Google Cloud adds an MCP server for AI agents to use cloud rules

Google Cloud made 2 parts of Organization Policy available to all users. Organization Policy is a tool for rules on cloud resources. The first part is an MCP server. MCP is a standard connection that lets AI agents use tools. The other part lets administrators write custom rules with a Gemini prompt. The post gives no test results and no use numbers.

How it works

Agents use tools to show the rules for a project. They also calculate the policy that applies and make new policies. An agent can make a dashboard of rules and exemptions.

Gemini can make test setups that follow a rule and that do not. A dry run uses them before the rule starts to apply.

The proof

Google gave no test results. Its one number is a catalog of more than 100 policies that come with the tool. Before, a person who sets rules had to know cloud resource schemas to write a custom rule.

What is not known

The post does not tell how frequently Gemini writes an incorrect rule. It gives no data on speed or time.

Insight: Each vendor makes its own rule layer for agents

HashiCorp made Terraform policy open to all users. It gives one set of rules to persons and AI agents. Google at this time lets agents read and write its cloud rules. Each vendor possibly makes a rule layer for agents. The rules of one tool do not include the other tools. Teams that use 2 or more tools can have rules in different systems.