Statute™
Statute is where the rules an organisation wants its agents to follow are authored, versioned and published as a constitution – and bound to what the foundation actually enforces, so policy and intent can't drift apart.
The rules an organisation wants its agents to follow live in documents, Slack threads and tribal knowledge. What's written down and what's enforced drift apart, and no one can point to the version that was in force when something happened.
Those rules are authored once, published as a versioned constitution, and bound to what the foundation actually enforces at evaluation time – so intent and enforcement stay the same thing, and every version is on the record.
Statute is how an organisation's intent enters the stack: a constitution the whole foundation evaluates against, versioned like code.
Statute treats governance the way engineering treats code: authored deliberately, versioned, published, and bound to the thing that enforces it.
Author & publish
Rules are authored in one place and published as a constitution – a single, reviewable artefact rather than scattered documents and config no one can trace.
Version
Every publish is a version. You can point to exactly which constitution was in force at any moment, and what changed between one and the next.
Bind to enforcement
The published constitution is what the foundation's policy engine evaluates against at call time – so a rule change is a versioned publish, not an untracked edit.
The value is the binding. A constitution isn't a document that hopes to be followed – it's the exact input the policy engine evaluates, versioned so intent and enforcement never quietly diverge.
constitution: acme-agents v3 finance agents: read ledgers, no transfers v4 finance agents: read ledgers, transfers ≤ $10k published 2026-06-30 · bound at evaluation # the foundation evaluates against the published version; # "what was in force" is always answerable.
Intent becomes a versioned, enforced artefact – not a rule someone remembers differently.
SYNTHERA is the trust layer for multi-agent systems: every agent gets a verifiable identity, scoped authority and a tamper-evident record, so software from different teams, vendors and frameworks can act on each other’s behalf without custom glue between every pair.