Skip to main content
NEW: RSAC 2026 NHI Field Report. How Non-Human Identity became cybersecurity's central axis
Back to Integrations
Permission AnalysisPro

GCP Permission Analyzer

Connect a read-only service account to map IAM bindings and roles on discovered GCP credentials.

Key Features

  • Service-account-based authentication
  • Organization, folder, and project-level analysis
  • Inherited IAM role resolution
  • Custom role binding detection
  • Workload Identity Federation support

Requirements

  • 1GCP organization admin or project owner
  • 2Ability to create a service account with read-only IAM roles
  • 3Cremit Argus account on Pro plan or above
Setup Time10 min

Step-by-step setup guide

The exact flow you follow inside the dashboard.

Overview

The GCP Permission Analyzer connects a read-only service account to Argus so that every discovered GCP credential can be analyzed for its real IAM bindings, including inherited organization and folder-level roles. Supports traditional service-account keys and Workload Identity Federation.

Prerequisites

  • GCP organization admin or project Owner
  • Permission to create service accounts and grant read-only IAM roles
  • A Cremit Argus account on the Pro plan or above

Step-by-Step Setup

Step 1: Create a dedicated service account

This account gives Argus read-only access to IAM metadata.

  • In GCP Console, go to IAM & Admin > Service Accounts > Create service account
  • Name it 'cremit-argus-analyzer' and give it a description
  • Skip granting project access in this step; we grant org-scope roles next

Step 2: Grant the analyzer's required roles

Apply these at the organization level so the analyzer can see inherited bindings.

  • iam.securityReviewer: read IAM policies across all resources
  • iam.serviceAccountViewer: list service accounts and their metadata
  • iam.roleViewer: read custom role definitions
  • Optionally resourcemanager.folderViewer and resourcemanager.projectViewer for inventory

Step 3: Choose authentication method

Pick either a service-account key (simpler) or Workload Identity Federation (no long-lived keys).

  • Service-account key: create a JSON key, download it, keep it secret
  • Workload Identity Federation: set up a provider that trusts Argus's OIDC identity, then exchange tokens at runtime

Step 4: Register the analyzer in Argus

  • In Argus, go to Security > Permission Analysis > Add GCP Analyzer
  • Paste the service account email
  • Upload the JSON key OR paste the Workload Identity Federation configuration
  • Click Test Connection, then Register

Verification

To confirm the integration is configured correctly:

  • The analyzer shows Connected under Registered Analyzers
  • Analysis Results shows parsed bindings for existing GCP credentials
  • Inherited roles from org and folder levels are listed, not only project-level
  • GCP audit logs confirm the service account makes read-only IAM calls

Troubleshooting

Issue: Analysis misses inherited roles at the organization level.

  • Solution: iam.securityReviewer must be granted at organization scope, not project scope. Re-grant at the org node and wait up to 7 minutes for IAM propagation.

Issue: 'Workload Identity Federation token exchange failed.'

  • Solution: The audience claim in the provider configuration must match exactly what Argus issues. Regenerate the config from Argus and re-upload.

Key Benefits

  • Read-only access via a dedicated, single-purpose service account
  • Resolves inherited roles, so 'why does this key have BigQuery access?' is answerable
  • Workload Identity Federation support avoids long-lived service-account keys entirely
  • Auto re-analysis when bindings change keeps the picture current

Get started now

Set up the GCP (Service Account) integration in minutes and start improving your security today.

GCP (Service Account) Integration