Adobe Deploy Integration Skill

Adobe Deploy Integration

Overview

Deploy Adobe-powered applications to three platforms: Vercel (serverless), Google Cloud Run (containers), and Adobe App Builder (native Adobe Runtime). Each with proper OAuth credential management.

Prerequisites

Adobe OAuth Server-to-Server credentials for production
Platform CLI installed (vercel, gcloud, or aio)
Application tested in staging environment

Instructions

Option A: Adobe App Builder (Native Adobe Hosting)

App Builder deploys serverless Runtime actions directly to Adobe infrastructure:

# Login to Adobe I/O CLI (requires IMS auth since AIO CLI v11)
aio login

# Select your project and workspace
aio console project select
aio console workspace select Production

# Deploy all actions, static assets, and event registrations
aio app deploy

# Check deployed actions
aio runtime action list

# View action logs
aio runtime activation list --limit 10
aio runtime activation logs <activationId>

// app.config.yaml — App Builder configuration
application:
  actions: actions
  web: web-src
  runtimeManifest:
    packages:
      my-adobe-app:
        actions:
          process-image:
            function: actions/process-image/index.js
            runtime: nodejs:20
            inputs:
              ADOBE_CLIENT_ID: $ADOBE_CLIENT_ID
              ADOBE_CLIENT_SECRET: $ADOBE_CLIENT_SECRET
            annotations:
              require-adobe-auth: true
              final: true

Option B: Vercel Deployment

# Set Adobe credentials as Vercel environment variables
vercel env add ADOBE_CLIENT_ID production
vercel env add ADOBE_CLIENT_SECRET production
vercel env add ADOBE_SCOPES production

# Deploy
vercel --prod

// vercel.json
{
  "functions": {
    "api/**/*.ts": {
      "maxDuration": 60
    }
  },
  "env": {
    "ADOBE_CLIENT_ID": "@adobe_client_id",
    "ADOBE_CLIENT_SECRET": "@adobe_client_secret",
    "ADOBE_SCOPES": "@adobe_scopes"
  }
}

// api/firefly/generate.ts — Vercel serverless function
import type { VercelRequest, VercelResponse } from '@vercel/node';
import { getAccessToken } from '../../src/adobe/client';

export default async function handler(req: VercelRequest, res: VercelResponse) {
  if (req.method !== 'POST') return res.status(405).end();

  try {
    const token = await getAccessToken();
    const fireflyResponse = await fetch(
      'https://firefly-api.adobe.io/v3/images/generate',
      {
        method: 'POST',
        headers: {
          'Authorization': `Bearer ${token}`,
          'x-api-key': process.env.ADOBE_CLIENT_ID!,
          'Content-Type': 'application/json',
        },
        body: JSON.stringify(req.body),
      }
    );

    const result = await fireflyResponse.json();
    return res.status(fireflyResponse.status).json(result);
  } catch (error: any) {
    return res.status(500).json({ error: error.message });
  }
}

Option C: Google Cloud Run

# Store credentials in Secret Manager
echo -n "${ADOBE_CLIENT_ID}" | gcloud secrets create adobe-client-id --data-file=-
echo -n "${ADOBE_CLIENT_SECRET}" | gcloud secrets create adobe-client-secret --data-file=-

# Build and deploy
gcloud builds submit --tag gcr.io/${PROJECT_ID}/adobe-service

gcloud run deploy adobe-service \
  --image gcr.io/${PROJECT_ID}/adobe-service \
  --region us-central1 \
  --platform managed \
  --set-secrets="ADOBE_CLIENT_ID=adobe-client-id:latest,ADOBE_CLIENT_SECRET=adobe-client-secret:latest" \
  --set-env-vars="ADOBE_SCOPES=openid,AdobeID,firefly_api" \
  --min-instances=1 \
  --timeout=60s

Health Check Endpoint (All Platforms)

// api/health.ts
export async function GET() {
  const checks: Record<string, any> = {};

  // Test Adobe IMS token generation
  try {
    const start = Date.now();
    const token = await getAccessToken();
    checks.adobe = {
      status: 'healthy',
      latencyMs: Date.now() - start,
      tokenLength: token.length,
    };
  } catch (error: any) {
    checks.adobe = {
      status: 'unhealthy',
      error: error.message,
    };
  }

  const overall = Object.values(checks).every(
    (c: any) => c.status === 'healthy'
  ) ? 'healthy' : 'degraded';

  return Response.json({
    status: overall,
    services: checks,
    timestamp: new Date().toISOString(),
  });
}

Output

Application deployed to chosen platform
Adobe credentials injected via platform secret management
Health check endpoint validates IMS connectivity
Serverless function timeout configured for Adobe API latency

Error Handling

| Issue | Cause | Solution | |-------|-------|----------| | aio app deploy auth error | Not logged in to AIO CLI | Run aio login | | Vercel function timeout | Adobe API takes > 10s | Increase maxDuration in vercel.json | | Cloud Run cold start timeout | Token generation on cold start | Set min-instances=1 | | Secret not found | Wrong secret name | Verify with gcloud secrets list or vercel env ls |

Resources

Next Steps

For webhook handling, see adobe-webhooks-events.

Agent Skills: Adobe Deploy Integration

Install this agent skill to your local

Skill Files