DSH Session Lab: Session Evidence Packaging, Skill Assessment, and Trajectory Comparison

DSH Session Lab is a community plugin maintained by developer zhangguiping-xydt, providing standardized workflows for processing session evidence and skills in DeepSeek Harness. Its core features include packaging session exports into portable evidence bundles and verifying their integrity; extracting and evaluating reusable skills from successful sessions; and performing controlled trajectory comparisons. The plugin supports CLI installation and requires a Node.js and Python environment. Note that it is not an official product, the raw data contains sensitive content, and it does not support Host session import or exact model replay. It is primarily used for evidence verification, skill extraction, and differential analysis.