Separate incoming updates from agent work
You treat delivery and model execution as separate decisions. A resource subscription receives updates but never grants an agent permission to run.
Unit example (nearest native match)
See the idea in context
source := ax.NewAxMCPEventSourceWithPolicy(client, "inventory", "tenant:demo", "authenticated", ax.AxMCPSubscribeAll())
routes := []ax.AxEventRoute{{
ID: "resource-wake", Action: "wake",
TargetID: "inventory-agent", RequireAuthenticated: true,
Match: map[string]ax.Value{"types": ax.Array("mcp.resource.updated")},
}}
runtime, err := ax.NewAxEventRuntime(routes, nil)
if err != nil { panic(err) }
runtime.RegisterTarget(ax.AxEventTarget{
ID: "inventory-agent", RetrySafety: "idempotent",
MapInput: func(event ax.AxEventEnvelope, _ *ax.AxEventContinuation) (ax.Value, error) {
return map[string]ax.Value{"uri": event.Data.(map[string]ax.Value)["uri"]}, nil
},
Invoke: invokeAgent,
})
runtime.AddSource(source)
if err := runtime.Start(); err != nil { panic(err) }- Discover the endpoint
The live catalog identifies concrete resources and supported event capabilities.
- Apply subscription policy
Your application chooses which updates it actually wants.
- Route through the inbox
An explicit event route decides whether the update only records state or wakes work.
Run itIn your own project
go get github.com/ax-llm/ax/packages/go
import axllm "github.com/ax-llm/ax/packages/go"
client := axllm.NewAI("openai", map[string]axllm.Value{"apiKey": os.Getenv("OPENAI_API_KEY")})
classify := axllm.NewAx("review:string -> sentiment:class \"positive, negative, neutral\"", nil)Set OPENAI_APIKEY in your environment before running provider-backed code.
From a clone of the ax repo:
npm run example -- go src/examples/go/mcp/resource_wake_agent.goActive practice
Answer 2 in a row to learn this · attempt 1