Skip to main content

Quickstart: Zero-Trust Security via Access Policies

Enable authorization for Go AI with access policies.

This quickstart demonstrates how to secure agent-to-agent communication using Catalyst access policies. You'll run two apps — a caller that invokes a server running a Go AI agent graph — and then apply a deny policy to block unauthorized access.

note

This tutorial uses Catalyst Cloud to get started quickly with zero infrastructure setup. Catalyst Cloud uses a tunnel to route service invocation requests from the cloud to your local dev machine. For production or on-premises requirements, Diagrid also offers a self-hosted enterprise option where traffic stays within your private network.

You will learn how to:

  • Use Catalyst service invocation for agent-to-agent communication
  • Apply an access control configuration to enforce zero-trust authorization
  • Observe a 403 PermissionDenied response when a caller is blocked by policy
Caller to server through Catalyst service invocationCatalystCaller appcallerApp IDAccesspolicyserverApp IDGo AIserver
Caller to server through Catalyst service invocationCatalyst checks the access policy before it forwards the call. A denied call gets a 403 and never reaches the server.

Your apps run entirely on your local machine — you are not deploying code to Catalyst. diagrid dev run opens a secure tunnel between Catalyst Cloud and your local processes and registers each app with an ID. When the caller sends a service invocation request, it travels up to Catalyst Cloud (not directly to the server). Catalyst evaluates the access policy there, in the cloud: if the policy allows it, the request is forwarded back down through the tunnel to the local server; if the policy denies it, Catalyst returns a 403 PermissionDenied immediately and the request never reaches your local server process.

1. Prerequisites​

2. Log in to Catalyst​

diagrid login

Confirm your identity:

diagrid whoami

3. Create project files​

Create a new directory for the quickstart:

mkdir go-ai-access-policies
cd go-ai-access-policies

Create the app folders:

mkdir server
mkdir caller

server/go.mod​

Create server/go.mod with the following content:

module server

go 1.26.4

require (
	github.com/diagridio/go-ai v0.2.0
	github.com/diagridio/go-ai/adapters/langchaingo v0.1.1
	github.com/tmc/langchaingo v0.1.15-0.20251029190607-e35755df7084
)

server/main.go​

Create server/main.go with the following content. It runs a two-node Go AI agent graph behind a /research HTTP endpoint.

note

The graph's nodes call LangChainGo's fake model, which returns scripted replies in order, so the quickstart needs no LLM API key.

// Server app: exposes a go-ai agent graph via HTTP.
package main

import (
	"context"
	"encoding/json"
	"fmt"
	"log"
	"net/http"

	"github.com/diagridio/go-ai/adapters/langchaingo"
	"github.com/diagridio/go-ai/agent"
	"github.com/tmc/langchaingo/llms/fake"
)

// logged prints each node's reply as it runs.
func logged(name string, node agent.NodeFunc) agent.NodeFunc {
	return func(ctx context.Context, s agent.State) (agent.State, error) {
		out, err := node(ctx, s)
		if err == nil {
			fmt.Printf(">>> %s: %v\n", name, out["messages"])
		}
		return out, err
	}
}

// research builds the two-node graph for one topic. LangChainGo's fake LLM
// stands in for a real model: it returns its scripted replies in order, so no
// API key is needed.
func research(ctx context.Context, topic string) ([]any, error) {
	model := fake.NewFakeLLM([]string{
		fmt.Sprintf("Found 42 papers on '%s'", topic),
		fmt.Sprintf("Summary: AI diagnostics up 40%% for '%s'", topic),
	})
	node := func(name string) agent.NodeFunc {
		return logged(name, langchaingo.ModelNode(model,
			langchaingo.WithInputKey("topic"),
			langchaingo.WithOutputKey("messages"),
		))
	}

	graph, err := agent.NewGraph("research").
		AddNode("gather_sources", node("gather_sources")).
		AddNode("summarize", node("summarize")).
		AddEdge(agent.START, "gather_sources").
		AddEdge("gather_sources", "summarize").
		AddEdge("summarize", agent.END).
		WithReducer("messages", agent.AppendReducer).
		Compile()
	if err != nil {
		return nil, err
	}

