dsh-live-stream

dsh-live-stream tackles the efficiency gap in the data-decomposition step of live-stream e-commerce workflows. After every broadcast, operators and analysts must manually cross-calculate metrics such as GMV, average watch time, conversion rate, and average order value. Parameter-passing errors are common, and the process is difficult to reuse across sessions. This plugin packages the entire breakdown into a single structured workflow tool: one call inside an Agent session finishes the calculation and returns a cleanly structured result that slots directly into a larger analytics pipeline.

On the capability side, the plugin registers a tool named live_stream. It accepts user-supplied parameters, performs precise calculations on the live-stream data, and returns a detailed, structured result set. The implementation is written in pure Node.js and makes zero outbound network calls, so it runs reliably in intranet, offline, or sandboxed environments. There is no risk of an unreachable API stalling the pipeline, and no additional runtime dependencies are introduced.

Who it is for: e-commerce operations teams that want to fold live-stream metrics into an automated analytics flow; developers who treat post-broadcast breakdown as a standard step in an Agent workflow; and any team with a hard requirement that data never leaves the local environment and who prefer a zero-dependency approach.