| name | vendix-cloud-operations |
| description | Vendix cloud operations guidance for using the private production runbook, SSH, and AWS CLI to inspect or administer cloud resources. Trigger: When consulting keys/README.md, using AWS CLI for Vendix cloud inventory, checking production resource locations, or deciding whether to use SSH versus AWS CLI.
|
| license | MIT |
| metadata | {"author":"rzyfront","version":"1.0","scope":["root"],"auto_invoke":["Using AWS CLI to inspect or administer Vendix cloud resources","Using SSH to inspect Vendix production infrastructure","Consulting or updating keys/README.md production runbook","Checking production resource locations, IPs, distributions, buckets, ECR, RDS, Route53, or Secrets Manager metadata"]} |
| allowed-tools | Read, Edit, Write, Glob, Grep, Bash |
Vendix Cloud Operations
Purpose
Use this skill when an agent needs to operate or inspect Vendix production cloud resources through the private runbook, AWS CLI, or SSH.
This skill does not replace vendix-ec2-maintenance. Use this skill for cloud inventory, resource orientation, and choosing the right operational tool. Use vendix-ec2-maintenance for EC2 incidents, deploy failures, disk pressure, Docker cleanup, or server-level troubleshooting.
Core Rules
- Treat
keys/README.md as the private operational runbook when it exists locally.
- Do not copy private runbook snapshots into committed files.
- Do not store secret values in skills, AGENTS files, tickets, PRs, or chat responses.
- Prefer AWS CLI for resource discovery before hardcoding IPs, distribution IDs, bucket names, or endpoints.
- Prefer SSH only for host-local state: Docker containers, Nginx, disk, logs, ports, and local health checks.
- If a value can drift, resolve it live with AWS CLI and state when it was verified.
- Keep cloud operations read-first. Ask before making production changes.
Source Map
| Need | Primary Source |
|---|
| Current private operational snapshot | keys/README.md |
| Backend deploy behavior | .github/workflows/deploy-backend-ec2.yml |
| Frontend deploy behavior | .github/workflows/deploy-s3.yml |
| EC2 instance metadata and public IP | aws ec2 describe-instances |
| DNS records | aws route53 list-resource-record-sets |
| CloudFront aliases and origins | aws cloudfront list-distributions / get-distribution |
| Frontend and asset buckets | aws s3api list-buckets, get-bucket-website, get-bucket-location |
| Backend image repository and tags | aws ecr describe-repositories, describe-images |
| Database metadata | aws rds describe-db-instances |
| Secret inventory by name only | aws secretsmanager list-secrets |
| Email identity metadata | aws ses list-identities, get-identity-verification-attributes |
| Cloud metrics and log groups | aws cloudwatch, aws logs describe-log-groups |
| Runtime host state | SSH + systemctl, ss, nginx -T, docker, curl localhost |
Tool Selection
| Situation | Use |
|---|
| Need the current EC2 public IP | AWS CLI EC2 query |
| Need to know whether DNS points to EC2 or CloudFront | Route53 + dig |
| Need CloudFront distribution, aliases, origins, or deploy status | AWS CLI CloudFront |
| Need frontend deploy target | Workflow + S3/CloudFront AWS CLI |
| Need backend deploy target or current image tags | Workflow + ECR AWS CLI |
| Need DB engine/version/public exposure metadata | AWS CLI RDS |
| Need names of production secrets | AWS CLI Secrets Manager list operation |
| Need cloud metrics or log group inventory | AWS CLI CloudWatch / Logs |
| Need container status, logs, disk, Nginx, or localhost health | SSH to EC2 |
| Need EC2 deploy failure diagnosis or cleanup | vendix-ec2-maintenance |
Standard Workflow
- Read the relevant source map entry before acting.
- If
keys/README.md exists, use it for the latest local runbook and verified command patterns.
- Use AWS CLI to confirm cloud resource identity and current location.
- Use SSH only after cloud-level discovery shows the target host.
- For any production write/change, summarize target, command, expected effect, and rollback path before running it.
- After updating the private runbook, verify it avoids secret values and stale hardcoded facts.
AWS CLI Patterns
Use focused queries and projected output so responses stay safe and readable:
aws ec2 describe-instances \
--region us-east-1 \
--filters Name=instance-state-name,Values=running,stopped \
--query 'Reservations[].Instances[].{InstanceId:InstanceId,Name:Tags[?Key==`Name`]|[0].Value,State:State.Name,PublicIp:PublicIpAddress,InstanceType:InstanceType,KeyName:KeyName}' \
--output table
aws cloudfront list-distributions \
--query 'DistributionList.Items[].{Id:Id,DomainName:DomainName,Status:Status,Enabled:Enabled,Aliases:Aliases.Items}' \
--output table
aws ecr describe-images \
--region us-east-1 \
--repository-name vendix-backend \
--query 'sort_by(imageDetails,& imagePushedAt)[-5:].{tags:imageTags,pushed:imagePushedAt,digest:imageDigest}' \
--output json
aws secretsmanager list-secrets \
--region us-east-1 \
--filters Key=name,Values=vendix/production \
--query 'SecretList[].{Name:Name,LastChangedDate:LastChangedDate,Description:Description}' \
--output table
aws logs describe-log-groups \
--region us-east-1 \
--query 'logGroups[].{Name:logGroupName,Retention:retentionInDays,StoredBytes:storedBytes}' \
--output table
For this skill, use Secrets Manager list/describe style operations only. Do not include commands that print stored secret payloads in cloud inventory or runbook guidance.
SSH Patterns
Resolve the target host from AWS before connecting. Use SSH for local runtime facts:
sudo docker ps --format 'table {{.Names}}\t{{.Image}}\t{{.Status}}\t{{.Ports}}'
sudo docker logs vendix-backend --tail 100
sudo systemctl is-active nginx
sudo ss -tulpn | grep -E ':(80|443|3000|6379)'
curl -fsS http://localhost:3000/api/health
Do not make SSH the first discovery step when AWS CLI can identify the resource directly.
Runbook Updates
When updating keys/README.md:
- Replace stale facts with "resolve with AWS CLI" commands where possible.
- Include a "last verified" timestamp for snapshots.
- Keep actual secret values out.
- Mention whether a value was verified by AWS CLI, SSH, DNS, or HTTP.
- Keep the runbook operational and copy-paste friendly.
Related Skills
vendix-ec2-maintenance - EC2 incidents, deployment failures, disk pressure, and Docker cleanup.git-workflow - Branching, commits, and PR rules.skill-sync - Required after creating or changing skills.
