jinhuang712/dsh-survey
DeepSeek Harness 的问卷式批量提问插件:一次可提10个以上问题(单选/多选/是否切换/对比/开放式),可跳过单个问题,全屏覆盖,提交后双栏回顾。
Project Overview项目介绍
This is a native DSH tool plugin built exclusively for DeepSeek Harness, that adds the do_a_survey tool to your agent environment. It lets AI agents send an entire questionnaire to users in one go, instead of asking one question at a time across multiple round trips. To install, you can run the dsh plugin command pointing to the GitHub release tag, the main branch, or a local cloned directory. No compilation is required because build artifacts are committed to the repo, and you just need to restart your DSH web profile and refresh the page to activate the tool.
This plugin is designed for scenarios where agents need to collect multiple user inputs before starting work, such as confirming project scope, selecting configuration options, or gathering user preferences. It supports four presentation modes to fit different use cases: compact for single questions, inline for multiple questions embedded in the conversation thread, overlay for comparison questions, and grid for many simple yes/no or single-choice questions. All answers are returned as structured data, so agents do not need to parse natural language responses to get the correct selection. Each question also includes a skip option for users that do not want to answer it.
The plugin uses the existing DSH web client’s React instance and UI primitives, so it does not bundle its own copy of React, keeping the plugin size small. It supports both English and Chinese interface languages, following the DSH web UI’s language setting automatically. It also ships with a companion skill that includes a usage guide for agents and a fallback recipe for environments that do not support bundled plugins. It is released under the permissive MIT license, so you can modify and redistribute it freely. To uninstall, you just remove the entry from your profile’s configuration files and run pnpm remove to clean up dependencies.
这是一个面向 DeepSeek Harness 开发的原生工具插件,提供 do_a_survey 工具,允许 AI 代理一次性发送完整问卷,而不是逐次提问引发多轮往返。用户可以在同一个 UI 卡片中回答所有问题并一次性提交,所有回答以结构化格式返回给代理,大幅减少不必要的交互,提升了任务启动效率。
该插件适合需要用户提前确认多个决策点才能启动任务的场景,比如项目范围确认、配置选项选择等。它支持四种展示模式:紧凑单题、嵌入对话、全屏对比、网格矩阵,可适配单题、多题、对比题、大量简单选择题等不同场景,每个问题都支持跳过选项。
插件遵循 MIT 许可协议发布,可通过 DSH 命令行直接安装,不需要本地编译,已将构建产物提交到仓库,支持从 GitHub 版本标签、主分支或本地检出目录三种安装方式。安装后永久生效,附带使用说明技能文件,方便代理查阅调用规范。
请帮我安装这个 DSH 插件。安装前先完成【兼容性检查 + 安全性检查】,检查通过再动手。
插件:dsh-survey(jinhuang712/dsh-survey)
仓库:https://github.com/jinhuang712/dsh-survey
本站详情页:https://www.yhbd.top/plugins/jinhuang712-dsh-survey/
本站登记:类型 plugin · 归类 原生 DSH 插件 · 许可证 MIT · ⭐ 4 · 最近提交 2026-09-26 · 主语言 JavaScript
按下面顺序执行,每步先把结论告诉我,再进入下一步:
【1 兼容性检查】
① 我这边:DSH 版本、Node 版本、操作系统、当前 profile(web / desktop)。
② 读它的 README、package.json、插件 manifest,列出它要求的 DSH 版本 / Node 版本 / 操作系统 / 外部依赖 / 需要另外先装的运行时。
③ 逐条比对,结论只写「满足 / 不满足 / 未知」三种;不满足的给出可行替代方案。
④ 检查是否和我已装的插件冲突:命令名重复、skill / tool 重名、端口占用、重复注册的 MCP server。
【2 安全性检查】
① 仓库可信度:和上面「本站登记」是否一致;star / fork 数、创建时间、最近提交,是否归档或长期停更。
② 安装脚本:逐行看 package.json 的 preinstall / install / postinstall,以及 install.sh、setup.ps1 之类脚本。出现 curl|bash、下载后直接执行、混淆代码、访问与插件功能无关的域名,立刻停下来告诉我,不要继续装。
③ 依赖:列出新增依赖,标出无人维护、或与知名包拼写近似的可疑包(typosquatting)。
④ 权限与副作用:它会读写哪些目录、访问哪些域名、需要哪些 DSH 权限(filesystem / network / shell / clipboard 等),以及怎么卸载和回滚。
⑤ 如果它要求 sudo / 管理员权限,或权限明显超出功能所需,先停下来问我。
【3 安装】
上面两步没有「不满足」和「高危项」时才执行;用官方推荐方式安装,不要自行提权。
【4 汇报】
用表格输出:检查项 / 结论 / 依据 / 是否需要我决策。拿不准的一律写「未知」并说明要我怎么确认——不要猜,也不要替我决定。
Send this message to DSH in your current session: it verifies compatibility and security first (answering met / not met / unknown item by item) and only installs once everything checks out — it will stop and ask you if it finds a high-risk item. The box scrolls; the copy is the full prompt. CLI install commands may not be accurate across systems, so DSH is the safer route.把上面这条消息直接发给当前会话里的 DSH:它会先核对兼容性与安全性(逐条给「满足 / 不满足 / 未知」),确认没问题再安装,有高危项会停下来问你。框内可滚动,复制到的是完整提示词;安装命令不一定准确,发给 DSH 更稳。
