dashboard-data

Build truthful self-contained dashboard snapshots from connected MCP/data tools or user-provided datasets. Use for /dashboard runs that query real metrics, render charts or operational summaries, and may be refreshed by scheduled jobs.

Ir para a instalação

Repositório: juspay/xyne-spaces
Última atividade na origem: 17 de agosto de 2026 às 18:46
Idioma detectado do SKILL.md: inglês
Estrelas: 143
Forks: 17

Opções de instalação

Por padrão, está selecionado o prompt que primeiro revisa a origem. Você pode mudar para um comando direto ou baixar uma cópia local.

Instalar com Codex ou Claude Copie este prompt, cole no Codex, Claude ou outro assistente e deixe que ele revise a página da skill e instale para você.

Um comando direto ignora o prompt de revisão. Verifique a origem antes de executá-lo.

npx skills add https://github.com/juspay/xyne-spaces --skill dashboard-data

O comando permanece em uma só linha. Role horizontalmente para revisá-lo antes de copiar.

Prefere uma cópia local? Baixe os arquivos disponíveis atualmente no SkillsMP.

Revise os arquivos de origem

Leia o SKILL.md e os arquivos complementares exibidos pelo SkillsMP antes de decidir se vai instalar.

SKILL.md

Instruções da origem · Visualização somente leitura

Voltar à pesquisa Voltar à instalação

Dashboard data

Follow this order; data retrieval is a hard gate before visual composition.

Identify the requested measures, dimensions, time window, timezone, and comparison period.

Inspect available read-only data tools. Query the authoritative source and retain the exact query/window needed to explain the snapshot.

If no suitable tool or supplied dataset is available, stop. Name the missing connection. Never invent placeholder values, convert guesses into metrics, or deliver a sample dashboard.

Check basic integrity: units, aggregation level, missing intervals, duplicate series, timezone, and whether counters require rate/increase rather than raw values.

Build a self-contained HTML snapshot with no runtime fetches, tokens, cookies, internal URLs, or embedded credentials.

Show Data as of <timestamp with timezone>, the reporting window, and a human-readable source label. Do not expose secret query parameters.

Use stable card order, chart domains, colors, labels, and DOM structure across refreshes. A scheduled refresh should change the data, not randomly redesign the dashboard.

Present unavailable metrics as unavailable; do not coerce missing data to zero.

QA desktop and mobile layouts, chart legibility, empty/error states, console output, and network requests before delivery.

If the user explicitly asks for a recurring refresh, schedule the same complete /dashboard brief only after the first verified snapshot succeeds. Preserve the original conversation target. Never schedule implicitly, and never let a scheduled dashboard schedule another run.

Prefer KPI cards only for decision-driving headline values. Use lines for trends, bars for comparisons, tables for exact lookup, and annotations for threshold crossings. Include a concise textual takeaway adjacent to each complex chart.

Treat the artifact as a verified snapshot. The delivered HTML must remain useful offline and must never contact the underlying data source when opened or shared.