Assurance schema

SOV

Sovereign Operational Verification

A verification and assurance schema for AI agents, tools, skills and delegated capabilities.

SOV connects identity, declared authority, evidence and operational status into one inspectable record.

Sovrentic is building an inspectable assurance scheme for agentic resources.

Try the verification preview

SOV-1 Declaration is self-declared and not assessed. It is not a safety or compliance claim. The advanced SOV profiles are proposed scoped assessments.

01 / The SOV model

From declaration to continuous verification

  1. 01

    Declare

    Define what the resource is and what it is allowed to do.

  2. 02

    Verify

    Create a persistent identity, hash and public verification record.

  3. 03

    Assess

    Assess the resource against a defined SOV profile.

  4. 04

    Monitor

    Track changes, validity, revocation and reassessment triggers.

02 / Product family

Start with SOV-1. Grow into assurance.

Explore the SOV offer

SOV-1 Declaration

Self-declared · not assessed

Proposed · Draft SOV v0.2

Proposed free declaration profile for a defined agentic resource. Draft SOV v0.2, not yet frozen.

Structural checks on a record the owner declares: machine-readable identity and declared authority envelope. No assessment, behavioural competence or safety claim.

SOV-Agent — Proposed

Resource profile · proposed

Planned advanced assessment: a scoped assessment of an agent configuration under an identified programme, including delegated authority, tools and organisational data.

SOV-Tool — Proposed

Resource profile · proposed

Planned advanced assessment: a scoped assessment of a tool and its relevant contract and behaviour, including access scope, data handling and execution boundaries.

SOV-Skill — Proposed

Resource profile · proposed

Planned advanced assessment: a scoped assessment of a reusable skill and its declared dependencies and conditions.

SOV-Agent, SOV-Tool and SOV-Skill are resource profiles, not successive safety levels. SOV-C is a bespoke assurance and programme-design engagement, not a higher level.

Every SOV result states what was checked, what remains outside scope and what changes would require review. Scope, evidence and inclusions for each offer are set out on the Offer page, and the reasoning behind the schema is in the SOV technical paper.

03 / Illustrative record

What a SOV-1 Declaration looks like

SOV-1

Synthetic illustration

SOV-1 Declaration

self-declared · not assessed

Proposed · Draft SOV v0.2
Resource
Invoice Review Agent
Version
0.4.2
Resource type
AI Agent
Manifest hash
sha256:illustrative
Ledger status
Optional anchoring · not implemented
Assurance class
DECLARATION
Assessment status
NOT_ASSESSED
Profile identifier
SOV-1-FOUNDATION (frozen v0.1 identifier)

Synthetic illustration. Self-declared by the resource owner, not assessed or independently issued by Sovrentic. Public ledger anchoring is an optional integrity and transparency mechanism that is not implemented; it would not prove the truth of an assessment, and sensitive evidence is intended to remain off-chain.

04 / Binding rule

Every SOV credential is bound to a specific resource, version, purpose, authority envelope and operating context.

The SOV schema and all related product concepts are in development. No assurance record is issued today, and no payment flow, blockchain transaction, regulatory endorsement or accreditation claim is implied. See status and independence and what an SOV record cannot establish.

Start with the preview

Describe a resource. See how SOV would read it.

Try the verification preview