---

name: deploy-staging description: Deploy latest code to staging, wait for CI/CD pipelines, and verify health. Use after committing changes that need to go live on staging.

Deploy to Staging

Check, trigger, and monitor staging deployments across both NeoMud repos. Verifies everything is healthy after deploy.

Instructions

$ARGUMENTS

Step 1: Check Pipeline Status

Check all active and recent deploy workflows across the NeoMud repo:

gh run list --limit 8 --json name,status,conclusion,headSha,createdAt \
  --jq '.[] | select(.name | test("Deploy|deploy")) | "\(.name) | \(.status) \(.conclusion) | \(.headSha[0:7]) | \(.createdAt[11:16])"'

And the NeoMud-Platform repo:

cd /c/Users/lbarnes/IdeaProjects/NeoMud-Platform && \
gh run list --limit 5 --json name,status,conclusion,headSha \
  --jq '.[] | select(.name | test("Deploy|deploy")) | "\(.name) | \(.status) \(.conclusion) | \(.headSha[0:7])"'

Step 2: Wait for In-Progress Pipelines

If any deploy workflows are in_progress, poll until completion:

RUNID=<id> && while true; do
  STATUS=$(gh run view $RUNID --json status,conclusion --jq '.status + " " + .conclusion')
  echo "$(date +%H:%M:%S) $STATUS"
  if echo "$STATUS" | grep -q "completed"; then break; fi
  sleep 20
done

Step 3: Handle Failures

If a deploy failed:

1. Check logs: gh run view <id> --log-failed 2>&1 | tail -20
2. Common issues:
   • SSH auth: VPS_SSH_KEY_STAGING secret may be wrong
   • SCP permission: Run ssh root@159.203.127.47 'chown -R neomud:neomud /srv/neomud-web-stage/ /srv/neomud-assets/'
   • Flaky test: Re-run with gh run rerun <id> --failed
   • GHCR pull denied: PAT on VPS needs read:packages + repo scopes
3. Re-run: gh run rerun <id> --failed

Step 4: Pull New Images on VPS (if needed)

If the deploy workflow doesn't auto-recreate containers:

ssh root@159.203.127.47 'su - neomud -c "cd ~/neomud-staging && docker compose pull && docker compose up -d --force-recreate"'

Step 5: Verify Staging Health

After deploy completes, verify all services:

# API health
curl -s https://stage-api.neomud.app/api/v1/health

# Web marketplace
curl -s -o /dev/null -w "web: %{http_code}\n" https://stage.neomud.app/

# Maker
curl -s -o /dev/null -w "maker: %{http_code}\n" https://stage-maker.neomud.app/

# Game server health
curl -s -o /dev/null -w "game: %{http_code}\n" https://stage.neomud.app/health

# Container status
ssh root@159.203.127.47 'docker ps --format "{{.Names}} {{.Status}}" | grep -E "stg|world"'

Step 6: Verify WASM Client Content

Check the deployed WASM has the latest code AND matches the local build:

# What does VPS neomud.js reference?
VPS_WASM=$(ssh root@159.203.127.47 'grep -o "[0-9a-f]\{20,\}\.wasm" /srv/neomud-web-stage/client/neomud.js')
echo "VPS references: $VPS_WASM"

# Does that file exist on VPS?
ssh root@159.203.127.47 "test -f /srv/neomud-web-stage/client/$VPS_WASM && echo 'EXISTS' || echo 'MISSING'"

# Does local build match?
LOCAL_WASM=$(grep -o '[0-9a-f]\{20,\}\.wasm' client/build/dist/wasmJs/productionExecutable/neomud.js 2>/dev/null)
echo "Local references: $LOCAL_WASM"
[ "$VPS_WASM" = "$LOCAL_WASM" ] && echo "✓ MATCH" || echo "✗ MISMATCH — VPS has stale code"

IMPORTANT: Never declare deployed unless VPS hash matches local build hash. Never manually SCP — fix the pipeline instead.

Staging URLs

• Marketplace: https://stage.neomud.app/
• API: https://stage-api.neomud.app/api/v1/health
• Maker: https://stage-maker.neomud.app/
• Game health: https://stage.neomud.app/health

VPS Access

• SSH: ssh root@159.203.127.47
• Deploy user: neomud
• Staging compose: /home/neomud/neomud-staging/
• Web files: /srv/neomud-web-stage/
• WASM client: /srv/neomud-web-stage/client/
• World bundles: /srv/neomud-assets/