Decision-First Dashboard skill
A skill that decides whether a dashboard is needed and builds it around a confirmed decision, never inventing rankings or scores.
- 58 ★ on the repository
- Last commit
- MIT
Minimal access
Why this level
- The skill builds HTML and SVG locally from your data and sends it nowhere.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/decision-first-dashboardDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The skill builds HTML and SVG locally from your data and sends it nowhere.
This is third-party code. Review the repository files before installing.
What it does
Decision-First takes dashboard screenshots, Figma frames, existing panel code or verified metrics and first asks whether a dashboard is needed at all. Until the user confirms the decision and action, routing and rendering do not proceed. The skill then selects metrics that can change the decision, keeps the rest for drilldown and chooses legal visual forms. A compiler enforces the rules: without a grounded ordering basis there is no ranking, incomparable dimensions stay off a radar chart and a composite score is allowed only with a defined scoring model. The finished file is checked against the declared output contract and carries verifiable data provenance.
Who it is for. Analysts and product designers who overload dashboards with dozens of metrics and want only decision-relevant ones on the first screen.
Good fit when
- A seventy-metric dashboard must shrink to what changes the decision.
- A Power BI or Tableau screenshot needs a review for superfluous charts and unsupported conclusions.
- Before commissioning a panel you want to know whether a report or a table is enough.
- Data must be shown with the source of every number.
Not a fit when
- You need a quick pretty panel with no questions about the decision.
- The data is unverified and the errors lie in the sources themselves.
- You want a ready service rather than an agent-side process.
Example request
Review this dashboard: first decide whether it should exist and ask only what is needed to confirm the decision and action.Limitations
The skill may decline to build a panel: without a confirmed decision and action, rendering will not start, and for some requests it returns a polite refusal. Verification confirms the file matches the inputs and the contract, while correctness of the data itself remains with the author. It contains JS scripts and JSON schemas, so Node.js is needed. Installing through `npx skills` places the skill in the chosen agent directory.
How to disable. Remove the skill with your skill manager or delete the `decision-first-dashboard` directory from the agent skills folder.
FAQ
How do I install Decision-First?
Run `npx skills add fourleaf13-hue/decision-first-dashboard`. Then attach a screenshot or panel code and describe which decision it supports.
What does the skill accept as input?
Dashboard screenshots, Figma frames, existing panel code, structured data and source-backed facts from other material.
Why does the skill refuse a radar chart?
It needs dimensions on a shared scale. Mixed units without a defensible comparison do not go on one chart, and the skill suggests another form.
Does the skill invent rankings and scores?
No. A ranking needs a grounded ordering basis and a composite score needs a defined scoring model with scale, weights and bands.
What does final verification check?
That the delivered HTML and SVG match the approved structure and contract, and the provenance of the data. It does not prove the input numbers are correct.
Discussion
No questions or reviews yet.
Related
SkillsA dashboards-as-code CLI: an agent describes panels in YAML and TSX, DAC validates them, builds SQL from metrics and serves them
dbt Charts
dbt-charts
Declarative YAML on top of SQL for dashboards: you describe how to show data, agents write YAML, people audit. The dbt-charts package and the dct CLI, beta
Lightdash skills: metrics, charts and dashboards as YAML, branch previews, dbt SQL rules and a safety check before a server upgrade
MCP Dashboards: interactive charts in chat
MCP Dashboards
32 tools draw interactive charts, KPI widgets and dashboards right inside the chat in Claude Desktop, VS Code and other clients
Skill and CLI for tables: cleaning with recorded decisions, level-based statistics, a PDF report and a separate audit.
A data visualization skill for agents: turns data into polished interactive HTML charts and full reports with one visual language
README in short
The project contrasts generating dashboards with governing them by rules: first the question of necessity, then evidence selection, legal forms and result verification. A refusal or clarification beats a polished but unsupported file.