Skip to content

OWASP MCG Project Draft

Jump to bottom
Michael Schwartz edited this page Nov 10, 2025 · 8 revisions

Project Name

OWASP Model Context Governance (MCG) Practitioner Guide & Toolkit

Project Classification

OWASP Incubator Project
Category: Application Security / Capability Governance / Continuous Assurance

1. Overview

Traditional governance programs such as Identity Governance & Administration (IGA) were designed to satisfy compliance requirements that required organization to certify human entitlements via periodic compliance reviews. These approaches cannot address the growing scale of machine-to-machine interactions, token-based authority, and autonomous software components that make real-time authorization decisions.

Model Context Governance (MCG) defines a new approach centered on capabilities. MCG says rather then inventory the people, inventory the locks--it's simpler.

MCG seeks to define an engineering framework for continuous, provable, and federated governance. It enables enterprises to govern through policies that are declarative, verifiable, and responsive to telemetry and threat data.

2. Design Principles

MCG is based on key architectural and operational principles derived from the Trust Governance Architecture:

  1. Inventory capabilities, not identities – manage risk by enumerating what can be done (action–resource pairs), not who performs it.
  2. Declarative governance with formal reasoning – express policies and token claims in verifiable form; use proofs to reduce subjective review.
  3. Continuous governance – treat each code, configuration, or deployment change as a governance event.
  4. Federation by policy schema – enable cross-domain interoperability via shared capability and token semantics, not centralized directories.
  5. Multi-layer visibility – connect the human, software, observability, and trust management layers into a unified governance fabric.
  6. Data provenance – use signed tokens (e.g., JWTs or verifiable credentials) to propagate trust and traceability.
  7. Actionability – governance outputs must enable safe policy change, rollback, and verification, not just reporting.

These principles make governance measurable, testable, and automatable across distributed systems.

3. Objectives

  1. Define a capability governance lifecycle for MCP servers and proxies, including discovery, modeling, issuance, enforcement, telemetry, and proof.
  2. Publish a practitioner guide explaining how to model capabilities, author declarative policies, and integrate formal reasoning techniques.
  3. Provide tool-agnostic templates for capability definitions, token schemas, assurance levels, and rollout checklists.
  4. Develop a federated policy reference model demonstrating interoperability between autonomous governance domains.
  5. Create a continuous governance playbook that embeds automated validation and telemetry feedback into DevSecOps pipelines.
  6. Establish the foundation for future formal specifications and verification tools for capability governance.

4. Deliverables

Deliverable Description Format Target Version Target Date
MCG Practitioner Guide v0.1 (MCP Edition) Initial draft describing the capability governance lifecycle, token flows, and policy examples for MCP servers/proxies Markdown / PDF 0.1 Month 3
Capability & Token Templates Editable templates defining capability models (action-resource), JWT claim sets, and trust assertions Markdown / JSON 0.2 Month 4
Continuous Governance Playbook Implementation guide for integrating observability, threat data, and verification pipelines Markdown / YAML 0.3 Month 5
Federated Policy Reference Model (MCP Use Case) Diagrams and schemas illustrating multi-domain governance via capability exchange Diagram / YAML 0.4 Month 6
MCG Practitioner Guide v1.0 Consolidated release with guide, templates, and reference models PDF / Web 1.0 Month 7
Launch Webinar & Slides Public presentation introducing Model Context Governance principles and toolkit PPTX / Video — Month 8
Maintenance & Expansion Roadmap Plan for extending MCG to additional systems (API gateways, AI agents, observability platforms) Markdown 1.1 Month 9

5. Scope

In Scope

  • Governance lifecycle for MCP servers and proxy systems
  • Capability and token-based policy modeling and enforcement
  • Integration of runtime telemetry and threat data into governance feedback loops
  • Federated policy exchange across autonomous domains
  • Declarative policy authoring and continuous validation

Out of Scope

  • Vendor-specific or proprietary integrations
  • Human identity access certification and entitlement reviews (traditional IGA)
  • Compliance frameworks outside technical governance and assurance

6. Target Audience

  • Application Security and DevSecOps engineers
  • Platform architects managing distributed authorization and policy systems
  • AI and agent governance engineers
  • GRC professionals building continuous assurance models
  • Researchers exploring federated and formal governance mechanisms

7. Project Roles

Role Name / Placeholder Responsibilities
Project Leader Michael Schwartz Direction, documentation, and community coordination
Co-Leader Manoj Kumar Technical design, contributor coordination
Community Manager Ravindra Neriyanuri Repository maintenance, contributor engagement
Technical Writers Community volunteers Content development, editing, and formatting
Review Board AppSec & Governance experts Peer review and technical validation
Design Lead TBD Visual design, diagrams, and presentation assets

8. Licensing

All materials are released under the Creative Commons Attribution 4.0 International License (CC BY 4.0), consistent with OWASP project policy.

9. Repository & Infrastructure

Clone this wiki locally