Initial commit: Ollama MCP server

MCP server exposing local Ollama models via LiteLLM proxy to Claude Code.
Tools: query_local_model, review_code, summarize, generate_boilerplate, list_models.
Deployed to k8s ai-inference namespace via ArgoCD.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

k8s/argocd-app.yaml

@@ -0,0 +1,43 @@
+---
+# ArgoCD Application - deploy this once to bootstrap:
+# kubectl apply -f k8s/argocd-app.yaml
+#
+# Pre-requisite: Add the repo to ArgoCD first:
+# argocd repo add https://repo.adservio.us/ai_approver/ollama-mcp.git \
+#   --username <gitea-user> --password <gitea-token>
+apiVersion: argoproj.io/v1alpha1
+kind: Application
+metadata:
+  name: ollama-mcp
+  namespace: argocd
+  finalizers:
+    - resources-finalizer.argocd.argoproj.io
+spec:
+  project: default
+  source:
+    repoURL: https://repo.adservio.us/ai_approver/ollama-mcp.git
+    targetRevision: main
+    path: k8s
+    directory:
+      exclude: argocd-app.yaml
+  destination:
+    server: https://kubernetes.default.svc
+    namespace: ai-inference
+  syncPolicy:
+    automated:
+      prune: true
+      selfHeal: true
+      allowEmpty: false
+    syncOptions:
+      - CreateNamespace=true
+    retry:
+      limit: 5
+      backoff:
+        duration: 5s
+        factor: 2
+        maxDuration: 3m
+  ignoreDifferences:
+    - group: apps
+      kind: Deployment
+      jsonPointers:
+        - /spec/replicas