Governance That Follows
Derivation Trees.

A portable semantic specification and cryptographic reference monitor for provenance-scoped governance conditions across agentic AI pipelines, datasets, and software supply chains.

Launch Reference Monitor Read Specification v0.1

5 Hardware KMS Authorities

W3C PROV-O Lineage Schema

O(V+E) Lattice Evaluation Depth

RLS Multi-Tenant Isolation

MGC Reference Monitor

KMS Cluster: us-east-2 · Status: Active

urn:sha256:d8a2...3f1 Base Dataset · Inscribed by ISSUER_ROOT

DENY

urn:sha256:7b91...8c4 Fine-Tuned Model Weights · Inherited Encumbrance

DENY

GATE: production-inference-boundary Enforcement Gate · Blocked Until Discharged

GUARDED

The Enforcement Gap

A static signature is only half the record.

Conventional SLSA and SBOM attestations record what an artifact was at the moment of build. When models are fine-tuned, prompts are composed, or weights are transformed, static signatures detach. Material Governance Conditions bind encumbrances directly to provenance trees.

What Conventional Signatures Capture

Static digests: Point-in-time SHA-256 hashes that break when any derivative artifact is compiled or fine-tuned.
Flat SBOM inventories: Flat component lists lacking causal lineage, parent-child inheritance, or semantic constraint propagation.
Single-key admin authority: Omnipotent master credentials where any admin key can sign, override, or erase security restrictions.
Disconnected PDF/JSON claims: Advisory attestation documents disconnected from live runtime authorization gates.

What Material Governance Enforces

Transitive downward inheritance: Restrictions attached to training data or base models flow automatically into child derivatives.
Bounded lattice algebra: Mathematical semilattice where DENY strictly dominates, eliminating ambiguous policy conflicts.
5 typed hardware KMS authorities: Cryptographically enforced role separation; discharge requires explicit hardware authority keys.
Zero-knowledge credential brokering: Ephemeral STS access tokens minted at the boundary only upon cryptographic verification.

Lineage is verifiable graph data. Material Governance ensures policy requirements follow that lineage through every derivation.

The Architectural Primitive

A "Lien" for Code, Models & Data

Comparing encumbrances on physical real property with provenance conditions on modern digital assets

Physical World Precedent

Real Estate Property Lien

Recorded on Title: Attached to the legal county property registry, completely independent of who currently resides in the building.
Runs with the Land: Survives transfer of ownership, tenant turnover, structural renovations, and subdivision into parcels.
Encumbers Title Closing: Prevents clean settlement or transfer until the underlying obligation is formally satisfied or released.
Typed Authority: Only specific authorized lienholders (tax authorities, mortgage lenders, contractors) have legal capacity to discharge.

Digital Supply Chain Analogue

Material Governance Condition (MGC)

Recorded on Provenance: Bound cryptographically to artifact digests via W3C PROV-O, regardless of which container or registry hosts it.
Runs with Derivation: Survives git forks, container builds, model fine-tunes, weight quantizations, and multi-agent pipeline handoffs.
Encumbers Deployment Gates: Prevents runtime execution, agent tool consumption, or production deployment until the condition is satisfied.
Typed Hardware KMS Authority: Dispositions require hardware-backed KMS signatures strictly matching designated typed authority roles.

Governed Processing Lifecycle

From Inscription to Hardware Enforcement

Every condition moves through a deterministic, mathematically provable pipeline verified at the infrastructure boundary.

STAGE 01

Condition Inscription

Security requirements, licensing obligations, and regulatory encumbrances are inscribed directly onto entity digests with typed authority signatures.

STAGE 02

Provenance Tracking

Graph traversal maps parent-to-child relationships across W3C PROV-O derivation edges, ensuring every fork, fine-tune, and build preserves full lineage.

STAGE 03

Downward Inheritance

Conditions propagate downward through derivation trees. Bounded lattice algebra resolves multiple parent conditions deterministically with zero policy ambiguity.

STAGE 04

Typed Disposition

Hardware-backed KMS keys matching explicit disposition roles (Issuer, Security, Risk Owner, Adjudicator) verify satisfaction before production gates unlock.

