrades.
Blocked or High-Risk
- #82134 (Discord Suppression): Blocks the bot's ability to communicate in Discord guilds.
- #59878 (Stale Session Lanes): Causes permanent silence for affected users until manual intervention occurs.
- #81940 (iMessage LaunchAgent): Renders the iMessage channel unusable for users utilizing the default macOS installation method.