Turn freeform dashboard requests into an interactive HTML prototype and a Tableau .twbx workbook by running an 8-step skills pipeline: analyze requests, extract brand tokens, acquire CSV data, plan layouts, build mockups, generate implementation specs, and assemble the final Tableau file. A workflow orchestrator skill recommends the next step.
Extracts design tokens (palette, type, spacing) into DESIGN-TOKENS.md for the tableau-dashboard-plugin workflow, so the mock and the Tableau workbook match the brand. Reads the analyst's branding/ (a branding.md spec and/or an org template .twb, plus logo/icons), or runs a short brand interview when none exists. Every value guessed from a Tableau default is flagged. Skippable only once branding/branding.md exists. Use when the user wants to set up branding/design tokens, or when tableau-route reports brand is next. Step 4 of 8 in the workflow.
Acquires the data for a tableau-dashboard-plugin project and writes the DATA-MODEL.md handoff artifact. CSV mode (the default, zero-credential path) profiles the analyst's data/*.csv into documented field names + types and validates that headers match exactly; the scaffold/sample-data/ demo CSVs are the fallback. The published-ds route (datasources.json + .env) samples published sources through the VizQL Data Service into data/*.csv. Use when the user wants to acquire or model dashboard data, build DATA-MODEL.md, or when tableau-route reports data is next. Step 3 of 8 in the workflow.
Scaffolds a new tableau-dashboard-plugin project. Creates a scaffold/ folder of demo examples (EXAMPLE-DASHBOARD-REQUEST.md, EXAMPLE-datasources.json, .env.example, branding/EXAMPLE-branding.md, starter sample-data/) and initializes the STATE.md manifest, recording the target Tableau Desktop version so the build step never has to re-ask. Use when the user wants to start a new Tableau dashboard project, scaffold the workflow, or when tableau-route reports a fresh project. Step 1 of 8 in the workflow.
Turns a free-form dashboard request into a structured PRD.md for the tableau-dashboard-plugin workflow. Reads the analyst's DASHBOARD-REQUEST.md (or text pasted into the terminal, or the scaffold/ demo example) and authors a schema-complete PRD.md. Optional and idempotent — if a PRD.md already exists it offers to refine rather than overwrite. Use when the user wants to structure a dashboard request, write a PRD, or when tableau-route reports intake is next. Step 2 of 8 in the workflow.
Renders the interactive HTML demo (mock.html) for the tableau-dashboard-plugin workflow at the plan's screen size, populated from the real sample CSVs, so the analyst can share a realistic demo and validate direction before any Tableau work. Builds strictly from DASHBOARD-PLAN.md — every KPI, chart, filter, and interaction — and ends with a coverage checklist proving each plan element (screen size, every filter, every interaction) was rendered, so omissions are visible before the stakeholder sees it. Enforces readable slot sizing (no compressed, out-of-bounds, or empty-space-heavy charts). Reads DASHBOARD-PLAN.md and the sample CSVs, and DESIGN-TOKENS.md if present. Use when the user wants to build the mock, render the demo, or when tableau-route reports mock is next. Step 6 of 8 in the workflow.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
A Claude Code plugin that turns a free-text dashboard request into an interactive HTML demo and a Replace-Data-Source-ready Tableau .twbx — through a guided, approval-gated workflow split across 8 single-job skills plus a router.
The original release was a single monolithic skill (/tableau-dashboard-creator) that loaded the entire workflow — every step's instructions, references, and snippets — into one long conversation. That is expensive: context grows step over step, and on a smaller plan you run out of budget before you reach the workbook.
This plugin breaks that monolith into 8 independent skills, each running in its own fresh conversation and handing off to the next through files on disk. No single skill ever loads the whole workflow — only its own slice of the contract. The result is dramatically lower per-step token use, so the full pipeline is practical to run even on a $20 Claude Pro license, not just higher tiers.