Mathematical Rigor

Bounded Lattice Semantics

Conventional policy engines suffer from conflicting rule precedence, circular definitions, and order-dependent evaluation bugs. MGC solves policy conflict by structuring governance verdicts as a formal bounded meet-semilattice:

Status(v) = Local(v) ⊓ ⨅_{u ∈ Parents(v)} Inherited(u)
Order: DENY ⊏ REVIEW ⊏ MONITOR ⊏ PERMIT

Under this algebraic structure, DENY is the strict zero element (⊥). If any ancestor in an artifact's derivation history is encumbered by an active blocking condition, no downstream child can escape it without cryptographic discharge by an authorized authority.

reference-monitor-eval.sh

# 1. Evaluate downstream artifact against derivation tree
curl -X POST https://app.materialcondition.com/api/evaluate \
  -H "Authorization: Bearer $MGC_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "target_entity": "urn:sha256:7b91c04f981e...",
    "depth_limit": 64
  }'

# 2. Reference Monitor Evaluation Response
{
  "status": "EVALUATED",
  "overall_verdict": "DENY",
  "lattice_meet": "LOCAL_PERMIT ⊓ INHERITED_DENY = DENY",
  "blocking_encumbrance": {
    "condition_id": "MGC-2026-004",
    "root_ancestor": "urn:sha256:d8a2048f...",
    "lineage_depth": 2,
    "required_authority": "ADJUDICATOR"
  }
}

Cryptographic Separation of Duties

Hardware-Enforced Authority Roles

No single administrator key can authorize, modify, and discharge conditions. AWS KMS hardware security modules enforce physical boundary separation across 5 distinct authority roles.

ISSUER_ROOT

Authorized to inscribe initial material conditions onto root datasets, base models, and source repositories. Anchors initial compliance scope.

kms:alias/mgc-issuer-root

SECURITY_RESOLUTION

Dedicated authority for discharging security vulnerability encumbrances (CVEs, poisoned weights, secret leaks) after remediation verification.

kms:alias/mgc-sec-resolution

RISK_OWNER

Authorized to grant policy waivers, adjust acceptable risk tolerances, and approve controlled deployments under operational supervision.

kms:alias/mgc-risk-owner

BREAK_GLASS

Emergency high-privilege override key for mission-critical incidents. Time-bounded, dual-custody authorization with mandatory immutable audit logging.

kms:alias/mgc-break-glass

ADJUDICATOR

Independent judicial authority for multi-party dispute resolution, licensing reconciliations, and clearing encumbrances on downstream forks.

kms:alias/mgc-adjudicator

Operational Environments

Engineered for High-Stakes Pipelines

Material Governance Conditions provide cryptographically provable guardrails across modern computing and automated agent environments.

Autonomous Systems

Agentic AI Pipelines

Prevent autonomous multi-agent pipelines from consuming compromised tools, poisoned context vectors, or unvetted external APIs. Downward inheritance blocks child subagents when parent inputs carry active restrictions.

Atomic TOCTOU-immune credential issuance per tool invocation
Transitive encumbrance blocking prompt-injected memory states
Hardware-attested clearance required before agent tool dispatch

DevSecOps & SBOMs

Software Supply Chain Security

Transform passive SBOM inventories into active runtime enforcement gates. Inscribe material conditions on base container images and third-party libraries that travel transitively into deployed production microservices.

Transitive inheritance through multi-stage Docker builds
Deterministic blocking of compromised upstream dependencies
Cryptographic audit trail linking binary digests to source commits

Model Development

Foundation Model Fine-Tuning

Enforce dataset licensing terms, copyright conditions, and safety guardrails across the entire model training lifecycle. Conditions attached to raw training corpora follow model weights through fine-tuning, LoRA adapters, and quantization.

License encumbrances run with derived neural weights
Prevent unauthorized commercial deployment of research-only data
Provable lineage satisfying EU AI Act documentation mandates

Governance & Audit

Regulatory Compliance & Forensics

Provide external auditors and regulatory authorities with tamper-evident cryptographic proofs. Replace manual spreadsheet attestations with machine-evaluable DAGs signed by hardware security modules.

Standardized W3C PROV-O schema for cross-platform auditability
Dual-custody KMS logging on all Break-Glass emergency overrides
Deterministic semilattice evaluation eliminating human interpretation bias

Enterprise Isolation

Cryptographic Boundaries & Database Security

MGC is engineered to protect critical supply chain infrastructure with mathematical isolation and zero ambient credentials.

Row-Level Security (RLS)

PostgreSQL Row-Level Security with FORCE ROW LEVEL SECURITY strictly enforces multi-tenant boundaries at the database engine layer. Application connection roles lack rolbypassrls, making cross-tenant data leakage mathematically impossible.

