Cloud Resource Manager MCP server

Cloud Resource Manager MCP server gives an agent access to the containers that hold Google Cloud resources: organizations, folders, projects, and the tags applied to them. Through Neotask, an agent can create a new project inside a given folder, read the metadata attached to an existing project or folder, update labels and tags, and generally help keep the hierarchy of an organization's cloud footprint organized. This matters most for teams whose Google Cloud presence has grown past a handful of projects, where knowing which folder a project sits under, or which tags govern its policies, becomes something worth automating rather than tracking by memory. Instead of a cloud admin manually checking the console to confirm a project's parent folder before applying a policy, the agent can read that structure directly, make the requested metadata change, and confirm the update took effect. It's an administrative layer over the structure of a cloud environment rather than over the workloads running inside it, which keeps its blast radius limited to organization, folder, and project bookkeeping. For a company that acquires teams or spins up new business units regularly, that bookkeeping otherwise falls to whoever remembers to do it manually, and this integration turns it into something an agent can maintain continuously instead.

What you can automate

Create projectProvision a new project under a chosen folder or organization node with an initial set of labels.
Read project metadataRetrieve a project's parent folder, labels, tags, and current lifecycle state.
Update tagsApply or change tags on a project or folder for policy enforcement or billing attribution.
List foldersEnumerate the folders under an organization to map out the full hierarchy.
Move projectReassign a project to a different parent folder within the hierarchy after a reorganization.
Inspect policy scopeConfirm which container a given organization policy is attached to before assuming it applies elsewhere.

Real workflows

Onboarding a new team's project

When a new team joins the company and needs its own Google Cloud project, a platform engineer asks the agent to set one up under the correct department folder with the standard tags applied. The agent creates the project, applies tags matching the team's cost center and environment, and confirms the metadata matches what similar existing projects already have in that same folder. It also checks that the folder's existing organization policies will apply to the new project by virtue of its placement. This keeps the new project consistent with the rest of the hierarchy without someone manually copying settings from an older project.

Cleaning up a drifted hierarchy

An admin notices that several projects seem to be sitting in the wrong folder after a reorganization last quarter, but isn't sure which ones without checking each individually. The agent reads the metadata for every project under the affected folders, compares parent folders against the current org chart the admin shares, and lists eleven projects that appear misplaced relative to their owning department. Once the admin confirms the correct destination for each one, the agent moves each flagged project to the right folder and reports back a short summary of what changed.

Frequently asked questions

Can the agent restructure an entire organization on its own?

It can create, read, and update containers and their metadata, but a large restructuring should still be reviewed step by step rather than left entirely to automated judgment, since folder moves can affect which policies apply.

Does this touch IAM permissions or policies directly?

Its scope is the resource containers themselves and their metadata, such as tags and parent relationships. Policy enforcement and IAM bindings are typically handled through related, separate tooling rather than this integration.

Can it tell me which folder a specific project belongs to?

Yes, reading a project's metadata to confirm its parent folder is one of its core capabilities, and it can do this across many projects at once when asked.

Is tagging something the agent can do in bulk?

It can apply tag updates to multiple projects in sequence when asked, provided each update is specified clearly enough to avoid ambiguity between similarly named projects.

What happens if a requested project name is already taken?

Project creation will fail the same way it would through the console, since project identifiers need to be unique within the applicable scope, and the agent will report the failure rather than silently retry with a different name.

Does it help with cost allocation across projects?

Indirectly, since consistent tags on projects and folders make it easier for billing exports to be split accurately by team or environment, though it doesn't produce billing reports itself.

Can it flag projects that don't follow naming or tagging conventions?

Yes, comparing a project's actual labels and name against a defined convention is a straightforward check once the convention is described, and the agent can list every project that falls outside it.