How to use the SLO Alert Configuration Wizard agent skill

What does the SLO Alert Configuration Wizard skill make the agent do?

Once installed, the agent acts as a step-by-step setup wizard for Service Level Objective (SLO) alerts in Google Cloud Monitoring. It asks for the target service, the reliability target, the indicator being measured, and the alert condition. It then writes Terraform (.tf) files that create the alerting policies using the google_monitoring_alert_policy and condition_prometheus_query_language resources.

The skill can also run gcloud commands to look up your App Hub services or underlying resources, so the agent can fill in some details for you instead of asking.

• Models an SLO with four components: Service Scope, Service Level, SLI, and Alert Condition.
• Generates PromQL-based SLO alerting policies as Terraform output only.
• Adds a user_labels block to each alert policy so the skill's policies can be identified later.
• Does not cover standard, non-SLO alerting policies.

How do you install the skill?

You can install the skill in one of two ways. Download the ZIP file from the skill's page on this site and unzip it into your personal skills folder, ~/.claude/skills/, or into a project's .claude/skills/ folder. Alternatively, in the Claude apps, open Settings, go to Capabilities, and upload the skill under Skills.

After installing, the skill's folder should contain SKILL.md and a references folder with app_hub_labels.md, attribution.md, promql_templates.md and service_metrics.md.

• Personal install: unzip into ~/.claude/skills/
• Project install: unzip into .claude/skills/ inside your project
• Claude apps: Settings → Capabilities → Skills, then upload the skill

What do you need before you start?

The skill assumes you have a Google Cloud project with Cloud Monitoring enabled. App Hub and Asset Inventory are optional but let the agent discover services for you. You also need the Google Cloud SDK (gcloud) for the discovery commands, and Terraform to apply the generated configuration.

Check these before you start. Without a Google Cloud project or a Terraform workflow, the output is not useful, and the lookups need gcloud access.

• A Google Cloud project with Cloud Monitoring
• Optional: App Hub and Asset Inventory enabled
• Google Cloud SDK (gcloud) for resource discovery
• Terraform to apply the generated configuration

How do you trigger the skill with a natural request?

The skill is meant to activate when you ask to configure an SLO or Service Level Objective. You do not need special syntax. Describe what you want in plain language and include whatever details you already know, such as the service name, target percentage, or metric.

The agent will ask for any missing information in a single reply, and it will state a recommended best practice with each question. It will not write Terraform until the required information is complete.

• Example: "Configure an SLO for my Cloud Run service checkout in project my-project."
• Example: "Set up a 99.9% availability SLO alert for my App Hub workload backend."
• Example: "Use best practices for my SLO alert on the frontend service."

What happens during the wizard?

The agent works through four steps. In Step 1 it defines the ServiceScope, discovering services with gcloud if you did not name them. In Step 2 it defines the ServiceLevel target, such as a 99.9% target over a rolling 28-day period, which the skill recommends as a starting point. Step 3 covers the ServiceLevelIndicator, where you name the metric or describe the latency or availability condition. Step 4 sets the alert condition.

If you ask for best practices, the agent keeps any values you already gave and fills the blanks with its recommended defaults. Every generated alert policy includes an alert_strategy block with an auto_close setting and leaves notification_channels empty unless you supply one.

What safety and network behaviour should you expect?

This skill runs scripts and uses the network. Its gcloud commands read your Google Cloud resource information, such as App Hub services and Cloud Run services, so it needs access to your project. Before running any gcloud, REST API, or MCP command, the skill requires the agent to read its attribution reference. Terminal gcloud commands are prefixed with an environment variable that tags usage with the skill name and version, and the skill says it sets this inline rather than using gcloud config set.

If a command fails, for example because an API is not enabled or permission is denied, the skill tells the agent to stop troubleshooting and ask you for the missing information. Review the generated Terraform before applying it, because the skill produces configuration files only and does not check that alerts work in production.

• Runs scripts and uses the network
• Discovery commands run against your own Google Cloud project
• Output is configuration files only; test alerts yourself after applying

What tips will get better results?

Give the agent as much as you can in the first message: the service, the target percentage, the metric, and the threshold. The fewer gaps there are, the fewer rounds of questions. If you are unsure of the metric, say so and the agent may infer one from the service type if the skill's references define it.

For App Hub resources, say so explicitly or include the App Hub URI. The skill then uses App Hub labels and checks the label mappings in its references before generating configuration, rather than guessing.

• State the App Hub context explicitly when the resource is in App Hub.
• Provide the exact metric name, such as run.googleapis.com/request_count, when you know it.
• Ask for best practices if you want defaults for anything you have not decided.