theorycraft-skills

Architecture discovery and design skill — the front door to the theorycraft suite. Takes a user's idea ("I want to build X to do Y") and runs a deep Socratic Q&A to challenge assumptions and stress-test the idea before producing architecture documentation, diagrams, cost analysis, and T-shirt effort estimates. Q&A depth is adaptive — 20+ rounds for complex ideas, fewer for well-scoped ones. Output format is flexible based on complexity. Draws on theorycraft-cloud, theorycraft-azure, theorycraft-aws, theorycraft-gcp, theorycraft-kubernetes, and stack skills. Trigger whenever a user describes something they want to build, phrases an idea as "I want to build X" or "we need a system that does Y", asks for an architecture review, or needs a design doc or RFC. Trigger before any other theorycraft skill when the user is starting from an idea rather than a specific architectural question.

AWS-specific cloud architecture extension. Produces deep AWS service recommendations, AWS Well-Architected Framework analysis, and architecture diagrams (Mermaid for topology, SVG for detailed component diagrams). Extends theorycraft-cloud — run that skill's analysis first, then this skill adds AWS-specific service selection, IAM and VPC design, CDK/Terraform guidance, and visual diagrams. Use this skill whenever a user asks about AWS architecture, AWS service selection, AWS networking (VPC, Transit Gateway, PrivateLink, Direct Connect), AWS identity (IAM, IRSA, AWS Organizations), AWS data services, AWS security (GuardDuty, Security Hub, SCPs), or wants an AWS architecture diagram. Trigger for any question that is AWS-specific or where the user's estate is AWS-first.

Azure-specific cloud architecture extension. Produces deep Azure service recommendations, Azure Well-Architected Framework analysis, and architecture diagrams (Mermaid for topology, SVG for detailed component diagrams). Extends theorycraft-cloud — run that skill's analysis first, then this skill adds Azure-specific service selection, naming conventions, ARM/Bicep/Terraform guidance, and visual diagrams. Use this skill whenever a user asks about Azure architecture, Azure service selection, Azure networking (VNet, peering, Private Endpoints, ExpressRoute), Azure identity (Entra ID, Managed Identity, Workload Identity), Azure data services, Azure security (Defender, Sentinel, Policy), or wants an Azure architecture diagram. Trigger for any question that is Azure-specific or where the user's estate is Azure-first.

Cloud architecture theorycrafting — deeply analyses any cloud question, idea, or design challenge and produces a structured recommendation covering the best possible approach given constraints, anti-patterns to avoid, well-architected framework alignment, business size and use case fit, FinOps, security and threat modelling, and multi-cloud options where relevant. Use this skill whenever a user asks how to architect, design, or choose a cloud solution; asks "what's the best way to do X in cloud"; wants to evaluate a cloud technology or pattern; mentions AWS, Azure, GCP, Kubernetes, serverless, multi-cloud, cost optimisation, cloud security, or cloud migration; or is theorycrafting or stress-testing a cloud design idea. Trigger even for vague or early-stage ideas — the skill is built to handle ambiguity and draw out requirements.

Container and Kubernetes architecture extension focused on architecture diagrams and design patterns. Produces Mermaid topology diagrams and SVG component diagrams for container-based architectures. Extends theorycraft-cloud for cloud-level analysis and complements theorycraft-kubernetes for operational depth — this skill's focus is visual architecture representation and high-level design patterns for containerised systems. Use this skill when a user wants a container or Kubernetes architecture diagram, wants to visualise a microservices topology, wants to design a multi-service containerised architecture, asks about service mesh design, multi-cluster topology, GitOps pipeline architecture, or container networking layout. Trigger whenever the question involves drawing, designing, or visualising a Kubernetes or container-based architecture.

GCP-specific cloud architecture extension. Produces deep GCP service recommendations, Google Cloud Well-Architected Framework analysis, and architecture diagrams (Mermaid for topology, SVG for detailed component diagrams). Extends theorycraft-cloud — run that skill's analysis first, then this skill adds GCP-specific service selection, GCP networking (VPC, Shared VPC, Private Service Connect, Cloud Interconnect), GCP identity (IAM, Workload Identity, Service Accounts), GCP data services (BigQuery, Cloud SQL, Spanner, Bigtable), and visual diagrams. Use this skill whenever a user asks about GCP architecture, Google Cloud service selection, BigQuery, GKE, Cloud Run, Pub/Sub, Cloud Spanner, Vertex AI, or wants a GCP architecture diagram. Trigger for any question that is GCP-specific or where the user's estate is GCP-first.

