name	deploy-guide
description	Guide user through actual deployment steps for their application. This skill should be used when a project is ready to deploy to production. Walks through pre-deployment checks, deployment execution, and post-deployment verification. Supports VPS/Docker, Cloudflare Pages, fly.io, and Hostinger Shared Hosting.
allowed-tools	Read, Grep, Glob, Write, Bash, WebFetch

deploy-guide

Guide the user step-by-step through deploying their application to the chosen hosting target. Performs pre-deployment checks, provides deployment commands, and verifies successful deployment. Creates deployment documentation for future reference. BUILDER role with GUIDE approach. Executes deployment steps with user confirmation. - WILL run deployment commands (with user approval) - WILL perform pre-deployment checks - WILL verify deployment success - WILL create deployment documentation - WILL troubleshoot common deployment issues - Deployed application - .docs/deployment-log.md (deployment record and runbook) - Post-deployment verification results

Understand deployment target and project state.

Read .docs/deployment-strategy.md if it exists to understand: - Chosen deployment target - Deployment workflow - Environment configuration

If not present, gather information conversationally.

"I can see from your deployment strategy that you're deploying to {target}.

Let me verify your project is ready for deployment, then we'll proceed step by step."

"I don't see .docs/deployment-strategy.md. No problem - let me understand your deployment target.

Where are you deploying?

VPS with Docker - Your Hostinger VPS using Docker Compose
Cloudflare Pages - Static/JAMstack deployment
Fly.io - Containerized full-stack deployment
Hostinger Shared Hosting - PHP + MySQL deployment

Which target? [1/2/3/4]"

Based on target, gather: - VPS: Hostname/IP, SSH username, project path - Cloudflare: Project name, build command, output directory - Fly.io: App name, region preference - Shared: FTP/SSH credentials, remote path

Verify project is ready for deployment. Pre-deployment Verification:

Code Readiness:

All changes committed to git
Working branch merged to main (or deploy branch)
No uncommitted changes
Build passes locally

Configuration:

Environment variables documented
Production config separate from development
Secrets not committed to git

Testing (if applicable):

Tests passing
No critical bugs open

Infrastructure:

Target environment accessible
Required services running (database, etc.)
DNS configured (if first deployment)

```bash # Check git status git status

Verify on correct branch

git branch --show-current

Check build passes

{build_command}

Verify tests pass (if configured)

{test_command}

</verification-commands>

<prompt-to-user>
Running pre-deployment checks...

{checklist_results}

**Issues Found:** {count}
{issue_details}

Ready to proceed with deployment? [yes/no/fix issues]
</prompt-to-user>
</phase>

<phase id="2" name="deploy">
<action>Execute deployment based on target.</action>

<deployment-targets>

<vps-docker>
<name>VPS with Docker (Hostinger)</name>

<pre-steps>
1. Ensure Docker Compose file is ready
2. Verify SSH access to VPS
3. Confirm project directory exists on VPS
</pre-steps>

<deployment-process>
**Step 1: Connect to VPS**
```bash
ssh {username}@{host}

Step 2: Navigate to project

cd /var/www/{project_name}

Step 3: Pull latest code

git pull origin main

Step 4: Build and restart containers

docker compose pull
docker compose up -d --build

Step 5: Verify containers running

docker compose ps

Step 6: Check application logs

docker compose logs --tail=50

Step 7: Clean up

docker system prune -f

For first-time deployment:

Create project directory:

sudo mkdir -p /var/www/{project_name}
sudo chown {username}:{username} /var/www/{project_name}

Clone repository:

cd /var/www/{project_name}
git clone {repo_url} .

Create production .env:

cp .env.example .env
nano .env  # Configure production values

Configure Caddy (reverse proxy):

{domain} {
    reverse_proxy localhost:{port}
}

Reload Caddy:
```
sudo systemctl reload caddy
```

Cloudflare Pages

Ensure Cloudflare account connected
Verify build configuration
Confirm project name

**Option A: Git-Connected (Automatic)**

If connected to GitHub:

Push to main branch
Cloudflare automatically deploys
Monitor build in Cloudflare dashboard

Option B: Direct Deploy (Manual)

Using Wrangler CLI:

# Install/update Wrangler
npm install -g wrangler

# Login to Cloudflare
wrangler login

# Build project
{build_command}

# Deploy
wrangler pages deploy {output_dir} --project-name={project_name}

For first-time deployment:

Create project in Cloudflare Pages dashboard
Connect to GitHub repository (recommended)
Configure build settings:
- Build command: npm run build
- Build output directory: out or dist or .next
Set environment variables in dashboard

