Separate the action.
Challenge the outcome.

Infrastructure between existing agents and existing software. Designed around explicit permissions, scoped claims and independent evidence.

FIG. 02 / PRODUCTION EXECUTION PATH
APPROVED INTENTCUSTOMER TRUST BOUNDARYFRESH AUTHORIZED READBACK
01 / EXECUTION

User prompt

Existing user intent.

02 / EXECUTION

Existing agent

Your agent remains the agent.

03 / EXECUTION

ANL gateway

Typed SDK / HTTP; optional MCP tools.

04 / EXECUTION

Control server

Tenant registry, policy and current qualification.

05 / CUSTOMER

Local worker

Permissions and bounded actions.

06 / CUSTOMER

Target app

Existing, unmodified software.

↳ INDEPENDENT VERIFIER

Expected effects + forbidden effects + fresh evidence.
Independently implemented checks, not an action pathway confirming itself.

VERIFIED / FAILED / UNKNOWN

Discover in the sandbox.
Execute in production.

01

Start with intent

EXISTING PROMPT → EXISTING AGENT

The existing agent receives the user’s prompt. ANL sits beneath it, not in its place. The gateway resolves a scoped business operation, not an unrestricted UI instruction.

02

Observe in isolation

SANDBOX / DISCOVERY

In an authorized, disposable test environment, the worker encodes permitted DOM and ARIA structure into a typed observation IR. Windows UIA is a future direction, not finished native support.

03

Experiment safely

SANDBOX / CAUSAL HYPOTHESES

Bounded experiments induce candidate capabilities from partial, observable business transitions. These are causal hypotheses, not universal decompilation or access to hidden internals.

04

Synthesize a contract

SANDBOX / FALSIFIABLE SCOPE

Describe typed inputs, permission policy, preconditions, expected effects, forbidden effects and evidence requirements. Every claim must be narrow enough to challenge.

05

Attempt to falsify

SANDBOX / ADVERSARIAL TESTS

Challenge wrong targets, stale observations, false success toasts and forbidden side effects. An independently implemented verifier evaluates the contract rather than trusting the action pathway.

06

Qualify, then activate

CERTIFICATION / HUMAN DECISION

A scoped capability certificate records qualification and its limitations. Human activation is separate and explicit. Qualified does not automatically mean active.

07

Publish typed tools

REGISTRY / CURRENT QUALIFICATION

Expose only explicitly activated and currently qualified capabilities as typed SDK / HTTP operations and, optionally, MCP tools. Certification is a capability state, not a promise about every future invocation.

08

Execute within bounds

PRODUCTION / ACTIVE + CERTIFIED + AUTHORIZED

The server coordinates an approved intent. The customer-side worker independently enforces local authorization and performs bounded UI actions. No unapproved exploratory production writes.

09

Challenge fresh readback

PRODUCTION / INVOCATION VERDICT

Fresh authorized observations are checked for expected and forbidden effects. Return VERIFIED, FAILED or UNKNOWN with evidence tiers and limitations. UI readback alone is not an external backend authority.

10

Suspend on drift

LIFECYCLE / RECERTIFICATION

Changed interfaces, policies or failed qualification can suspend a capability. Recertification and explicit activation restore eligibility; stale certificates must not permit production execution.

The server coordinates.
The worker enforces.
The verifier challenges.

CERTIFICATION / CONTROL PLANEA

Dedicated ANL server

Tenant-scoped registry, policies, discovery and capability qualification. Explicit human activation before publication.

REGISTRYPOLICIESCERTIFICATION
— LOCAL TRUST BOUNDARY —
CUSTOMER / EXECUTION PLANEB

Endpoint worker

Encodes permitted interface observations and translates approved intents into bounded UI actions. Enforces local permissions.

OBSERVEAUTHORIZEEXECUTE
↳ INDEPENDENT VERIFIER

Fresh authorized readback. Expected and forbidden effects. Evidence-bound verdict.

VERIFIED / FAILED / UNKNOWN

Not all evidence
is equal.

E2

Fresh UI readback

The expected state was read back through a fresh authorized observation. This is visible-interface evidence, not proof of an authoritative backend state.

UNKNOWN IS A VALID RESULT.

Local credentials.
Explicit authority.

01

The initial test bed

A synthetic browser invoice / finance sandbox is the initial MVP. No published benchmarks, completed security audits or certified real-world transactions are claimed.

02

What ANL is not

Not an agent replacement, universal decompiler, RPA robot or generic MCP generator. It is an in-progress capability qualification and outcome verification layer.