OpenClaw Digest: Enhancing Agent Orchestration and Channel Reliability

Merged PRs

• fix(agents): add trajectory flush timeout diagnostics Original PR
• fix: recover stale subagent completion announces Original PR
• fix(agents): ignore duplicate embedded run clears Original PR
• fix(config): allow bundled provider timeout overlays Original PR
• Add Telegram progress preview flows Original PR
• fix: forward-port Tideclaw alpha release fixes Original PR
• fix: include gateway plugin commands in TUI autocomplete Original PR
• fix(whatsapp): clarify inbound group diagnostics Original PR
• fix(agents): preserve bare Telegram reply context Original PR
• fix: retry config snapshot after rejection Original PR
• Fix stuck Codex-native subagent tasks after blocked spawn Original PR
• fix(agents): skip dormant completion wake probes Original PR
• Guard final delivery session refresh Original PR
• fix(codex): preserve plugin tool auth profiles Original PR
• fix(codex): honor Docker bind write policy Original PR
• plugins: Allow trusted plugin keyed state Original PR
• fix(whatsapp): periodic delivery-queue drain so enqueued items don't wait for next reconnect Original PR
• docker: support optional pip packages in local builds Original PR
• Preserve queued Telegram topic followups Original PR
• fix(codex): guard against stale codex app snapshots leading to plugin invocation failure Original PR
• fix(migrate): use resolved provider for options Original PR
• fix(deepseek): normalize mcp union tool schemas Original PR
• Fix Telegram forum topic parallel flow Original PR
• Improve Telegram groups config shape diagnostics Original PR
• test(config): keep agent model overrides selection-only Original PR
• Fix memory plugin CLI help dispatch Original PR
• codex: surface deferred dynamic tool names Original PR
• fix(cli): include loopback tools in cli prompts Original PR
• fix(ui): keep delete confirm in viewport Original PR
• fix(memory-wiki): preserve fs-safe write diagnostics Original PR

Key Changes

Agent Orchestration and Subagent Delivery

Significant effort was directed toward the reliability of subagent completion announces. A critical fix was implemented to recover stale subagent completion announces by retrying requester wake and forcing a message-tool handoff when direct-agent replies are invisible. Additionally, the system now skips dormant completion wake probes, routing late announcements directly through the requester-agent handoff to reduce no_active_run log noise.

In the Codex extension, a fix ensures that Codex-native subagent tasks are marked terminal when a spawnAgent call fails or is blocked, preventing tasks from remaining stuck in a running state in the registry.

Telegram and Channel Enhancements

Telegram integration received substantial upgrades to support forum topics more effectively. The update fixes parallel flow issues by replacing global flush chains with per-topic queues and introducing a GroupFairQueue to prevent a single topic from monopolizing the shared group send pipe.

Other channel improvements include:

• Telegram Context: Preservation of bare mention reply context in the runtimeOnly prompt path, ensuring the model sees the replied-to message body.
• WhatsApp Reliability: Introduction of a periodic delivery-queue drain (every 30s) to prevent outbound messages from sitting in the queue until the next reconnect.
• WhatsApp Diagnostics: Clarified inbound group diagnostics to better distinguish between received and routed messages.

Codex and Sandbox Hardening

Security and policy enforcement in the Codex app-server were strengthened. The system now honors Docker bind write policies, ensuring that :rw bind mounts are included in writableRoots even when workspaceAccess is set to ro.

Further Codex improvements include:

• Guarding against stale app snapshots to prevent plugin invocation failures.
• Preserving plugin tool auth profiles.
• Surfacing deferred dynamic tool names in developer instructions.

Configuration and CLI Improvements

Configuration validation was refined to allow bundled provider timeout overlays (e.g., models.providers.openai.timeoutSeconds) without requiring a full provider redeclaration. CLI usability was improved by making root help config-sensitive, ensuring that the correct memory plugin commands (like ltm for LanceDB) are advertised based on the active configuration.

Impact

These changes collectively reduce operational noise and increase the robustness of complex agent workflows. The improvements to subagent delivery and Codex task reconciliation directly address high-severity issues where completed work failed to surface or tasks remained ghost-running in the registry.

For Telegram users, the transition to per-topic concurrency and fair-share queuing removes serialization bottlenecks, allowing multiple forum topics to operate in parallel without blocking one another. The WhatsApp periodic drain eliminates significant delivery delays, ensuring a more responsive user experience. Finally, the hardening of the Codex sandbox and configuration validator provides a more flexible yet secure environment for deploying custom model providers and bind-mounted data sources.