	out, err := agent.Execute(ctx, graph, agent.State{"topic": topic}, agent.RunOptions{})
	if err != nil {
		return nil, err
	}
	messages, _ := out["messages"].([]any)
	return messages, nil
}

func writeJSON(w http.ResponseWriter, status int, body any) {
	w.Header().Set("Content-Type", "application/json")
	w.WriteHeader(status)
	json.NewEncoder(w).Encode(body)
}

func main() {
	http.HandleFunc("POST /research", func(w http.ResponseWriter, r *http.Request) {
		var req struct {
			Topic string `json:"topic"`
		}
		if err := json.NewDecoder(r.Body).Decode(&req); err != nil {
			writeJSON(w, http.StatusBadRequest, map[string]string{"error": err.Error()})
			return
		}
		messages, err := research(r.Context(), req.Topic)
		if err != nil {
			writeJSON(w, http.StatusInternalServerError, map[string]string{"error": err.Error()})
			return
		}
		writeJSON(w, http.StatusOK, map[string]any{"messages": messages})
	})

	http.HandleFunc("GET /health", func(w http.ResponseWriter, r *http.Request) {
		writeJSON(w, http.StatusOK, map[string]string{"status": "ok"})
	})

	log.Println("Server listening on :8002")
	log.Fatal(http.ListenAndServe(":8002", nil))
}

This server runs the graph in process with agent.Execute, which needs no Dapr workflow runtime. For durable execution, run the graph with goai.NewRunner instead, as shown in Durable Go agents.

caller/go.mod​

Create caller/go.mod with the following content:

module caller

go 1.22

caller/main.go​

Create caller/main.go with the following content. This app invokes the server's /research endpoint through Catalyst service invocation:

// Caller app: invokes the server app through Catalyst service invocation.
package main

import (
	"bytes"
	"encoding/json"
	"fmt"
	"io"
	"log"
	"net/http"
	"os"
	"time"
)

func getenv(key, fallback string) string {
	if v, ok := os.LookupEnv(key); ok {
		return v
	}
	return fallback
}

// Catalyst injects these environment variables via `diagrid dev run`
var (
	daprHTTPEndpoint = getenv("DAPR_HTTP_ENDPOINT", "http://localhost:3500")
	daprAPIToken     = getenv("DAPR_API_TOKEN", "")
	client           = &http.Client{Timeout: 30 * time.Second}
)

func writeJSON(w http.ResponseWriter, status int, body any) {
	w.Header().Set("Content-Type", "application/json")
	w.WriteHeader(status)
	json.NewEncoder(w).Encode(body)
}

func callServer(w http.ResponseWriter, r *http.Request) {
	var req struct {
		Topic string `json:"topic"`
	}
	if err := json.NewDecoder(r.Body).Decode(&req); err != nil {
		writeJSON(w, http.StatusBadRequest, map[string]string{"error": err.Error()})
		return
	}

	// Call the server app's /research endpoint via Catalyst service invocation.
	url := daprHTTPEndpoint + "/v1.0/invoke/server/method/research"
	payload, _ := json.Marshal(map[string]string{"topic": req.Topic})
	httpReq, _ := http.NewRequestWithContext(r.Context(), http.MethodPost, url, bytes.NewReader(payload))
	httpReq.Header.Set("dapr-api-token", daprAPIToken)
	httpReq.Header.Set("Content-Type", "application/json")

	fmt.Printf(">>> Calling server via Catalyst: %s\n", url)
	resp, err := client.Do(httpReq)
	if err != nil {
		writeJSON(w, http.StatusBadGateway, map[string]string{"error": err.Error()})
		return
	}
	defer resp.Body.Close()
	text, _ := io.ReadAll(resp.Body)
	fmt.Printf(">>> Response status: %d\n", resp.StatusCode)
	fmt.Printf(">>> Response body: %s\n", text)

	result := struct {
		StatusCode int `json:"status_code"`
		Body       any `json:"body"`
	}{resp.StatusCode, string(text)}
	if resp.StatusCode == http.StatusOK {
		result.Body = json.RawMessage(text)
	}
	writeJSON(w, http.StatusOK, result)
}

func main() {
	http.HandleFunc("POST /call-server", callServer)
	http.HandleFunc("GET /health", func(w http.ResponseWriter, r *http.Request) {
		writeJSON(w, http.StatusOK, map[string]string{"status": "ok"})
	})

	log.Println("Caller listening on :8001")
	log.Fatal(http.ListenAndServe(":8001", nil))
}

