How do I install the Pulsebot App Builder skill?

Run npx skillmds@latest add timeplus-io/pulsebot-app-builder in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

What does the Pulsebot App Builder skill do?

Build real-time Timeplus data processing and analysis applications in Pulsebot. Creates pure frontend HTML/JavaScript apps that connect directly to Timeplus Proton via @timeplus/proton-javascript-driver (UMD), visualize live streaming data with @timeplus/vistral (UMD), and follow the Timeplus UI style guide. No npm build or bundler required — output is a single self-contained HTML file. It is listed under Web & Frontend on SkillMD.

Is Pulsebot App Builder safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Pulsebot App Builder?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Pulsebot App Builder free to use?

Yes. Installing skills from SkillMD is free. This skill is licensed under Apache-2.

Who published Pulsebot App Builder?

timeplus-io (@timeplus-io) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Pulsebot App Builder by timeplus-io

Name: Pulsebot App Builder
Author: timeplus-io

Pulsebot App Builder

Use this skill whenever the user asks to build a data processing application, pipeline visualizer, real-time dashboard, streaming analytics app, or any frontend tool that queries or visualizes data from Timeplus Proton.

Overview

You will produce a single self-contained HTML file that:

Queries Timeplus Proton directly via @timeplus/proton-javascript-driver loaded from unpkg CDN
Visualizes streaming data using @timeplus/vistral loaded from unpkg CDN
Follows the Timeplus App Style Guide (dark theme, brand colors, clean layout)
Requires no npm install, no build step — open the HTML file directly in a browser

Key architecture: Both @timeplus/proton-javascript-driver and @timeplus/vistral ship UMD builds. Import them via <script src="https://unpkg.com/..."> tags. The Proton driver connects directly to the agent proxy running at localhost:8001.

Step-by-Step Workflow

Step 1 — Clarify Requirements

Before writing any code, confirm:

What stream(s) or table(s) to query (name, schema if known)
What kind of visualization: time series, bar/column, table, single value, or multi-panel
Whether the query is streaming (SELECT ... FROM stream_name) or historical (SELECT ... FROM table(stream_name))
Any filters, aggregations, or window functions needed

If the user doesn't know the schema, suggest running DESCRIBE stream_name or SHOW STREAMS first.

Step 2 — Use the HTML Template

All Timeplus apps follow this single-file HTML template. Read references/STYLE_GUIDE.md for style rules and references/VISTRAL_API.md for chart configuration options.

<!DOCTYPE html>
<html lang="en">
<head>
  <meta charset="UTF-8" />
  <meta name="viewport" content="width=device-width, initial-scale=1.0" />
  <title>My Timeplus App</title>

  <!-- 1. React (required by Vistral) -->
  <script src="https://unpkg.com/react@18/umd/react.production.min.js"></script>
  <script src="https://unpkg.com/react-dom@18/umd/react-dom.production.min.js"></script>

  <!-- 2. Vistral peer dependencies — load BEFORE vistral -->
  <script src="https://unpkg.com/lodash@4/lodash.min.js"></script>
  <script src="https://unpkg.com/ramda@0.29/dist/ramda.min.js"></script>
  <script src="https://unpkg.com/@antv/g2@5/dist/g2.min.js"></script>
  <script src="https://unpkg.com/@antv/s2@2/dist/s2.min.js"></script>

  <script>
    // Prevent jsx-runtime from crashing on process.env access
    window.process = window.process || {
      env: { NODE_ENV: "production" }
    };

    // Some builds also expect global
    window.global = window.global || window;
  </script>

  <!-- 3. Vistral UMD — exposes window.Vistral -->
  <script src="https://unpkg.com/@timeplus/vistral/dist/index.umd.min.js"></script>

  <!-- 4. Proton JavaScript Driver UMD — exposes window.ProtonDriver -->
  <script src="https://unpkg.com/@timeplus/proton-javascript-driver/dist/index.umd.js"></script>

