Run your diagnostic

How to start

The first diagnostic is free and takes about 30 minutes. There's no API key required to read a report, and it publishes nothing unless you choose to. Head to the start page to point Verigent at your agent, or run the tester from the command line:

npx verigent

Then restart your agent's session so the MCP server it just registered actually loads, and tell it to sit the Verigent test. Its client may ask you to approve mcp__verigent tool calls the first time it uses them (HTTPS to verigent.ai only) — approve them, or add mcp__verigent to its allow list.

Either way, Verigent pulls a set of tasks, runs them against your agent, scores the results, and mints a VG key and a report.

What your first run measures

Your opening test is a genuine read straight away — a provisional snapshot across the cognitive pillars (Model, Backbone and Agent): 24 dimensions with deterministic, server-checked proofs. It's a real picture on day one, and it only fills in from there.

• Cross-run memory starts scoring on your second run — Verigent plants something in one session and checks whether a later one recalls it, so it physically needs a prior run to measure honestly. Until then it shows as pending, not failed.
• The Sovereignty pillar — real on-chain payments and signatures — is measured inside continuous verification, where your agent performs the actions instead of describing them.
• Every continuous cycle fills in more of the battery, so both your coverage and your headroom climb week over week.

After the run

Your first diagnostic is a free one-shot read — a snapshot. To keep it re-tested as you ship changes, continuous verification runs on a schedule (a paid upgrade), tracking every change automatically. The score isn't a verdict to defend — it's a baseline you keep raising.

Once you have a report in front of you, read Reading your report & fixing what it finds to make sense of every part of it and work the fix list in the right order.