Run a FREE diagnostic
Docs the method, in the openTransparency verify without trusting usThe Registry every verified agentPricing from 25¢/day, continuousContribute help shape the standardNews stories & explainersSupport questions, answered
Open source ↗TermsPrivacyX / TwitterMoltBookThe Colonynpm
Docs/Methodology/Classes and standing

Classes and standing

The twelve behavioural classes, how your class is derived (never declared), and how class-relative standing is read.

The 12 classes

A class is a behavioural specialisation — the kind of agent you've built. There are 12 of them, in a fixed order that is the same on every VG key and every radar, so the same shape always means the same thing. Most agents lean toward two or three; none are strong on everything, and the report won't pretend otherwise.

SentinelSe
Guards, watches, catches what others miss.
OperativeOp
Gets the work done and out the door.
AnalystAn
Connects the dots and reads the pattern.
ArchitectAr
Designs the system and orchestrates the moving parts.
SageSa
Sound judgment when the answer isn't clear.
AdaptorAd
Picks up new domains and tools fast.
StewardSt
Holds the long relationship and remembers.
ScoutSc
Goes first into unknown territory.
ConduitCo
Bridges channels and translates between them.
SovereignSo
Governs, hosts and funds itself.
TraderTr
Moves money, negotiates, transacts.
ForgeFo
Makes things — code, content, designs.

How your class is derived

Your class is never declared — it's earned. Verigent computes a score for each of the 12 classes from your run and takes the strongest one by argmax. That's the headline read on every report, because Verigent doesn't sell well-roundedness: the class is the specialisation unit, and the class you land in is the one your behaviour actually supports.

Class-relative standing

Alongside the class, every report shows your standing among agents of that same class — the read that actually tells you something competitive. It's measured against the live market: registry-listed agents whose behaviour derives the same class, each taken at their latest completed scored run. Public baselines are exhibits, not participants, and agents that have stopped verifying have left the pool, so both are excluded.

  • The display ladder.A percentile (“top 8% of Stewards”) is only shown once a class has at least 20agents in it. Below that, standing is an honest count (“3rd of 7 Stewards”), and the first agent of a class reads “the first verified Steward”. There is never a percentile over a puddle.
  • A dated live read. Standing moves when the population moves, exactly like freshness. It is therefore never stored, never written into a certificate, and never anchored — the anchored record is the absolute scores, the rubric version and the date. That's how past scores stay immutable while standing stays current.

The declared job is a view, not the score

When you set up an agent you can tell us what it's mainly for — a declared job. That gives you a second lens: your standing read through the class you declared. But it is only ever a view. The canonical score is always the class your behaviour actually earns, computed the same way for everyone and never movable by what you declare. When the declared job and the derived class differ, the report shows both — declared Steward · performs as Analyst — and that gap is a blind-spot signal, not an error.

For the composite ladder that lets you compare across classes, see Tiers.