skill-deployment-context

# Koji Deployment Context Discovery **Always resolve context before running queries. Never assume resource names.** ## Step 1: Locate `deployment_summary.yaml` (if available) This file is typically generated by the Koji deployment pipeline (e.g., as an Azure DevOps pipeline artifact) and stored at the repository root as `./deployment_summary.yaml`. It should contain `resource_group`, `aks_cluster_name`, `subscription_id`, and monitoring resource names/IDs. - If the file exists, open it and use its values as the source of truth for the deployment context. - If the file does **not** exist (e.g., in a new or manually provisioned environment), skip to Steps 2–4 and resolve the deployment context directly from Azure. - If the user needs help obtaining this file, offer to run `az` commands to discover the values and generate it for them. ## Step 2: If subscription is missing Run: `az account show --query id -o tsv` ## Step 3: If deployment does not match the summary Ask the user for the **resource group** or **AKS cluster name**. ## Step 4: Discover monitoring resources dynamically ```bash az monitor log-analytics workspace list --resource-group <RG> -o table az resource list --resource-group <RG> --resource-type Microsoft.Monitor/accounts -o table az resource list --resource-group <RG> --resource-type Microsoft.Dashboard/grafana -o table ``` ## Naming convention Resources follow `<prefix>-koji-*`: - Resource group: `<prefix>-koji-rg-<suffix>` - AKS cluster: `<prefix>-koji-aks` - Log Analytics: `<prefix>-koji-logs` - Azure Monitor: `<prefix>-koji-monitor` - Grafana: `<prefix>-koji-grafana` ## MCP tools for context discovery | Tool | Purpose | |------|---------| | `aks_cluster_get` | List/get AKS cluster details | | `aks_nodepool_get` | List/get node pool details | | `monitor_table_list` | List tables in a Log Analytics workspace | **Prerequisites:** Node.js (`npx`) and an active `az login` session.