Skip to content

Site search

Type to search Pages

For relying agent operators

Your agents cannot act on a claim your governors can't audit.

When your agents transact across organisational boundaries, they must rely on claims they cannot re-verify from source. Signura lets them verify one signed attestation at decision time and keep the proof.

What running relying agents on the exchange gives you.

Verify one proof, not many

Your agent queries the exchange and verifies a signed attestation cryptographically, instead of re-authenticating each counterparty and re-checking its credentials from source on every transaction.

Tell valid from stale

Each attestation carries freshness and revocation semantics — extending the proven OCSP status pattern to assertions — so your agent distinguishes a currently valid claim from a stale one in real time.

A named party behind each claim

Every attestation names the issuer who stands behind it, with explicit scope and liability terms, so reliance rests on an accountable party rather than an anonymous source.

Reliance metered per event

Each reliance is metered per event and priced below a fresh from-source verification, because the checking work has already been done and signed for.

How reliance runs

From admitting an issuer to an audited reliance.

Admit the issuers

Your governors set which issuers and attestation types your agents may rely on, with scope, value thresholds and expiry configured up front.

Query and verify

At the point of decision, your agent queries the exchange and verifies the cryptographic proof rather than re-checking the claim from its source.

Check it still holds

The freshness and revocation engine confirms the attestation is current, not one that was valid only when it was last cached.

Proceed or escalate

Within the configured limits the agent proceeds; a reliance outside its scope, threshold or expiry is routed to a human governor for a decision.

Keep the record

The reliance event is logged to the evidence ledger, so you can demonstrate afterwards precisely what your agent relied on and when.

The choice you actually face when agents cross boundaries.

Two bad options, and a third

When an AI agent acting for your company reaches a counterparty controlled by another organisation, it hits a dependency it cannot resolve alone: it must rely on a claim the other side makes — that a licence is valid, a policy is in force, a certification holds, a permission has been granted — and it cannot pause the transaction to check that claim from source. Historically there have been two responses, and neither survives fiduciary governance. The agent can accept the claim blindly, which invites impersonation and fraud. Or it can decline to proceed, which rejects valid transactions out of over-caution. Re-verifying every claim pairwise is the theoretical alternative, but it scales combinatorially with the number of counterparties and is frequently impossible in the time a decision allows.

What reliance changes

Signura replaces that dilemma with a checkable step. An issuer publishes a signed, revocable attestation carrying explicit scope and a named liable party. Your agent fetches the cryptographic proof, verifies it at the moment it acts, and confirms through the freshness and revocation engine that the claim still holds. It then proceeds within the limits your governors have set, or escalates a case that falls outside them. The distinction matters to the people who remain accountable: a feeling of trust is not something they can inspect, whereas a verification is. Every reliance event is logged as a tamper-evident record — the claim, its scope, the liable party and the proof verified — so a governance or liability review later can see exactly what underpinned a decision.

Why now, and what it does not promise

The substrate for this is no longer speculative: the W3C Verifiable Credentials 2.0 data model became a formal Recommendation in May 2025, providing a stable format and a revocation mechanism. What has been missing is transferable reliance with someone standing behind the claim. Signura supplies that as a neutral exchange, sitting between issuers and relying agents and taking neither side's part. It does not remove your accountability, decide your transactions, or guarantee a legal outcome; it makes the evidence for a reliance decision available where the decision is made and retainable afterwards.

What operators of relying agents ask.

Which inter-agent protocol should we back before adding a reliance layer?

You do not have to pick a winner. Signura is built on the ratified W3C Verifiable Credentials 2.0 substrate and sits above identity and credentialing primitives, so it works alongside the protocols your agents already speak rather than replacing them.

If a winning protocol builds reliance in natively, are we stuck?

The reliance runs on the open Verifiable Credentials substrate, not a single proprietary protocol. Signura's defensibility comes from being the neutral party both sides rely on and from the accumulating record of reliance, not from locking you into one protocol.

Does relying on an attestation just move liability onto us?

Each attestation names the issuer who stands behind it, with explicit scope and liability terms the issuer approved, so the accountable party is stated rather than left implicit. Signura is neutral and does not guarantee a legal outcome; it makes who relied on what, and who stood behind it, inspectable.

How does an agent tell a currently valid claim from a stale one?

The freshness and revocation engine applies revocation and freshness semantics at decision time, so your agent verifies that an attestation still holds now rather than trusting a value that was valid only when it was last cached.

Can we bound what agents rely on without approving each one by hand?

The governance console lets your governors set admissible issuers, attestation types, value thresholds and expiry once, and receive escalations only on ambiguous or high-risk reliance. Consequential decisions stay with a human; routine reliance runs within the limits you configured.

Tell us where your agents are stuck relying on counterparties.

Describe one cross-boundary reliance your agents cannot make blind, and we will walk through how it would resolve on the exchange.

Describe the reliance decision your agents can't make blind.

Whether you deploy relying agents or govern what they depend on, tell us what they must rely on — a person reads it and replies.