Observe Only
No meaningful authority or value exposure.
How much authority should you give an AI agent?
DN Agent Trust turns executed safety evidence into exposure classes, control scores and machine-readable risk profiles. Unclassified does not mean unsafe. It means DN does not yet have enough evidence to assign an exposure class.
5systems tracked
1classified
v1.3Risk Graph
provisionalsurveillance status
Public leaderboard
Machine-readable APIAgent trust profiles
| System | Exposure class | Evidence confidence | Verified controls | Evidence mode | |
|---|---|---|---|---|---|
| AlmanakAlmanak | E3: Meaningful Financial Authority | 99.8% | 62 | fork-onchain-open-position-executed | View profile |
| Virtuals EconomyOS / ACPVirtuals Protocol | Unclassified | 90% | 4 | documentation-only | View profile |
| Liquid Co-Invest / Co-Invest ComputerLiquid | Unclassified | 85% | 4 | documentation-only | View profile |
| Definitive FlashDefinitive | Unclassified | 80% | 4 | documentation-only | View profile |
| Coinbase AgentKitCoinbase | Unclassified | 78% | 7 | code-executed-upstream-unit-boundary-tests | View profile |
Low-Risk Informational
Read, analyze or recommend with minimal execution authority.
Bounded Commercial
Limited commercial actions inside tightly defined controls.
Meaningful Financial Authority
Material financial actions with verified limits, recovery and containment.
High-Assurance Autonomous Counterparty
Highest assurance tier. Requires deeper longitudinal and production evidence.
DN is measuring authority, not popularity.
A system can have a strong brand, large user base or impressive model and still be unsuitable for meaningful financial authority. DN separates identity, reputation, authorization, financial limits, recovery, containment and infrastructure evidence so one strong dimension cannot hide a failed critical control.