Portability. Each skill is a plain
SKILL.mdfile, so the content ports to Cursor, Codex, or any other agentic tool that reads skill files. The install and invocation documented here are specific to Claude Code's plugin system.
| Requirement | Details |
|---|---|
| Claude Code | The plugin targets Claude Code's plugin system (/plugin install, namespaced skills). |
| Python 3.9+ | The helper scripts use pandas 2.x (which requires 3.9+). |
| Python dependencies | pip install -r requirements.txt — pandas, python-dotenv, requests (published-datasource route), lxml (workbook XSD validation). |
| Tableau Desktop | Only needed to open the generated .twbx workbook. |
Add this repo as a plugin marketplace, then install the plugin. Works straight from GitHub, no clone, and picks up updates:
/plugin marketplace add https://github.com/laviDrori0702/tableau-dashboard-creator-skill.git
/plugin install tableau-dashboard-plugin@dashboard-creation-tool
Use the HTTPS URL (as above), not the owner/repo shorthand — the shorthand clones over SSH and fails if you don't have GitHub SSH keys configured. (If you cloned locally, point the marketplace at the checkout instead: /plugin marketplace add ./path/to/tableau-dashboard-creator-skill.)
For hacking on the plugin itself, load it for the current session only:
git clone https://github.com/laviDrori0702/tableau-dashboard-creator-skill.git
cd tableau-dashboard-creator-skill
claude --plugin-dir .
Plugin skills are namespaced by the plugin name. Typing /tableau-init finds nothing; the actual command is:
/tableau-dashboard-plugin:tableau-init
/tableau-dashboard-plugin:tableau-route ← "what step am I on / what's next?"
Every skill has disable-model-invocation: true, so Claude never auto-runs them — you drive the workflow explicitly, one step at a time. Not sure where you are? Run the router (…:tableau-route) and it tells you the single next skill to run.
Each step is owned by exactly one skill, reads a known set of artifacts, and writes exactly one primary artifact. Run them in order (the router enforces it); approve each before moving on.
| # | skill (/tableau-dashboard-plugin:…) | what it does | reads → writes | skippable |
|---|---|---|---|---|
| 1 | tableau-init | Scaffold the project; record target Tableau version | — → STATE.md + scaffold/ demo examples | no |
| 2 | tableau-intake | Structure the free-text request into a PRD | DASHBOARD-REQUEST.md / pasted text → PRD.md | yes |
| 3 | tableau-data | Acquire & profile the data | data/*.csv (or published-ds via VDS) → DATA-MODEL.md + data/*.csv | no |
| 4 | tableau-brand | Extract design tokens from branding | branding/ → DESIGN-TOKENS.md | yes |
| 5 | tableau-plan | Blueprint: screen size, slots, KPIs, charts, filters, interactions (stable ids) | DATA-MODEL.md (+ PRD.md, DESIGN-TOKENS.md) → DASHBOARD-PLAN.md | no |
| 6 | tableau-mock | Interactive HTML demo from real sample data | DASHBOARD-PLAN.md, data/*.csv → mock-version/v_N/mock.html | no |
| 7 | tableau-spec | Map every mock element to a concrete Tableau construct | mock.html, DASHBOARD-PLAN.md → mock-version/v_N/IMPLEMENTATION-SPEC.md | no |
| 8 | tableau-build (in development) | Generate the version-aware, XSD-validated workbook | IMPLEMENTATION-SPEC.md, DATA-MODEL.md, data/*.csv → mock-version/v_N/dashboard.twbx | no |
npx claudepluginhub lavidrori0702/tableau-dashboard-creator-skill --plugin tableau-dashboard-pluginBuild Vizro dashboards from concept to deployment. Enforces a 2-phase workflow covering requirements, layout design, visualization selection, implementation with Python, and testing.
Connect to Looker and interact with your data using LookML.
Write SQL, explore datasets, and generate insights faster. Build visualizations and dashboards, and turn raw data into clear stories for stakeholders.
Design effective KPI dashboards with metrics selection, visualization best practices, and real-time monitoring patterns. Use when building business dashboards, selecting metrics, or designing data visualization layouts.
Business analysis with data storytelling and KPI dashboard design
Designs effective KPI dashboards with proper metric selection, visual hierarchy, and data visualization best practices. Use when building executive dashboards, creating analytics views, or presenting business metrics.