Fly.io

Ensure fly.toml exists
Verify Fly CLI installed and authenticated
Confirm app created

**Step 1: Verify fly.toml** Ensure fly.toml is configured correctly.

Step 2: Deploy

flyctl deploy

Step 3: Monitor deployment

flyctl logs

Step 4: Verify running

flyctl status

Step 5: Open application

flyctl open

For first-time deployment:

Install Fly CLI:
```
curl -L https://fly.io/install.sh | sh
```
Authenticate:
```
flyctl auth login
```
Create app:
```
flyctl apps create {app_name}
```
Create fly.toml (or use flyctl launch)

Set secrets:

flyctl secrets set DATABASE_URL="..."
flyctl secrets set SECRET_KEY="..."

Create database (if needed):

flyctl postgres create
flyctl postgres attach

Hostinger Shared Hosting

Ensure FTP/SSH credentials available
Verify remote path
Confirm PHP version compatibility

**Option A: Using Git (if SSH available)**

# SSH to server
ssh {username}@{host}

# Navigate to public_html
cd ~/public_html/{subdirectory}

# Pull latest code
git pull origin main

# Install dependencies (if composer)
composer install --no-dev --optimize-autoloader

Option B: Using rsync

rsync -avz --delete \
  --exclude='.git' \
  --exclude='.env' \
  --exclude='node_modules' \
  ./ {username}@{host}:~/public_html/{subdirectory}/

Option C: Using FTP

Use FileZilla or similar:

Connect to FTP server
Navigate to public_html
Upload files (excluding .env, node_modules, .git)

For first-time deployment:

Create subdirectory (if not root):
```
mkdir -p ~/public_html/{subdirectory}
```

Create .htaccess (for PHP routing):

RewriteEngine On
RewriteCond %{REQUEST_FILENAME} !-f
RewriteCond %{REQUEST_FILENAME} !-d
RewriteRule ^(.*)$ index.php [QSA,L]

Configure database in cPanel
Create production .env on server

Set correct file permissions:

find . -type f -exec chmod 644 {} \;
find . -type d -exec chmod 755 {} \;

For each deployment step: 1. Show the command about to run 2. Ask for confirmation before executing 3. Show output 4. Verify success before proceeding 5. Offer to stop if errors occur

Verify deployment was successful.

**Accessibility:** - [ ] Application loads at expected URL - [ ] No 500 errors on main pages - [ ] Static assets loading correctly

Functionality:

Authentication works (if applicable)
Database connections working
API endpoints responding

Performance:

Reasonable load time
No console errors
SSL certificate valid

```bash # Check HTTP response curl -I https://{domain}

Check SSL certificate

curl -vI https://{domain} 2>&1 | grep "SSL certificate"

Check specific endpoints

curl https://{domain}/api/health

</verification-commands>

<prompt-to-user>
Verifying deployment...

{verification_results}

**Status:** {SUCCESS/ISSUES_FOUND}
{details}

{if success}
Your application is live at: https://{domain}

{if issues}
Issues detected. Would you like help troubleshooting? [yes/no]
</prompt-to-user>
</phase>

<phase id="4" name="create-deployment-log">
<action>Create .docs/deployment-log.md documenting the deployment.</action>

<deployment-log-template>
```markdown
# Deployment Log

## Latest Deployment

**Date:** {date}
**Target:** {deployment_target}
**Branch:** {branch}
**Commit:** {commit_hash}
**Deployed by:** {user}
**Status:** SUCCESS

### Pre-Deployment Checks
- [x] Code committed and pushed
- [x] Build passed locally
- [x] Tests passing
- [x] Environment configured

