Ops Deploy

Ops: Deploy

Deploy the application to remote environments.

Argument: $ARGUMENTS — environment (dev, staging, production) and optional target (frontend, backend, both; default: both)

Path Convention

• Frontend: Current project directory (.)
• Backend: ${BACKEND_DIR:-../backend-v2} — set BACKEND_DIR in .claude/settings.local.json if your backend is elsewhere

Safety

CRITICAL: Production deployments require explicit human confirmation before proceeding. Always ask for confirmation when $ARGUMENTS contains production.

Discovery

1. Read backend package.json to discover deploy:*, aws:signin:* scripts
2. Read frontend package.json to discover available scripts
3. Read .env.{environment} files to find GraphQL URLs for post-deploy verification

CI/CD Path (Preferred)

The standard deployment path is via CI/CD — pushing to environment branches triggers auto-deploy. Manual deployment instructions below are for when CI/CD is not suitable.

Frontend Deployment

EAS Update (OTA — over-the-air JavaScript update)

Use for JS-only changes (no native module changes).

1. Verify EAS CLI:

   eas whoami
   

   If not authenticated: eas login

2. Copy environment file:

   cp .env.{environment} .env
   

3. Deploy OTA update:

   STAGE={env} NODE_OPTIONS="--max-old-space-size=8192" eas update --auto --channel={env} --message="Manual update"
   

4. Verify deployment:

   eas update:list --branch {env} --limit 3
   

EAS Build (Native binary — only when app.config.ts or native modules change)

1. eas build --platform all --non-interactive --no-wait --profile={profile}
   

2. Check build status:

   eas build:list --limit 5
   

Backend Deployment

Full Deploy (Serverless Framework)

1. Verify AWS credentials (discover the profile from backend package.json or remote-routine ~/.aws/config):

   cd "${BACKEND_DIR:-../backend-v2}"
   aws sts get-caller-identity --profile {aws-profile}
   

   If this is an interactive local session and credentials are expired, refresh the backend's local AWS signin flow. In a headless or remote-routine session, use the environment-backed assume-role profile and do not start an SSO browser/device flow.

2. Deploy all functions:

   cd "${BACKEND_DIR:-../backend-v2}"
   bun run deploy:{env}
   

3. Verify (use the GraphQL URL from .env.{environment}):

   curl -sf {graphql-url} -X POST \
     -H "Content-Type: application/json" \
     -d '{"query":"{ __typename }"}' \
     -w "\nHTTP %{http_code} in %{time_total}s\n"
   

Single Function Deploy

Discover available function names from backend package.json deploy:function:* scripts:

cd "${BACKEND_DIR:-../backend-v2}"
FUNCTION_NAME={fn} bun run deploy:function:{env}

Post-Deploy Verification

After any deployment:

1. Health check the deployed environment (use ops-verify-health skill)
2. Check logs for errors in the first 5 minutes (use ops-check-logs skill)
3. Monitor Sentry for new issues (use ops-monitor-errors skill)
4. Run browser UAT smoke test against the deployed environment (use ops-browser-uat skill)

Report deployment result as a table: