AI-native Kubernetes ops with Starling
Bridge the SRE context gap. Starling lets Claude, Cursor, and other AI agents scan manifests, read live clusters, and reason over pods, events, and metrics in real time.
# Install Starling CLI
brew install warble-tech/tap/starling
# Authenticate device
starling login
# Run MCP Server locally against current context
starling mcp --mode=readonly
// Server initialized. Listening on stdio...
CLI Interface
Starling provides a unified client-side interface to authorize, debug, and run specialized MCP servers against your Kubernetes environments.
Sign in via browser (device grant flow)
Cost metric: Free
$ starling login
Multi-Engine Capabilities
Five dedicated MCP servers bundled inside a single command-line companion.
Kubernetes (SteadyHelm)
Inspect namespaces, pods, container logs, HPA metrics, and events. Translate natural language commands into live API actions.
Observability (Prometheus)
Translate natural language queries into valid PromQL, fetch alert states, and introspect active rule definitions.
Supply Chain (Trivy)
Scan containers for CVEs, audit Terraform & Kubernetes configs for misconfigurations, detect secrets, and export findings.
Infrastructure (OpenTofu)
Analyze plans, inspect terraform state, summarize drift, and hook into cost estimation interfaces (Infracost) seamlessly.
ArgoCD Bridge
Connect GitOps workflows directly to agent context. Sync applications, execute rollbacks, and inspect cluster synchronizations.
Feature Expansion Roadmap
Terraform Plan Awareness
Consume JSON plans (terraform show -json) to detect drift and risk profiles prior to executing apply commands.
Policy Packs
Map configuration checks against common compliance profiles including CIS benchmarks, NIST, SOC2, and custom regulatory standards.
Autofix Patch Suggestions
Return ready-to-commit code patches to remediate insecure security rules, missing data encryption, or open ports.
Step into the formation
Experience our pay-as-you-go credit-based MCP server live demo and walkthrough slides today.