Loading...
Loading...
Build and maintain a trust network of contacts, organizations, and stakeholders. Every agent carries a trust score that reflects its compliance history, reliability, and transparency. Trust scores power badge tiers, gate access to sensitive operations, and give downstream consumers a verifiable signal of agent quality.
Every newly registered agent starts with a base trust score of 50. The score is continuously adjusted based on three weighted components.
Compliance (40%)
Policy adherence, audit completeness, rule violations
Reliability (35%)
Uptime, error rate, response latency, heartbeat consistency
Transparency (25%)
Logging coverage, explainability, event reporting
// Trust score calculation (simplified)
trustScore = (
compliance * 0.40 +
reliability * 0.35 +
transparency * 0.25
)
// Score adjustments
+5 Policy evaluation passed (no violations)
+3 Heartbeat reported on schedule
+2 Full event logging for action
-10 Policy violation detected
-15 Enforcement action issued
-5 Missing heartbeat (per interval)
-3 Incomplete event loggingAgents earn a "PROTECTED BY DRD" badge tier based on their current trust score. Badge tiers unlock access to higher-privilege operations and signal trust to downstream consumers.
Bronze
Score 60-69Basic verification. Limited to read-only API access and standard policy evaluation.
Silver
Score 70-84Verified agent. Access to content registration, enforcement APIs, and webhook configuration.
Gold
Score 85-94Trusted agent. Full API access, priority rate limits, and ability to issue sub-agent credentials.
Government
Score 95-100Highest trust. Reserved for government-certified or auditor-verified agents. Regulatory fast-track.
Categorize contacts by engagement level to prioritize outreach and manage relationships.
Active engagement, high priority
Regular contact, moderate priority
Infrequent contact, low priority
Inactive or churned contacts
The reputation graph maps trust relationships between agents. When Agent A vouches for Agent B, a directed edge is created in the graph. Transitive trust is calculated using a dampened PageRank-style algorithm.
{
"ok": true,
"data": {
"agentId": "01956abc-...",
"trustScore": 87,
"connections": [
{
"agentId": "01956def-...",
"direction": "outbound",
"trustWeight": 0.92,
"vouchedAt": "2026-01-15T10:00:00Z"
},
{
"agentId": "01956ghi-...",
"direction": "inbound",
"trustWeight": 0.78,
"vouchedAt": "2026-01-20T08:30:00Z"
}
],
"transitiveReach": 42
}
}Inactive agents lose trust score over time. This ensures that only actively maintained and monitored agents retain high-privilege access.
| Inactivity Period | Score Penalty | Effect |
|---|---|---|
| 7 days | -2 per day | Warning notification sent |
| 14 days | -3 per day | Badge tier may downgrade |
| 30 days | -5 per day | Agent flagged for review |
| 60 days | Score frozen at 30 | Agent suspended automatically |
Suspended agents can rebuild their trust score through a structured compliance recovery program. Recovery is gradual and requires demonstrated good behavior.
Acknowledge Violation
Agent owner acknowledges the enforcement action and submits a remediation plan.
Probation Period
Agent operates in read-only mode for 7 days. All actions are logged but not enforced.
Supervised Restoration
Agent regains write access with enhanced monitoring. Trust score rebuilds at 2x normal rate.
Full Reinstatement
After 30 days of clean operation, the agent regains its previous badge tier (capped at Silver).
The public agent registry is a searchable directory of all registered and verified agents. No authentication is required to query the registry.
GET /api/v1/registry?search=content+scanner&tier=gold&limit=10
{
"ok": true,
"data": [
{
"agentId": "01956abc-...",
"name": "Content Scanner v2",
"description": "Scans uploaded content for policy compliance",
"trustScore": 87,
"tier": "gold",
"status": "active",
"verified": true,
"capabilities": ["content-analysis", "image-classification"],
"badgeUrl": "https://badge.drd.io/01956abc.svg"
}
],
"meta": { "total": 1, "cursor": null }
}Import trust signals from external identity and reputation systems. The Trust Connector bridges third-party verification into the DRD trust graph.
GitHub Verified
Import repository verification status and maintainer identity.
Google Cloud IAM
Map GCP service account roles to DRD trust levels.
Auth0 / Okta
Import identity provider trust signals and MFA status.
Custom OIDC
Connect any OIDC-compliant identity provider.
{
"source": "github",
"agentId": "01956abc-...",
"credential": {
"type": "github_verified",
"repositoryUrl": "https://github.com/acme/content-scanner",
"verificationToken": "ghp_..."
}
}
// Response
{
"ok": true,
"data": {
"imported": true,
"trustBoost": 8,
"newScore": 73,
"source": "github_verified"
}
}Receive real-time notifications when trust scores change. Configure alert rules to trigger on specific thresholds, tier changes, or anomalous score movements.
Alert Event Types
trust.score.changedtrust.tier.upgradedtrust.tier.downgradedtrust.decay.warningtrust.suspendedtrust.recovery.startedtrust.vouch.receivedtrust.vouch.revoked{
"agentId": "01956abc-...",
"rules": [
{
"event": "trust.tier.downgraded",
"channel": "slack",
"webhookUrl": "https://hooks.slack.com/services/..."
},
{
"event": "trust.score.changed",
"condition": "delta < -10",
"channel": "email",
"recipients": ["admin@acme.com"]
}
]
}Meeting
Call
Demo
Support
Audit
import { DRD } from '@drd/sdk';
const drd = new DRD({ apiKey: process.env.DRD_API_KEY });
// Add a contact
const contact = await drd.trustNetwork.createContact({
name: 'Jane Smith',
organization: 'Acme Corp',
email: 'jane@acme.com',
role: 'CTO',
temperature: 'hot',
});
// Log an interaction
await drd.trustNetwork.logInteraction({
contactId: contact.id,
type: 'meeting',
summary: 'Discussed compliance requirements for Q2',
followUpDate: '2026-03-01T10:00:00Z',
});
// Get pending follow-ups
const followUps = await drd.trustNetwork.getFollowUps();