Cloud Platforms

Pluggable cloud platform knowledge modules. Each resource file contains platform-specific patterns for one cloud provider.

How It Works

1. Detect platform: Read the project's AGENTS.md tech stack declaration to identify which cloud platform(s) are in use
2. Load the right module: Read the corresponding resource file(s) below
3. Apply patterns: Use platform-specific CLI commands, service names, and operational patterns from the loaded module

Platform Detection Signals

Signal in AGENTS.md / Project	Platform	Resource File
GCP, Google Cloud, GKE, Cloud SQL, Cloud Build, gcloud, Artifact Registry	Google Cloud Platform	gcp-reference.md
Azure, AKS, Azure SQL, Azure DevOps, az CLI, ACR, Azure Monitor	Microsoft Azure	azure-reference.md
AWS, EKS, RDS, CodePipeline, aws CLI, ECR, CloudWatch	Amazon Web Services	aws-reference.md

If AGENTS.md does not declare a cloud platform, ask the user before proceeding with platform-specific operations.

If the project uses multi-cloud, load all relevant modules and apply patterns per service based on where it runs.

Module Structure

Each reference file follows a consistent structure:

1. Core Platform — account/project structure, IAM, networking, secrets, tagging
2. Container Orchestration — managed Kubernetes (GKE/AKS/EKS) patterns
3. Managed Database — managed SQL service patterns and reliability
4. CI/CD — platform-native build and deploy services
5. Observability — monitoring, logging, alerting, dashboards
6. CLI Reference — essential CLI commands for diagnostics and operations
7. Security and Compliance — audit logging, encryption, network security

Integration

• Used by roles: devops-engineer role, sre-engineer role, cloud-architect role
• Used by workflows: analyze-prod skill, bugfix skill, infra-change skill, deploy-production skill, ml-pipeline skill
• Complements: deployment-procedures skill (rollback procedures)