Agent Operations Platform
Audit
Every governed action — tool call, model call, lifecycle change, policy decision — emits an immutable event with the full context that produced it. The vocabulary is uniform across the three control points, which means a single query can reconstruct an incident regardless of where it happened.
Workflow
Capture an action.Stream it everywhere.
The same loop applies to every event the platform produces. Captured at the source, enriched with context, streamed in real time, and exported to wherever your operators, SOC, or compliance team already lives.
- 01
Audit
Every governed action emits an immutable event at the source — Registry, Gateway, or Broker.
- 02
Observe
Query, tail, and filter from the UI, CLI, or API. Full caller chain on every event.
- 03
Stream
Subscribe to the live event stream for real-time monitoring, alerts, and downstream automation.
- 04
Export
Project events to your SIEM, lakehouse, or compliance evidence store. Dome remains the system of record.
Consistent operations
One vocabulary, three control points
A unified event model is what turns three independent enforcement points into a single accountable system. The Registry, Gateway, and Broker each emit their own event classes, but the shape is the same — same caller chain, same decision trace, same export targets.
Each control point emits a typed event at the moment of decision or transition.
Caller chain, scope, policy version, and request digest attached automatically.
Live tail, query, and subscription APIs serve the events in real time.
Projected to SIEM, lakehouse, or compliance evidence store — Dome stays the system of record.
Unified vocabulary
agent.* events from the Registry, tool.* from the Gateway, llm.* from the Broker, policy.* across all three. One schema. One query language. One audit story per incident.
Caller chain
Every event records the agent, the end-user act-as claim, the parent agent if chained, and the rule version that produced the decision. Reconstruct any path through the system.
Multiple sinks
Query from the UI or CLI, subscribe via the API for real-time downstream automation, or export to Splunk, Snowflake, S3, or any sink your operations team already runs.
Options
What audit lets you do
Beyond capturing events, the audit layer surfaces operational controls — what you can configure via the CLI, API, or Terraform to shape how events flow and persist.
Real-time subscriptions
Stream events as they happen via gRPC or webhook. SOC platforms and SOAR playbooks can trigger off Dome events without polling.
Query-time filtering
Filter by agent, workspace, decision, time window, or full-text. Save common queries; subscribe to filtered streams.
Multi-sink export
Mirror events to Splunk, Microsoft Sentinel, Datadog, Snowflake, BigQuery, or S3. Configure transformation and retention per sink.
Retention policies
Hot for query, warm for compliance, cold for archive. Per-tenant retention windows with automatic tiering and lifecycle.
Compliance evidence packs
Generate signed evidence bundles for any time window — useful for SOC 2, ISO 27001, and customer audit responses.
SDK-emitted events
Agents can emit custom events from inside the process via the SDK. These flow through the same pipeline as platform-emitted events, with the same caller chain attached.
In practice
From query to export
Audit is queryable from the UI, the CLI, and the API. The same events can be exported to SIEM, lakehouse, or compliance evidence stores.
# What did this agent do recently?
dome audit query --agent support-triage --since 1h
# Live tail
dome audit stream --agent support-triage
# Filter by decision
dome audit query \
--workspace support \
--decision deny \
--since 24h# Export to Splunk
dome audit export splunk \
--hec https://splunk.example.com/services/collector \
--token $SPLUNK_HEC_TOKEN
# Stream to Snowflake or S3
dome audit export s3 \
--bucket dome-audit-prod \
--prefix workspace=support/Caller chain
Full context on every event
An audit event records the full caller chain: this agent, on behalf of this user, calling this tool, with these arguments, against this rule version, with this outcome. That is the level of detail compliance and incident response need.
{
"event": "tool.call",
"ts": "2026-06-02T14:03:22.418Z",
"agent": {
"id": "agt_3mw7kp",
"name": "support-triage",
"workspace": "support"
},
"actas": {
"subject": "jane.smith@example.com",
"verified_by": "oidc"
},
"tool": "zendesk/update_ticket",
"arguments_digest": "sha256:9b...",
"decision": "allow",
"policy": {
"version": "v23",
"rule": "workspace/support/zendesk.cedar#L12"
}
}