v1.3.0: AI Agents and Missions

AI Agents, Missions, and the dispatch machinery connecting them.

Full notes: 1.3.0 on GitHub.

Agent entity

Registration, identity, role assignments and lifecycle endpoints, mirroring Runner's shape — same secured repositories, controllers, DTOs, role enum and assignment subclasses.

Missions

Configurable at organization, stack, namespace and module scope. Each carries a mission type and an optional sidecar name, and is created through the same REST, dashboard and Terraform channels as every other resource.

Three types ship: AutoDiagnose on failure, ApprovalRecommend when a job awaits approval, and SummarizeJob on success.

Dispatch

MissionMatcher and MissionDispatcher run as competing consumers triggered by saga events, walk the relevant missions, resolve matching agent supply, and dispatch. Each mission type has an explicit typed endpoint and DTO rather than a generic invoke shim.

Authorization

The supply/demand conjunction: the scope owner controls mission creation, the agent owner controls supply creation, and dispatch enforces both.

Agent process and MCP

An Agent process with a Claude sidecar connects over AgentHub (SignalR), carrying dispatch, heartbeat, log streaming and run lifecycle. An MCP server exposes resources, tools and prompts to the sidecar; license-gated to Enterprise.

Attribution

Agents authenticate as Service Principals with an additional agent_id claim, so every REST call an agent makes is attributable to that agent.

Snap CD

Intelligent GitOps for Infrastructure as Code. Automate, orchestrate, and scale your infrastructure deployments with confidence.


© 2026 Snap CD. All rights reserved.

An unhandled error has occurred. Reload 🗙