DSH-Feedback-Bridge

When working with DeepSeek Harness plugins, the hardest part of reporting a missing feature, a bug, or a new idea is rarely spotting the problem itself. It is turning the scattered conversation context into a well-structured, fully contextualized community report. Manually stitching together messages, re-describing the scenario, then scrolling through GitHub Discussions to check whether someone already filed a similar issue is both time-consuming and prone to leaving out key details.

DSH Feedback Bridge compresses that workflow into a few steps. It groups feedback sources by complete exchange rounds (from your prompt through the full model reply, including tool results and errors), so you can filter, confirm, and quote them into specific fields. It calls the model already selected in the current conversation to help polish a draft, runs a similarity check against existing discussions, and auto-saves unfinished work so it survives a page refresh or a DSH restart. The finished text can be submitted directly to the official GitHub Discussions or exported as a Markdown file for manual publishing.

It is built for DSH Web users and developers who regularly accumulate rich conversation context and want to convert it into high-quality community posts without leaving the harness. Whether you are filing a feature request, documenting a reproducible error, or sharing a workflow suggestion, the plugin keeps the process fast, reviewable, and traceable.