  <style>
    /* Timeplus design tokens */
    :root {
      --tp-bg-primary: #0f1117;
      --tp-bg-secondary: #1a1d27;
      --tp-bg-card: #1e2235;
      --tp-bg-hover: #252a3a;
      --tp-accent-primary: #7c6af7;
      --tp-accent-secondary: #4fc3f7;
      --tp-accent-success: #4caf82;
      --tp-accent-warning: #f7a84f;
      --tp-accent-danger: #f76f6f;
      --tp-text-primary: #e8eaf6;
      --tp-text-secondary: #9ea3b8;
      --tp-text-muted: #5c6380;
      --tp-border: #2e3450;
      --tp-font-mono: 'JetBrains Mono', 'Fira Code', monospace;
      --tp-font-sans: 'Inter', system-ui, sans-serif;
      --tp-radius-lg: 12px;
      --tp-shadow: 0 2px 12px rgba(0,0,0,0.4);
    }
    * { box-sizing: border-box; margin: 0; padding: 0; }
    body { background: var(--tp-bg-primary); color: var(--tp-text-primary); font-family: var(--tp-font-sans); font-size: 14px; line-height: 1.5; }
    .tp-app { display: flex; flex-direction: column; height: 100vh; overflow: hidden; }
    .tp-header { display: flex; align-items: center; gap: 12px; padding: 0 24px; height: 56px; background: var(--tp-bg-secondary); border-bottom: 1px solid var(--tp-border); flex-shrink: 0; }
    .tp-header-title { font-size: 16px; font-weight: 600; }
    .tp-header-badge { font-size: 11px; padding: 2px 8px; background: var(--tp-accent-primary); color: white; border-radius: 99px; font-weight: 500; }
    .tp-status { display: flex; align-items: center; gap: 6px; font-size: 12px; color: var(--tp-text-muted); margin-left: auto; }
    .tp-status-dot { width: 7px; height: 7px; border-radius: 50%; background: var(--tp-text-muted); }
    .tp-status-dot.connected { background: var(--tp-accent-success); }
    .tp-status-dot.error { background: var(--tp-accent-danger); }
    .tp-main { flex: 1; overflow: auto; padding: 20px 24px; display: grid; gap: 16px; }
    .tp-card { background: var(--tp-bg-card); border: 1px solid var(--tp-border); border-radius: var(--tp-radius-lg); padding: 16px; box-shadow: var(--tp-shadow); }
    .tp-card-title { font-size: 13px; font-weight: 600; color: var(--tp-text-secondary); text-transform: uppercase; letter-spacing: 0.04em; margin-bottom: 12px; }
    .tp-error { background: rgba(247,111,111,0.1); border: 1px solid var(--tp-accent-danger); border-radius: 8px; padding: 12px 16px; color: var(--tp-accent-danger); font-size: 13px; display: none; }
    .chart-container { min-height: 300px; }
  </style>
</head>
<body>
  <div class="tp-app">
    <header class="tp-header">
      <span class="tp-header-title">My Stream App</span>
      <span class="tp-header-badge">LIVE</span>
      <span class="tp-status">
        <span class="tp-status-dot" id="status-dot"></span>
        <span id="status-text">Connecting…</span>
      </span>
    </header>
    <main class="tp-main">
      <div class="tp-error" id="error-panel"></div>
      <div class="tp-card">
        <div class="tp-card-title">Live Stream</div>
        <div class="chart-container" id="chart"></div>
      </div>
    </main>
  </div>

  <script>
    // ---- Library globals ----
    // proton-javascript-driver UMD exposes window.ProtonDriver
    const { ProtonClient } = window.ProtonDriver;
    // vistral UMD exposes window.Vistral; React/ReactDOM are also on window
    const { Vistral, React, ReactDOM } = window;
    const { StreamChart } = Vistral;

    // ---- Configuration ----
    const SQL = 'SELECT * FROM my_stream';  // TODO: replace with your SQL

    // Connect to the Proton proxy running on the agent at localhost:8001
    const client = new ProtonClient({ host: 'localhost', port: 8001 });

    // ---- State ----
    let columns = [];
    let rows = [];
    let chartRoot = null;

    // ---- Helpers ----
    function setStatus(state, message) {
      document.getElementById('status-dot').className = 'tp-status-dot ' + (state || '');
      document.getElementById('status-text').textContent = message;
    }

    function showError(msg) {
      const panel = document.getElementById('error-panel');
      panel.style.display = msg ? 'block' : 'none';
      panel.textContent = msg || '';
      if (msg) setStatus('error', 'Error');
    }

    function inferColumns(row) {
      return Object.entries(row).map(([name, value]) => ({
        name,
        type: typeof value === 'number' ? 'float64'
            : String(value).match(/^\d{4}-\d{2}-\d{2}/) ? 'datetime64'
            : 'string',
      }));
    }

    // ---- Render chart ----
    function renderChart() {
      if (columns.length === 0) return;
      if (!chartRoot) chartRoot = ReactDOM.createRoot(document.getElementById('chart'));

      // TODO: adjust config for your data — see references/VISTRAL_API.md
      const config = {
        chartType: 'line',
        xAxis: columns[0].name,
        yAxis: columns[1]?.name ?? columns[0].name,
        temporal: { mode: 'axis', field: columns[0].name, range: 60 },
      };

      chartRoot.render(
        React.createElement(StreamChart, {
          config,
          data: { columns, data: rows, isStreaming: true },
          theme: 'dark',
        })
      );
    }