The caller uses DAPR_HTTP_ENDPOINT and DAPR_API_TOKEN, which are automatically injected by diagrid dev run. The service invocation URL follows the pattern:

{DAPR_HTTP_ENDPOINT}/v1.0/invoke/{target-app-id}/method/{endpoint}

dev.yaml​

Create dev.yaml in the project root:

version: 1
common:
  appLogDestination: console
apps:
  - appID: caller
    appDirPath: ./caller
    command: ["go", "run", "."]
  - appID: server
    appDirPath: ./server
    appPort: 8002
    command: ["go", "run", "."]
tip

The appPort field on the server tells Catalyst which local port the app listens on. This is required for apps that receive service invocation requests. The caller does not need appPort because it only sends requests.

4. Install dependencies​

Download the server's modules and build both apps, so diagrid dev run doesn't download anything while it starts them:

cd server && go mod tidy && go build ./... && cd ..
cd caller && go build ./... && cd ..

5. Run with default allow policy​

Start both apps with Catalyst Cloud:

diagrid dev run -f dev.yaml --project go-ai-access-qs --approve
tip

Wait for the log to show Server listening on :8002 and Caller listening on :8001 before proceeding.

Open a new terminal and invoke the caller:

curl -X POST http://localhost:8001/call-server \
  -H "Content-Type: application/json" \
  -d '{"topic": "AI in healthcare"}'

You should see a successful response:

{
  "status_code": 200,
  "body": {
    "messages": [
      "Found 42 papers on 'AI in healthcare'",
      "Summary: AI diagnostics up 40% for 'AI in healthcare'"
    ]
  }
}

By default, Catalyst allows all agent-to-agent communication. The request traveled from your local caller → Catalyst Cloud → through the tunnel → your local server. The caller successfully invoked the server's Go AI graph.

6. Apply a deny policy​

Stop the running application by pressing Ctrl+C, then disconnect the apps from Catalyst:

diagrid dev stop --id caller --project go-ai-access-qs
diagrid dev stop --id server --project go-ai-access-qs

Deny every caller on the server app:

diagrid app access grant server --project go-ai-access-qs --caller '*' --action deny --wait
tip

If the command reports local app connection already exists, wait a few seconds and run it again.

Passing '*' as the caller sets the policy's default action instead of naming one app, so this denies all incoming service invocation requests to the server, regardless of which app is calling it. Because authorization is enforced in Catalyst Cloud — not by the server process — a denied request never reaches your local server. The 403 is returned by Catalyst before any traffic enters the tunnel.

7. Verify access is denied​

Restart both apps:

diagrid dev run -f dev.yaml --project go-ai-access-qs --approve

Open a new terminal and invoke the caller again:

curl -X POST http://localhost:8001/call-server \
  -H "Content-Type: application/json" \
  -d '{"topic": "AI in healthcare"}'

This time, the caller receives a 403 PermissionDenied error:

{
  "status_code": 403,
  "body": "{\"errorCode\":\"ERR_DIRECT_INVOKE\",\"message\":\"failed to invoke, id: server, err: rpc error: code = PermissionDenied desc = access control policy has denied access to id: spiffe://<trust-domain>/ns/<project-id>/caller operation: research verb: POST\"}"
}

The access policy blocked the caller from invoking the server. The server's Go AI graph never executed.

8. Clean up​

Stop the running application by pressing Ctrl+C.

Delete the Catalyst Cloud project:

diagrid project delete go-ai-access-qs

Summary​

In this quickstart, you:

  • Set up two apps communicating through Catalyst service invocation
  • Verified that agent-to-agent calls succeed with the default allow policy
  • Applied a deny-all access policy that blocked the caller with a 403 PermissionDenied error

Next steps​

Last updated on