w0 scoreDefined by WARI →Deterministic

Make your agent readiness measurable.

The w0 score turns the interaction between agents and your digital infrastructure intoone governed number, built from deterministic audits and held to the evidence underneath it. The model behind it is theWebzero Agent Readiness Index (WARI).

Know where you stand. See what is holding you back. Track what improves.
Get your w0 scoreSee how it works
01 / The gap

Agents are becoming users. Almost nothing measures infrastructure for them.

The agentic web adds a layer of interaction that existing analytics cannot see. Page views and load times describe a human session. They say nothing about whether an agent could discover you, understand what you offer, reach an interface it can use, or act inside boundaries you set.

Fig. 01Two visitors
Two agents at a laptop, browsing
Two agents on your product page. Analytics counted one session and could not say what either came for.
Analytics tell you
How many people visited
Which pages they viewed
Where they dropped off
How fast the page loaded
Nothing tells you
Whether an agent can discover, understand, reach and safely act on your infrastructure.
The w0 score makes that layer visible.
02 / The score

One score, and a complete view of what makes it.

Three layers, audited independently and reported separately. Discovery and action surface earn points. Governance gates them: a failed control caps the reading whatever the other layers say. Weightings, bands and signal definitions are set byWARI.

w0 score / readingproperty redacted · audit 2026-08-28run 04 of 04
w0 score
12/ 100
Not agent ready
Band Critical. Baseline audit, no previous reading.
Discovery21
Can agents find and understand the infrastructure?
Action surface6
Can they reach an interface and act through it?
Governance14
Is that interaction controlled, bounded and auditable?
Illustrative reading. Client scores, domains and audit data are never published, including on this site.
03 / Evidence

A score is only useful if you can explain it.

We do not score what a company says about its infrastructure. Every point comes from an observable signal, and every signal keeps the evidence that produced it. Signal, evidence, score, in that order and never out of it.

Example. Four of the 34 signals in a full audit.
Signal
Evidence
Scores into
robots.txt
Crawler and agent access observed, per user agent
✓Discovery
llms.txt
Structure and content read as an agent would read it
✓Discovery
OpenAPI · MCP · A2A
Protocol implementation exercised, not just declared
✓Action surface
Access controls
Permissions, rate limits and confirmation steps tested
✓Governance
A signal we could not observe is reported as not evaluated, never inferred.
04 / What it measures

The infrastructure agents actually depend on.

Earns points
Discovery

How easily an agent can find the property, understand what it offers and learn what it can do there.

robots.txtsitemapllms.txtagent cardsstructured datasemantic structuredeveloper docspricing exposureAPI discovery
Earns points
Action surface

Whether usable interfaces and protocols exist for an agent to act through, and whether they work when called.

browser operabilityOpenAPIMCPA2AWebMCPAPI accessauth discoverySDK availabilityagentic commerce
Gates the score
Governance

Whether agent interaction is bounded, consented, rate-limited and auditable.

agent access controlspermission boundariesrate limitshuman confirmationauditabilityinjection resistance
05 / The console

Signal by signal, with the audit history beside it.

Readiness is a tracked number, not a one-time report.

w0 score / consoleproperty redacted · audit 01 · 2026-08-28GOVERNED
34
signals audited
6
passing
9
partial
19
failing
Signals by layer
Three of each layer shown. Every row keeps its evidence.
Discovery3 of 12 passing
✓robots.txtagents permitted
!llms.txtpresent, thin
✕agent cardnot published
Action surface1 of 14 passing
✓OpenAPIpublished, v3.1
!MCP serverread only
✕auth discoveryno documented path
Governance2 of 8 passing
✓rate limitsdeclared, enforced
✓audit trailper agent action
✕human confirmationnot implemented
Score history
Illustrative, remediation between each audit.
12Baseline
38Audit 02
61Audit 03
84Audit 04
Movement across four audits
Discovery+52
Action surface+79
Governance+64
Behaviour, measured separately
Open a fixed deposit
Simulated in Agentrek, not scored here. Stopped at funding.
Open the run →
What to fix next
Ranked by score movement per unit of engineering effort.
CriticalAdd a human confirmation step before an agent commitshuman confirmation: not implemented, and this layer gates the scoreGovernance
HighPublish a documented authentication pathauth discovery: no documented path, so every authenticated journey stops hereAction surface
MediumPublish an agent cardagent card: not published, so capability has to be inferred from the pageDiscovery
LowFill out llms.txtllms.txt: present, thinDiscovery
Fix the infrastructure. Re-run the audit. Measure the change.View your readiness report
06 / The loop

Infrastructure is the foundation. Journeys are the proof.

The w0 score measures what your infrastructure exposes.Agentrek puts real agents on it and records what happens. Two readings, kept separate on purpose: a simulation result never merges into the score, and the score never stands in for observed behaviour.

Measure
w0 score
Deterministic audit of discovery, action surface and governance.
Fix
Remediate
Ranked by score movement per unit of engineering effort.
Simulate
Agentrek
Real agents and swarms run the journeys on the fixed infrastructure.
Learn
Evidence
What the agents did, where they stopped and why, with the trace.
Improve
Re-measure
The identical audit, re-run, so the delta is proven rather than asserted.
05 / Why measure

Don't guess whether your infrastructure is ready for agents. Measure it.

Traditional website audit
w0 score
Human-centric
Agent-centric
Static checklist
Evidence-backed measurement
Page-level analysis
Infrastructure-level readiness
Tells you what is present
Measures what agents can actually use
One-time report
A tracked score
Generic recommendations
Signal-level evidence
Never connects to agent behaviour
The foundation agent journeys are built on
Methodology is disclosed to clients under engagement. No client score, signal or audit is published, including on this site.

Know your score. Build for the agents.

We audit the property and return the score with the evidence under every signal.

Get your w0 scoreRead the WARI overview