- Only 4 stars - very few users, little community feedback星标只有 4,几乎没人在用,遇到问题缺少社区反馈
DSH walks through these 9 checksDSH 会逐条核对这 9 项
Compatibility兼容性
- DSH, Node, OS and profile requirementsDSH 版本 / Node 版本 / 操作系统 / profile 是否满足要求
- External dependencies and runtimes (Electron / Python / Docker, ...)外部依赖与运行时(Electron / Python / Docker 等)是否齐备
- Conflicts with installed plugins: command names, skill / tool names, ports, duplicate MCP registration与已装插件是否冲突:命令名、skill / tool 重名、端口占用、重复 MCP 注册
Security安全性
- Repo matches the facts registered here; archived or abandoned?仓库是否与页面登记一致,是否归档或长期停更
- Safety of preinstall / install / postinstall and install.sh / setup.ps1preinstall / install / postinstall 与 install.sh、setup.ps1 是否安全
- curl|bash, download-then-execute, obfuscation, unrelated domains → stop immediatelycurl|bash、下载即执行、混淆代码、无关域名 → 立刻停止
- Typosquatting or unmaintained packages among the new dependencies新增依赖里有没有 typosquatting 或无人维护的包
- Requested permissions vs. what the feature actually needs申请了哪些权限、是否超出功能所需(filesystem / network / shell / clipboard)
- Any sudo / admin requirement, plus uninstall and rollback是否要求 sudo / 管理员权限,以及卸载与回滚方式
Anything uncertain must be marked unknown with a note on how to confirm it. This site's signal screen is a static snapshot, not a security audit.拿不准的必须标「未知」并说明要我怎么确认。本站的信号筛查是静态快照,不能替代安全审计。
Or use CLI install (for developers)或使用命令行安装(适合开发者)
CLI Install命令行安装
dsh plugin --profile web add "github:jinhuang712/dsh-survey#v1.1.0"
把 jinhuang712/dsh-survey 加入你的 DSH 配置(web profile)即可启用。
READMEREADME
dsh-survey
Ask the user ten questions at once instead of ten times in a row.
do_a_survey is a tool plugin for DeepSeek Harness. The model sends a whole questionnaire; the user answers it in one card and submits once.
| Single | Multi | Yes/No | Compare | Open |
|---|---|---|---|---|
| Numbered rows | Checkbox | Radio pair | Side-by-side blocks | Multi-line input |
When you'd want this
- The agent needs five decisions before it can start. Without this it asks one, waits, asks the next — five round trips before any work happens. Here it asks once and you answer the lot in a single card.
- You want the answers back as data, not prose. Each reply arrives as
{ id, selected, custom?, skipped? }, so the model never has to parse "yeah the first one, and skip the last question" out of a sentence. - Some of it you genuinely don't care about. Every question has a skip, so a survey that asks more than you want to answer costs you a click, not a negotiation.
When the model calls do_a_survey, the Web UI renders the survey by mode:
compact— single-question cardinline— embedded in the conversation columnoverlay— fullscreen, for compare questionsgrid— matrix of many simple questions
All text supports Markdown (code blocks, blockquotes, inline code, bold) and color ({color:red}text{/color}). A readable two-column recap follows the submit.
Preview
Screenshots of the real toolview, captured from the shipped bundle.
Grid matrix — fullscreen overlay for many simple questions, one card each:

Inline — survey embedded in the conversation column, filled and submitted together:

Overlay compare — fullscreen side-by-side comparison with Markdown:
Showing the opening section of the README — the full document lives in the repository以上为 README 开头摘要,完整文档在仓库内 · View the full README on GitHub →在 GitHub 查看完整 README →
GanyuanRan/Aegis
shanliuling/skills-link
Jayden-X-L/forkprobe
Soulize/hard-flash
dushaobindoudou/dsh-acp
MichengAI/dsh-pua