Introduction to the DeepSeek Harness Plugin: dsh-plugin-asmemory

`dsh-plugin-asmemory` is an Action-State Memory Engine plugin for DeepSeek Harness (DSH), designed to address the limitation that generic memory cannot answer what happened and why. It enables agents to log typed events (states and actions), provides trend, anomaly, and causal analysis capabilities, and supports exporting data to DataLens. Installation and activation are straightforward, relying on the Python 3.10+ standard library, and the plugin provides agents with a memory layer for the physical and operational worlds through seven MCP tools, helping them more accurately understand their own behavior and environmental changes.