IaaS (Infrastructure as a Service) architecture extension. Produces deep VM-based architecture recommendations, infrastructure design patterns, migration guidance, and architecture diagrams (Mermaid for topology, SVG for detailed infrastructure diagrams). Extends theorycraft-cloud — run that skill's analysis first, then this skill adds IaaS-specific depth: VM sizing and families, availability sets and zones, scale sets and autoscaling, storage for VMs, networking (NSGs, load balancers, bastion), Windows and Linux workload patterns, lift-and-shift migration strategy, and hybrid connectivity. Use this skill whenever a user asks about VM-based architectures, IaaS design, VM sizing, availability zones for VMs, scale sets, auto scaling groups, managed disks, bastion hosts, Windows Server workloads, lift-and-shift migrations, or hybrid cloud connectivity. Trigger for any architecture question where compute is VM-based rather than containerised or serverless.

Kubernetes expert skill covering the full lifecycle — architecture, design decisions, workload patterns, security, observability, operations, and troubleshooting. Use this skill whenever a user asks about Kubernetes cluster design, workload configuration, RBAC, networking (CNI, Ingress, Gateway API, Network Policies), autoscaling (HPA, VPA, KEDA), storage, secrets management, policy enforcement, observability stack design, upgrade strategy, day-2 operations, debugging pods or nodes, or any incident involving a Kubernetes cluster. Covers managed K8s (AKS, EKS, GKE) and self-hosted (k3s, kubeadm). Presents CNCF tool options in context rather than prescribing defaults — recommends based on the user's environment, existing tooling, and constraints. Also covers FinOps for Kubernetes (Reserved Instances, Spot/Preemptible, node pool cost optimisation) and provider selection decisions (AKS vs EKS vs GKE). Trigger even for vague K8s questions.

PaaS (Platform as a Service) architecture extension. Produces deep PaaS design recommendations, managed service selection guidance, and architecture diagrams (Mermaid for topology, SVG for detailed component diagrams). Extends theorycraft-cloud — run that skill's analysis first, then this skill adds PaaS-specific depth: managed database selection (RDS, Azure Flexible Server, Cloud SQL), managed application platforms (App Service, Elastic Beanstalk, App Engine), managed messaging (Service Bus, SQS, Pub/Sub), API management, managed caching, and the key trade-offs between PaaS control vs operational simplicity. Use this skill whenever a user asks about managed cloud services, PaaS architecture, App Service, Elastic Beanstalk, App Engine, managed databases, managed queues, API gateway design, or wants to understand the right level of managed service abstraction for their workload. Trigger for any question about using managed services over self-managed alternatives.

Platform engineering skill covering the full lifecycle — IDP design, developer experience, golden paths, CI/CD toolchains, platform operations, SLOs, DORA/SPACE metrics, and Team Topologies org design. Handles both internal developer platforms and external platform products (PaaS/CPaaS). Q&A first to understand platform context, then opinionated recommendations with diagrams. Works standalone or as an extension of theorycraft-architecture. Use this skill whenever a user asks about platform engineering, internal developer platforms, IDPs, Backstage, Port, golden paths, self-service infrastructure, developer experience, CI/CD pipeline design, platform SLOs, DORA metrics, toil reduction, Team Topologies, stream-aligned teams, or how to build and run a platform team. Trigger for any question about building a platform for engineers or for customers.

Serverless architecture extension. Produces deep serverless design recommendations, event-driven pattern analysis, cold start and scaling guidance, and architecture diagrams (Mermaid for flows, SVG for component diagrams). Extends theorycraft-cloud — run that skill's analysis first, then this skill adds serverless-specific depth: FaaS selection (Lambda, Azure Functions, Cloud Run, Cloud Functions), event-driven patterns (choreography vs orchestration), cold start mitigation, stateful serverless (Durable Functions, Step Functions, Workflows), observability for serverless, and FinOps for consumption-based billing. Use this skill whenever a user asks about serverless architecture, FaaS, event-driven design, Lambda, Azure Functions, Cloud Run, Step Functions, Durable Functions, API Gateway + Lambda patterns, or wants a serverless architecture diagram. Trigger for any question involving serverless, event-driven, or consumption-billed compute patterns.