    // ---- Main query loop ----
    async function runQuery() {
      try {
        setStatus('', 'Connecting…');
        const { rows: stream, abort } = await client.query(SQL);
        setStatus('connected', 'Connected');

        // Expose abort so it can be called externally if needed
        window.stopQuery = abort;

        for await (const row of stream) {
          if (columns.length === 0) columns = inferColumns(row);
          rows = [...rows.slice(-999), row];  // keep last 1000 rows
          renderChart();
        }

        setStatus('', 'Stream ended');
      } catch (err) {
        showError(err.message);
      }
    }

    runQuery();
  </script>
</body>
</html>

Step 3 — Proton Connection

Read references/PROTON_DRIVER.md for the full driver API.

The driver connects directly to the Proton proxy at localhost:8001. UMD global: window.ProtonDriver.

const { ProtonClient } = window.ProtonDriver;

const client = new ProtonClient({ host: 'localhost', port: 8001 });

const { rows, abort } = await client.query('SELECT * FROM my_stream');

for await (const row of rows) {
  // row is a plain JavaScript object: { col1: val1, col2: val2, ... }
  console.log(row);
}

// To cancel a running streaming query:
abort();

Step 4 — Visualization with Vistral

Read references/VISTRAL_API.md for the complete component reference.

Vistral UMD global: window.Vistral. Key components: StreamChart, SingleValueChart, DataTable.

All charts are rendered via React (React.createElement). Always pass theme="dark".

Quick chart config reference:

Chart type	`chartType`	Best for	temporal `mode`
Time series	`'line'` or `'area'`	Scrolling metrics over time	`'axis'`
Horizontal bars	`'bar'`	Ranked categories, latest snapshot	`'frame'`
Vertical bars	`'column'`	Categorical comparison	`'frame'`
KPI tile	`'singleValue'`	One big number	`'frame'`
Live table	`'table'`	Raw rows / mutable state	`'key'`
Geo points	`'geo'`	Location data	`'key'`

Step 5 — Apply Timeplus Style Guide

Read references/STYLE_GUIDE.md for full rules. Core rules always applied:

Dark background #0f1117 for page, #1e2235 for cards
Purple #7c6af7 for primary accents
Inter font for UI, monospace for data values
Cards with border-radius: 12px and border #2e3450
Header 56px with LIVE badge and connection status indicator

Step 6 — Deliver the App

Write the complete single-file HTML, filling in the correct SQL and chart config
use workspace_create_app to create the app
Present the file — it opens directly in a browser with no server or build step needed

For PulseBot deployment:

workspace_create_app(
  session_id=session_id,
  task_name="My Stream App",
  html=open("my-app.html").read()
)

SQL Patterns

Load the timeplus-sql-guide skill for all SQL questions:

Streaming vs historical queries
Window functions (tumble, hop, session)
Aggregations, joins, filters
Schema introspection (DESCRIBE, SHOW STREAMS)

Checklist Before Delivering

Single HTML file — no separate JS/CSS files, no package.json, no build step
Script load order: React → ReactDOM → lodash → ramda → @antv/g2 → @antv/s2 → vistral → proton-javascript-driver
window.ProtonDriver destructured for ProtonClient
window.Vistral destructured for StreamChart (and other components as needed)
ProtonClient configured with { host: 'localhost', port: 8001 }
SQL uses correct Proton streaming syntax
columns inferred from first received row using inferColumns()
Chart config.temporal uses correct mode for the query type
Status indicator updates on connect, stream end, and error
abort() stored on window.stopQuery or similar
CSS uses Timeplus design tokens (CSS variables) from the style template

Pulsebot App Builder

File contents

Pulsebot App Builder

Overview

Step-by-Step Workflow

Step 1 — Clarify Requirements

Step 2 — Use the HTML Template

Step 3 — Proton Connection

Step 4 — Visualization with Vistral

Step 5 — Apply Timeplus Style Guide

Step 6 — Deliver the App

SQL Patterns

Checklist Before Delivering

Frequently asked questions

Pulsebot App Builder

File contents

Pulsebot App Builder

Overview

Step-by-Step Workflow

Step 1 — Clarify Requirements

Step 2 — Use the HTML Template

Step 3 — Proton Connection

Step 4 — Visualization with Vistral

Step 5 — Apply Timeplus Style Guide

Step 6 — Deliver the App

SQL Patterns

Checklist Before Delivering

Related skills

Frequently asked questions

How do I install the Pulsebot App Builder skill?

What does the Pulsebot App Builder skill do?

Is Pulsebot App Builder safe to use?

Which AI agents work with Pulsebot App Builder?

Is Pulsebot App Builder free to use?

Who published Pulsebot App Builder?