dsh-user-experience: A UX Walkthrough Plugin for DeepSeek Harness
dietcokewithsugar/dsh-user-experience
DSH plugin for persona-driven UX walkthrough, scanning frontend code to pinpoint issues and suggest fixes.
Have AI simulate target users to uncover UX problems during development and provide concrete suggestions.
dsh plugin --profile web add github:dietcokewithsugar/dsh-user-experience- Category
- Development & Operations
- Platform
- DSH-Plugin
- Author
- dietcokewithsugar
- Distribution
- Plugin
dsh-user-experience Key Features
dsh-user-experience Repository
dsh-user-experience Repo Summary
dsh-user-experience is a UX walkthrough plugin for DeepSeek Harness (DSH), maintained by DietCokewithSugar under the MIT license, last updated in 2026-08. It addresses the core problem that existing automated checks (axe, Lighthouse) can only verify absolute rules—contrast ratio, missing alt text—while UX issues are inherently relative: a confirmation dialog before deleting protects an occasional user but wastes the time of an operator who processes hundreds of records a day. This plugin makes target user personas a prerequisite for the walkthrough: every finding is anchored to an explicit persona, and no persona means no conclusions. By having AI walk through the product as those users, it surfaces experience problems during development and gives concrete, locatable, reviewable optimization suggestions—not post-launch user feedback. It is a pipeline, not a CLI: edit a front-end file and the walkthrough runs itself. The report card leads with plain language (which page, what happened, how bad), and folds file paths and rule IDs into a 'technical details' block you can copy straight to an AI in one click. It supports React + TypeScript / React + JavaScript / Vue 3, CSS/layout analysis, and optional browser evidence when the current Harness session can open the application.
What inputs and parsing engines does the dsh-user-experience DSH plugin support?
The plugin supports React + TypeScript (.ts/.tsx) and React + JavaScript (.js/.jsx), both parsed using the TypeScript compiler API (TSX). Vue 3 (.vue SFC) uses @vue/compiler-sfc for block splitting and @vue/compiler-dom for template AST; script blocks reuse the TypeScript engine, with line numbers remapped to the whole .vue file. CSS/SCSS/Sass/Less/PostCSS are analyzed conservatively for spacing, compact-layout, and decorative-content candidates, but visual conclusions still require a rendered page. When browser/screenshot tools and a runnable app are available, the agent inspects relevant routes and viewports, providing multi-level evidence (rendered and static) for review.
How do I install dsh-user-experience in DeepSeek Harness?
In DeepSeek Harness, enter the web profile and run the following command:
dsh plugin --profile web add github:dietcokewithsugar/dsh-user-experience
After installation, restart DSH or reload the web profile. GitHub plugins execute build scripts during installation; read the security note before installing, and pin a trusted commit for production use.
How does the dsh-user-experience walkthrough report present and interact?
The walkthrough report explains observed behavior and user impact in plain language. Once you confirm a finding, the card provides a task Prompt you can copy to another AI; it describes the observed phenomenon rather than prescribing code changes, and explicitly allows copy changes. Multi-level evidence appears in the technical details so every visual or interaction conclusion can be reviewed. The report card and confirmation workflow can use the developer's language (interface preview with sample data).
What is the core philosophy of dsh-user-experience?
The plugin makes target user personas a prerequisite for the walkthrough: every finding is anchored to an explicit persona, and no persona means no conclusions. By having AI walk through the product as those users, it surfaces experience problems during development, not post-launch user feedback. It is listed in awesome-dsh-plugin.