Ghenghis

Enables safe command execution for OpenClaw Agents with risk assessment, user approval, and audit logging for secure operations.

Monitors static layer file hashes to detect prompt cache invalidation, providing alerts on changes during session starts.

Monitors and manages auto-compaction failures in memory-lancedb-pro, triggering alerts via Telegram for manual intervention.

Automatically evaluates session outcomes by checking for errors and logs reflections to LanceDB, enhancing AI learning.

This skill automates compression strategy selection based on session context, optimizing data retention for various scenarios.

Enhances tool call safety with AI-driven risk classification, preventing harmful operations before execution.

Integrates OpenClaw with Claude Code to manage memory and prevent API call wastage through a robust circuit breaker mechanism.

Facilitates the extraction and integration of high-frequency memories into permanent behavior rules for AI agents using OpenClaw.

Automates memory management for LanceDB in OpenClaw, optimizing retrieval quality by cleaning and merging memory fragments weekly.