Team Skills Matrix Builder
Build a living skills matrix - people by skills, with an approved proficiency level in every cell - that managers actually keep current, with color-coded views, filters, and per-skill coverage counts.
A live, color-coded skills matrix where ratings are manager-approved before they count, coverage per skill is summarized at a glance, and the whole grid exports to CSV.
Before you start
- A list of the skills you care about, ideally grouped by area
- Your team roster (names + who their manager is)
- A rough idea of your proficiency levels (e.g., none / learning / can-do / can-teach)
- Free accounts on Vercel, Supabase, and Resend (the plan walks you through them)
The problem this kills
Somewhere in your shared drive is a "Skills Matrix" spreadsheet. It was beautiful the day it was made. Then one person changed jobs, two new people joined, someone added a column nobody else understood, and the colors stopped meaning anything. Now nobody trusts it, nobody updates it, and when leadership asks "do we have enough people who can run the close process?" you go ask around by hand.
The real problem isn't the spreadsheet - it's that a skills matrix has to be maintained and trusted. Self-ratings drift. "Can teach" means different things to different people. There's no record of who approved what, or when. So the grid rots.
This Implementation Plan builds you a proper tool instead: a place where the proficiency scale is defined once and means the same thing for everyone, where people can self-rate but a manager has to approve before a rating counts, and where you can instantly see how many people cover each skill.
What you'll build
A small, private web app for your team:
- A skills list grouped by area (e.g., "Finance: month-end close, AP, AR" / "Tools: Excel, Power BI, Salesforce").
- A team roster with reporting lines, so managers see their own people.
- A clearly defined proficiency scale that's spelled out on screen so a rating means the same thing to everyone.
- A rating workflow: people self-rate, managers review and approve - and upgrades like "can teach" get extra scrutiny before they're committed.
- A color-coded matrix view (people down the side, skills across the top) with filters by area, team, or proficiency.
- A coverage summary per skill: how many people are at each level, and where you're dangerously thin.
- A one-click CSV export of the whole matrix in clean columns.
What's inside the Implementation Plan
The plan is a single file you paste into an AI coding agent (Claude Code). It tells the agent exactly what to build, in order, with a copy-ready prompt at the end of every step.
Crucially, it starts by interviewing you about your business - your current process, the systems and spreadsheets you already use, the exact names of your skills and proficiency levels, your team size and structure, and your real approval rules. It reads a short tailored spec back to you and waits for your thumbs-up before building anything. You get a tool shaped around how your team actually works, not a generic template you have to bend yourself into.
Inside you'll find:
- The discovery interview (8-12 targeted questions) and how the agent turns your answers into the data model.
- The exact build steps: schema, login, the rating + approval workflow, the color-coded matrix, the coverage summary, and the CSV export.
- A "No API yet?" fallback so you can import your existing skills + roster from a Google Sheet / CSV today, and export the finished matrix as CSV - no integration required.
- A verification checklist so you can prove it works before you trust it.
The governance it includes (this is the point)
This isn't a free-for-all spreadsheet. The plan builds in real controls from the start:
- Login so only your team can get in.
- Row-level security so a manager only ever sees their own organization's people and ratings.
- A human-in-the-loop approval gate: a self-rating is a draft until a manager reviews and approves it. Upgrades (like "can teach") get explicit confirmation before they're committed to the matrix.
- A complete audit trail: who rated, who approved, what the old value was, and when - so the grid is defensible, not just decorative.
- Duplicate guards: one current rating per person-per-skill, with full history kept behind it, so the same entry can't quietly get processed twice.
Who it's for
Team leads, ops and BPM managers, and L&D folks who already track skills in a spreadsheet that nobody updates - and who want a live, trustworthy picture of the team's capabilities without hiring a developer or buying a heavyweight LMS. You don't need to know how to code. You need to know your team.
You've got this - paste the first prompt and let the agent interview you.