Skip to main content

CloudForge Documentation

CloudForge is an enterprise cloud governance platform that unifies posture findings, AI-powered risk scoring, policy enforcement, and automated remediation across AWS, Azure, and GCP.

SectionDescription
ArchitectureHigh-Level Design, Detailed Design, DR/BC
ADRs21 Architecture Decision Records
DiagramsSystem architecture and flow diagrams
API ReferenceOpenAPI schema download plus markdown endpoint reference (89 operations)
Posture ManagementMulti-cloud finding aggregation module
Runbooks9 operational runbooks
SecuritySTRIDE threat model

Architecture at a Glance

For the detailed component-level diagram, see Diagrams.

Key Capabilities

  • Posture Management -- Normalize findings from AWS Security Hub, Azure Defender, GCP SCC, Trivy, and Prowler into a unified schema
  • AI Risk Scoring -- LLM-powered severity re-scoring considering asset tier, environment, exposure, and blast radius
  • Dual-OPA Policy -- External OPA server for cloud provisioning + embedded Go SDK for AI agent governance
  • Automated Remediation -- Dispatcher routes findings to provider-specific handlers with approval workflows
  • FinOps Integration -- Multi-cloud cost aggregation with anomaly detection
  • Graph Analysis -- Security-graph context and attack-path analysis for triage and containment

Auth Architecture Today

  • Frontend SSO is a browser-owned Okta SPA PKCE flow that redirects to /callback when VITE_OKTA_ISSUER and VITE_OKTA_CLIENT_ID are configured.
  • The backend validates bearer JWTs via HS256 or RS256/JWKS and serves the public demo with static/demo auth modes.
  • Backend authorize/callback routes, refresh-token storage, and httpOnly session cookies are not implemented today.