mmakosiewicz

AI Visibility Dashboard Kit — AI skill for Claude Code

AI community

Look at the questions you track in AI assistants, find out which numbers are worth a dashboard, then have a coding agent build it.

How to install AI Visibility Dashboard Kit

This entry records only its repository, not the path inside it, so there is no exact command to give. Open mmakosiewicz/ai-visibility-dashboard-kit and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What AI Visibility Dashboard Kit does

Look at the questions you track in AI assistants, find out which numbers are worth a dashboard, then have a coding agent build it. Works in Claude or ChatGPT with the Ahrefs MCP server. Built on Ahrefs Brand Radar.

Alternatives in AI

AIUsage — One dashboard to manage all your AI subscriptions — track quotas, costs, accounts, Claude Code proxy, and Code 635 ★
ZotSeek — AI semantic search for Zotero, with a built-in MCP server for AI agents (Claude Code, Codex) 193 ★
Agf — Local-first TUI to find, search, and resume AI coding-agent sessions across Claude Code, Codex, Grok Build, Ki 148 ★

README

AI visibility dashboard kit

A free guide and toolkit that looks at the questions you track in AI assistants, tells you which numbers are worth putting on a dashboard, and then helps a coding agent build that dashboard for you. It works from [Ahrefs Brand Radar](https://ahrefs.com/brand-radar) data.

What it does

Most AI visibility dashboards show the same numbers for every question, whether or not they make sense. This starts from your own list of tracked questions.

You give it a Brand Radar report. It pulls the questions you track and a day of the answers AI assistants gave, and asks you for your brand and competitors.

It then does four things:

Groups your questions into topics in your own terms ("choosing a treatment", "finding a specialist"), not a generic template.
Works out what kind of question each one is. Some ask for a recommendation, some ask how to do something, some just want information. That decides what you can measure. Your brand can only "win" a question that asks for a recommendation. On a general information question AI rarely recommends anyone, so the useful number there is how often your website is cited as a source.
Tests every metric on real answers. Each one is labelled worth tracking, too little data or nothing to measure, with the calculation shown: "mention rate = answers that name you ÷ all answers = 23 ÷ 32".
Hands the approved plan to a coding agent, with the rest of this guide: how to store and grade answers, how each metric is calculated, and the mistakes that make AI visibility numbers wrong.

The numbers in the proposal come from a small sample. They show whether a metric is worth tracking, not what its real value is.

**Example.** On a 66-question report for a cancer charity, the proposal took about three minutes and a few cents. Only one topic turned out to be a real contest between organizations: there the charity was named in 72% of answers. The other 47 q