### Deployment Steps Executed
1. {step_1}
2. {step_2}
3. {step_3}

### Post-Deployment Verification
- [x] Application accessible
- [x] No errors in logs
- [x] Core functionality working

### URLs
- **Production:** https://{domain}
- **API:** https://{domain}/api

---

## Deployment Runbook

### Regular Deployment

```bash
{deployment_commands}

Rollback Procedure

{rollback_commands}

Environment Variables

Variable	Description	Where to Set
{var_1}	{desc}	{location}
{var_2}	{desc}	{location}

Deployment History

Date	Commit	Status	Notes
{date}	{hash}	SUCCESS	Initial deployment

Troubleshooting

Common Issues

Application not loading:

Check container status: docker compose ps
Check logs: docker compose logs
Verify Caddy config

Database connection failed:

Verify DATABASE_URL in .env
Check database container running
Test connection manually

SSL certificate issues:

Caddy auto-generates certificates
Check Caddy logs: sudo journalctl -u caddy
Verify DNS pointing to server

</deployment-log-template>
</phase>

<phase id="5" name="summarize">
<action>Provide deployment summary and next steps.</action>

<summary-template>
## Deployment Complete

**Application:** {project_name}
**Target:** {deployment_target}
**URL:** https://{domain}
**Status:** SUCCESS

---

### Deployment Record

Created: .docs/deployment-log.md

This file contains:
- Deployment runbook for future deployments
- Rollback procedure
- Environment variables reference
- Troubleshooting guide

---

### Workflow Status

**TERMINATION POINT - MANUAL DEPLOYMENT**

Your application is deployed. You can stop here if you don't need CI/CD automation.

**Next Options:**
1. **Stop here** - Use .docs/deployment-log.md for future manual deployments
2. **Add CI/CD** - Use **ci-cd-implement** skill to automate deployments

---

### For Future Deployments

Quick deployment:
```bash
{quick_deploy_command}

See .docs/deployment-log.md for full runbook.

Monitoring Recommendations

Check logs regularly: {log_command}
Monitor uptime with external service
Set up alerts for errors

Congratulations on your deployment!

**Container won't start:** ```bash # Check logs docker compose logs {service_name}

Rebuild from scratch

docker compose down docker compose build --no-cache docker compose up -d


**Port already in use:**
```bash
# Find process using port
sudo lsof -i :{port}

# Kill process or change port in docker-compose.yml

Out of disk space:

# Clean up Docker
docker system prune -a --volumes

**Application not accessible:** 1. Check if container is running: `docker compose ps` 2. Test locally: `curl localhost:{port}` 3. Check Caddy/reverse proxy logs 4. Verify firewall allows traffic: `sudo ufw status`

SSL certificate not working:

Verify DNS points to server
Check Caddy logs: sudo journalctl -u caddy -f
Wait for certificate propagation (up to 15 minutes)

**Connection refused:** 1. Check database container: `docker compose ps db` 2. Verify DATABASE_URL format 3. Check network: containers must be on same Docker network

Permission denied:

Verify user credentials in .env
Check database user has required permissions

**Build failed:** 1. Check build command matches local 2. Verify Node.js version in dashboard 3. Check build logs in Cloudflare dashboard

Environment variables:

Set in Cloudflare Pages dashboard
Redeploy after changing

**Deploy failed:** 1. Check fly.toml configuration 2. Verify app name matches 3. Check resource limits (memory, CPU)

App crashing:

Check logs: flyctl logs
Verify health check endpoint
Check memory usage: flyctl status

- Run pre-deployment checks before deploying - Ask for confirmation before executing commands - Verify deployment success - Create deployment-log.md documentation - Handle errors gracefully with troubleshooting guidance - Provide rollback instructions

- Deploy without user confirmation - Skip pre-deployment checks - Leave failed deployment without troubleshooting help - Expose sensitive credentials in logs or documentation - Skip post-deployment verification

Phase 5 of 7: Deployment

Status: Phase 0: Project Brief (project-brief-writer) Phase 1: Tech Stack (tech-stack-advisor) Phase 2: Deployment Strategy (deployment-advisor) Phase 3: Project Foundation (project-spinup) <- TERMINATION POINT (localhost) Phase 4: Test Strategy (test-orchestrator) - optional Phase 5: Deployment (you are here) <- TERMINATION POINT (manual deploy) Phase 6: CI/CD (ci-cd-implement) <- TERMINATION POINT (full automation)

Phase 5 of 7 in the Skills workflow chain. Expected input: Deployable project, .docs/deployment-strategy.md (gathered conversationally if missing) Produces: Deployed application, .docs/deployment-log.md

This is a TERMINATION POINT for projects not needing CI/CD automation.

This skill can be invoked standalone on any deployable project. It gathers deployment target information conversationally if not available in handoff documents.

- When project development is complete (or MVP ready) - When user is ready to deploy to production - For subsequent deployments (use deployment-log.md as runbook)

Users can invoke the **workflow-status** skill at any time to: - See current workflow progress - Check which phases are complete - Get guidance on next steps - Review all handoff documents

Mention this option when users seem uncertain about their progress.

deploy-guide

Install Skill

SKILL.md

deploy-guide

Verify on correct branch

Check build passes

Verify tests pass (if configured)

Check SSL certificate

Check specific endpoints

Rollback Procedure

Environment Variables

Deployment History

Troubleshooting

Common Issues

Monitoring Recommendations

Rebuild from scratch