Control before
execution.
Eraivo is built around one principle: money should move only after the right checks, limits, approvals, and records are in place.
Permission Limits
Agents and wallets can be limited by amount, asset, counterparty, chain, time window, and action type.
Approval Gates
Sensitive or high-value actions can be held until the right people approve them.
Kill Switches
Operators can pause or revoke agents and wallets when behavior no longer matches the mandate.
Known Actor
Every request is tied to an agent, operator, wallet, or system identity.
Defined Mandate
Eraivo checks whether the actor is allowed to take the requested action before execution.
Evidence Requirements
Legal, escrow, settlement, and compliance workflows can require counterparties, purpose, consents, and supporting evidence.
Scoped Access
Access is limited to the capabilities and workflows each operator is allowed to use.
Pre-Execution Checks
Requests can be checked for policy, counterparty risk, evidence, approval status, and expected outcome before value moves.
Blocked Outcomes
Actions outside policy are rejected or held, and the reason is recorded.
Settlement Review
Settlement workflows can verify obligations and approvals before execution.
Complete Record
Every decision, approval, hold, rejection, and execution is available for review.
Trust the record, not the operator.
Hash-chained audit trail
Every audit event links to the previous one, so tampering with history is detectable. Anchor bundles can be committed on-chain and re-verified by anyone.
Policy commitments
A mandate hashes deterministically. The commitment can live in an on-chain registry, so counterparties can verify which policy governed an action.
Registry-gated execution
The protocol gateway only executes for registered operators, behind timelock governance — no single actor can silently change the rules.
Serverless verification
Accounts, policy hashes, escrow state, audit anchors, and settlement proofs stay verifiable directly against the chain, even without Eraivo servers.
Every action on the platform is written to an append-only audit log — timestamped, attributable, and retained for a minimum of 90 days.
Security reviews are part of the release process for production workflows.