dsh-web
zhu1090093659/dsh-webDeepSeek Harness(DSH)Web 插件聚合生态包 · 一切皆插件,创意工坊分发(dsh-market.com)
Questionnaire-style batch questioning plugin for DeepSeek Harness: 10+ questions at once (single/multi/yes-no toggle/compare/open), per-question skip, fullscreen overlay, two-column recap after submit
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 |
{ id, selected, custom?, skipped? }, so the model never has to parse "yeah
the first one, and skip the last question" out of a sentence.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 questionsAll text supports Markdown (code blocks, blockquotes, inline code, bold) and color ({color:red}text{/color}). A readable two-column recap follows the submit.
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:
Compact — single-question card with rich question types:
From npm (the scoped name; the unscoped dsh-survey on the registry is not
this package):
dsh plugin --profile web add @jinhuang712/dsh-survey
# restart dsh web, then refresh the page
Or pinned to a release — build artifacts are committed, so there is nothing to build:
dsh plugin --profile web add "github:jinhuang712/dsh-survey#v1.1.4"
# restart dsh web, then refresh the page
Or track main to pick up unreleased commits:
dsh plugin --profile web add "github:jinhuang712/dsh-survey#main"
Or from a local checkout, linked so edits show up on refresh:
git clone https://github.com/jinhuang712/dsh-survey.git
dsh plugin --profile web add "link:$PWD/dsh-survey"
If your dsh profile directory is a pnpm workspace, pnpm asks for -w before
touching its root — pass it through: dsh plugin --profile web add -w ….
After install, the do_a_survey tool and its survey UI are available permanently.
The companion skill dsh-survey registers with the install (dsh.skills declaration):
references/dynamic-plugin-fallback.md) for environments without the bundleTell the model what you want to collect; it calls do_a_survey(mode, questions). mode is required:
| mode | When | Presentation |
|---|---|---|
"compact" | Exactly 1 question | Compact single-question card |
"inline" | Multiple questions, no compare | Embedded in the conversation column |
"overlay" | Compare questions or wide canvas | Fullscreen overlay (1180px) |
"grid" | Many simple questions (yes/no, single) | Fullscreen grid matrix, one card each |
Example:
{
"mode": "inline",
"questions": [
{
"id": "q1",
"header": "Runtime",
"question": "Which runtime does this ship on?",
"options": [
{ "label": "Node 20 LTS (recommended)", "description": "Matches the CI base image." },
{ "label": "Bun 1.1", "description": "Faster boot, fewer native addons." }
]
},
{
"id": "q2",
"header": "Scope",
"question": "Which surfaces need the rewrite?",
"multi_select": true,
"options": [{ "label": "Web client" }, { "label": "CLI" }]
},
{ "id": "q3", "kind": "boolean", "question": "Ship behind a flag first?" },
{ "id": "q4", "header": "Notes", "question": "Anything the changelog should call out?" }
]
}
header is an optional short heading shown beside the question number. A trailing (recommended) on a label becomes a badge.
| Type | Trigger | UI | Answer |
|---|---|---|---|
| Single | options + no multi_select | Numbered rows (1 / 2 / 3) | The chosen option's label |
| Multi | options + multi_select: true | Checkbox | Every chosen label, in option order |
| Yes/No | kind: "boolean" | Radio pair; segmented toggle in grid (omit options) | "yes" or "no" |
| Compare | kind: "compare" + compare: {left: {title,text}, right: {title,text}} | Side-by-side blocks (overlay recommended) | "left" or "right" |
| Open | no options, not boolean/compare | Multi-line input | Empty selected, body in custom |
Each answer is { id, selected, custom?, skipped? }. Question ids must be unique within one survey, and a survey nobody submits within 30 minutes fails on timeout rather than hanging the call.
{color:red}text{/color} (named / #hex / rgb()), usable in questions, options, compare blocksskipped: truemode: "overlay" centers fullscreen (mask + 1180px), breaking past the 748px conversation columnmode: "grid" fullscreen grid of many simple questions
lib/index.mjs): Cordis entry
defineTool registers do_a_survey with a 30-minute timeoutMswebServer.register serves /api/dsh-survey/submit|cancelexecute suspends on the answer, correlated by exec.callId; submit, cancel, abort, timeout and unload each release itsrc/ → lib/client.js): __ModuleLoader__.load bundle registering tool.call.toolview key=do_a_survey
runtime.js binds the host's React and UI primitives from the loader's require — they are never bundled, so the plugin shares the host's React instancestyles.css the stylesheet, injected once per pagei18n.js zh/en copy, markdown.js Markdown and colour, answers.js pick ↔ answer mappingcontrols.js the answer controls, model.js draft state and submit/cancelmodes/ one file per presentation: compact, survey (inline + overlay), grid, recapskills/dsh-survey/SKILL.md): usage guide + dynamic-plugin fallback recipe (references/dynamic-plugin-fallback.md)lib/client.js and its source map are build output, committed so the GitHub install stays one line. Edit src/, then:
pnpm install
pnpm build # esbuild src/index.js -> lib/client.js + lib/client.js.map
The host serves the map at /plugins/dsh-survey/client.js.map, so breakpoints land in src/.
It loaded at all:
__DSH_BOOT__ includes the dsh-survey client row, and
/plugins/dsh-survey/client.js returns 200cordis_inspect_query (Tool.listTools) lists do_a_surveyThen run a survey covering all four modes and check what the card does:
| Look at | Expect |
|---|---|
| a single-choice question | numbered seats 1 2, not radio dots — dots appear only on yes/no |
mode: "grid" | every card the same size, skip ✕ in each corner, a fenced code block flattened to inline rather than inflating the card |
| a compare question | both sides carry a panel; the chosen one lifts with a brighter border and an inverted number |
a (recommended) option | renders as a badge, and the marker is gone from the label — including in the recap after submit |
Markdown and {color:…} | bold, inline code and fenced blocks render; a coloured run stays on the same line as the text around it |
| the interface language | follows the Web UI's language setting, and switching it re-renders the card immediately |
dsh-survey insert row from the web profile's cordis.patch.ymldsh-survey dependency from the web profile's dsh.profile.bundles and run pnpm removeMIT