🌱 Add CLAUDE.MD

CLAUDE.md

@@ -0,0 +1,130 @@
+# CLAUDE.md
+
+This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
+
+## Project Overview
+
+This repository contains the **operator-controller**, which is the central component of Operator Lifecycle Manager (OLM) v1. OLM v1 is a Kubernetes operator lifecycle management system that provides APIs, controllers, and tooling for packaging, distributing, and managing Kubernetes extensions/operators.
+
+The project consists of two main components:
+- **operator-controller**: Manages ClusterExtension lifecycle (install, upgrade, uninstall)
+- **catalogd**: Serves operator catalogs and provides catalog content via HTTP API
+
+## Essential Commands
+
+### Build and Development
+```bash
+make build              # Build binaries for current GOOS/GOARCH
+make build-linux        # Build binaries for GOOS=linux
+make docker-build       # Build container images
+make generate           # Generate code (DeepCopy methods, etc.)
+make manifests          # Generate CRDs, RBAC, and other manifests
+make fmt                # Format Go code
+make lint               # Run golangci linter
+make verify             # Verify all generated code is up-to-date
+```
+
+### Testing
+```bash
+make test               # Run all tests (unit, e2e, regression)
+make test-unit          # Run unit tests only
+make test-e2e           # Run e2e tests (requires kind cluster)
+make test-experimental-e2e  # Run experimental e2e tests
+make test-regression    # Run regression tests
+make envtest-k8s-bins   # Download ENVTEST binaries for unit tests
+```
+
+### Local Development and Cluster Management
+```bash
+make kind-cluster       # Create local kind cluster
+make kind-load          # Load built images into kind cluster
+make kind-deploy        # Deploy to kind cluster
+make kind-clean         # Delete kind cluster
+make run                # Build, create cluster, and deploy (standard)
+make run-experimental   # Build, create cluster, and deploy (experimental)
+```
+
+### Release and Documentation
+```bash
+make release            # Run goreleaser (snapshot by default)
+make quickstart         # Generate release manifests and install scripts
+make crd-ref-docs       # Generate API reference documentation
+```
+
+## Architecture and Patterns
+
+### Core API Types
+- **ClusterExtension**: Represents an operator/extension to be installed
+- **ClusterCatalog**: References a catalog of available operators
+
+### Controller Architecture
+The system follows standard Kubernetes controller patterns:
+- Controllers use **controller-runtime** framework
+- **Reconciliation loops** manage desired vs actual state
+- **Finalizers** ensure proper cleanup during deletion
+- **Status conditions** track installation/upgrade progress
+
+### Key Components
+- **Resolver**: Determines which operator version to install based on constraints
+- **Applier**: Handles Helm-based installations and upgrades
+- **Authentication/Authorization**: Manages service account permissions and RBAC
+- **Catalog Metadata Client**: Fetches operator metadata from catalogs
+- **Bundle Utilities**: Processes operator bundle formats
+
+### Testing Strategy
+- **Unit tests**: Use ENVTEST with real Kubernetes APIs
+- **E2E tests**: Full cluster testing with kind
+- **Integration tests**: Test component interactions
+- **Regression tests**: Verify conversion between bundle formats
+
+## Development Environment
+
+### Required Tools
+- Go 1.24+
+- Docker or Podman for container builds
+- kubectl for cluster interaction
+- kind for local development clusters
+- make for build automation
+
+### Tool Management
+- Uses **bingo** for version-pinned tools (stored in `.bingo/`)
+- Tools include: kind, kustomize, controller-gen, golangci-lint, etc.
+
+### Code Generation
+Most Kubernetes resources are generated:
+- **CRDs** generated from Go struct definitions using controller-gen
+- **RBAC** manifests generated based on controller annotations
+- **DeepCopy methods** generated for API types
+- **Kustomize overlays** for different deployment configurations
+
+## Important Development Notes
+
+### Multi-Component Coordination
+- Changes often affect both operator-controller and catalogd
+- Shared utilities exist in `internal/shared/`
+- API definitions in `api/v1/` are used by both components
+
+### Testing Requirements
+- Always run `make verify` before submitting changes
+- E2E tests are crucial for Kubernetes integration
+- Unit tests use controller-runtime's envtest package
+- Coverage tracking enabled for both unit and e2e tests
+
+### Feature Gates
+The project uses feature gates for experimental functionality:
+- Standard vs Experimental builds via Go build tags
+- Feature flags defined in `internal/*/features/`
+- Different manifests for standard vs experimental deployments
+
+### Configuration Management
+- **Kustomize** used for manifest generation and customization
+- Multiple overlays: standard, experimental, e2e variants
+- Environment-specific configurations in `config/overlays/`
+
+### Container Images and Registries
+- Multi-arch container builds supported
+- Default registry: `quay.io/operator-framework`
+- Local development uses kind for loading images
+- E2E tests include registry setup for testing
+
+This is a complex, production-grade Kubernetes operator system requiring understanding of operator patterns, Helm integration, and catalog-based software distribution.