Ai

Mem0 OpenClaw Plugin v1.0.11 Expands Memory Automation and Gateway Integration

Mem0 OpenClaw Plugin v1.0.11 Expands Memory Automation and Gateway Integration

Mem0 OpenClaw Plugin v1.0.11 introduces a focused update aimed at improving how memory works inside OpenClaw. This release adds automatic skills-mode configuration after onboarding, exposes new runtime memory capability methods for gateway access, improves collection handling when embedding dimensions change, and broadens default recall behavior. The result is a more reliable memory workflow with less manual setup and better support for ongoing agent operations.

What Changed

One of the biggest additions in v1.0.11 is skills-mode auto-setup. The plugin now runs enableSkillsConfig() automatically after onboarding, which enables triage, recall with reranking and keyword search, and dream consolidation. It also sets tools.profile = "full" and disables the built-in session-memory hook to avoid conflicts with the skills-based workflow.

The plugin also adds a new memory runtime capability. It now exposes runtime.getMemorySearchManager() and resolveMemoryBackendConfig() through the registered memory capability, giving the OpenClaw gateway direct access to memory status and backend configuration.

Another practical change is dimension-aware collections. The OSS setup wizard can now detect when embedder dimensions have changed and automatically create a new collection such as mem0_1536d. The release also warns users that older memories may become inaccessible under the new embedder configuration, which helps avoid silent retrieval issues.

Documentation has also improved. The memory-triage and memory-dream skill files now include full tool reference sections, making it easier for operators to understand available tools and parameters without jumping between sources.

Why It Matters

This version is important because it reduces friction in setting up and running memory-heavy OpenClaw deployments. Automatic skills-mode setup means teams can move faster after onboarding and avoid misconfiguration between session memory and skill-driven memory flows.

The new runtime capability exposure is also significant for platform reliability. By letting the gateway inspect memory status and backend configuration directly, v1.0.11 improves observability and makes it easier to support diagnostics, automation, and operational health checks.

The dimension-aware collection update addresses a subtle but serious issue for vector-based systems. Embedding model changes can silently break retrieval if collection dimensions no longer match. Automating new collection creation makes that transition safer, while the warning about old memory access helps teams plan migrations more deliberately.

Finally, changes to defaults and memory editing behavior support more usable recall out of the box. Auto-capture and auto-recall now default to enabled, the search threshold has been lowered from 0.5 to 0.1 for broader retrieval, and skills now prefer memory_update instead of delete-and-add for in-place edits. Together, these changes make the memory layer more aggressive in finding context while preserving cleaner edit history and more atomic updates.

Official Source: https://github.com/mem0ai/mem0/releases/tag/openclaw-v1.0.11