Credential Brokering

Agents never hold long-lived cloud credentials or static API keys. The MGC Credential Broker verifies condition satisfaction atomically at the gate, vending short-lived AWS STS session tokens scoped strictly to authorized actions.

Immutable Audit Ledger

Every condition inscription, derivation edge creation, and typed KMS discharge event is appended to an immutable, cryptographically hashed audit ledger. Event histories are cryptographically bound to hardware KMS signatures.

Programmatic Interfaces

Developer & MCP Server Integration

Connect MGC directly into IDEs, agent orchestration platforms, and CI/CD pipelines via native Model Context Protocol (MCP) and REST endpoints.

Model Context Protocol (MCP) Server

Native MCP tools for Claude Desktop, Cursor, and Antigravity IDE enabling real-time condition evaluation directly inside developer workflows.

FastAPI REST Reference API

Low-latency endpoints for entity registration, derivation tree ingestion, condition discharge, and JSON-LD graph exports.

Python Core SDK

Type-safe Python package implementing W3C PROV models, lattice meet operators, and AWS KMS typed authority signing clients.

mcp_config.json

// Model Context Protocol Configuration
{
  "mcpServers": {
    "mgc-governance": {
      "command": "uvx",
      "args": ["mgc-mcp-server"],
      "env": {
        "MGC_ENDPOINT": "https://app.materialcondition.com",
        "MGC_API_KEY": "mgc_live_92f8..."
      }
    }
  }
}

Framework Lineage

Built within the Cognispace Framework

Governed under STD-001 v2.2, the Cognispace Framework establishes mathematical, architectural, and ethical standards for autonomous and algorithmic systems. Material Governance Condition (MGC) formalizes how governance constraints propagate through digital derivations without loss of fidelity.

Explore the Cognispace Framework →

Leadership & Architectural Vision

Engineered by practitioners committed to verifiable governance.

Material Governance Condition (MGC) is founder-led and focused. Every architectural decision, algebra definition, and hardware KMS implementation reflects a commitment to mathematical rigor, auditable provenance, and software supply chain integrity.

LinkedIn Profile

Dr. Donnel Hinkins

Founder & Chief Architect

Ph.D. in Information Technology • Founder, Cognispace, LLC • Creator, Cognispace Framework

Dr. Donnel Hinkins is the founder and chief architect of Material Governance Condition (MGC) and the founder of Cognispace, LLC. His work focuses on systems architecture, digital provenance intelligence, cybersecurity, and the structural modeling of complex distributed systems.

With an applied background spanning information technology, cybersecurity, and higher education, he has dedicated his career to examining how structural patterns emerge across technical and computational environments. His engineering approach centers on systems thinking: identifying mathematical invariants that govern information flow, authorization boundaries, and state transitions across complex pipelines.

He is the creator of the Cognispace Framework, a formal architecture for understanding how impulse, evaluation, and expression interact across cognitive, organizational, and autonomous software systems. Material Governance Condition (MGC) represents the formal provenance-governance realization of that foundation, defining how conditions, licenses, and security restrictions propagate downward through derivation graphs without loss of fidelity.

Under his architectural direction, MGC enforces strict operational boundaries: bounded semilattice resolution, hardware-isolated KMS authority roles, database-enforced row-level security, and auditable derivation provenance.

Dr. Hinkins holds a Ph.D. in Information Technology and directs architecture, formal semantics, and platform engineering across the Cognispace technology portfolio.

Information Technology Cybersecurity Systems Architecture Cryptographic Provenance Software Supply Chains Cloud Infrastructure Research & Development

We're building the team.

MGC is founder-led and focused. As we expand enterprise pilots and reference monitor deployments, we are looking for engineers and researchers who share our commitment to provable software supply chain security and autonomous agent governance.

If you are an engineer or researcher focused on formal methods, cryptographic authorization, or distributed graph systems, we invite you to connect.

Connect With Us

Engineering

Cryptographic Systems & Distributed State

Experience in distributed state machines, hardware KMS integrations, and low-latency authorization engines.

Research

Provenance Semantics & Formal Verification

Background in graph theory, lattice algebra, or programming language theory applied to artifact derivation trees.

Enterprise

Enterprise Architecture & Compliance

Experience working with DevSecOps teams, regulated enterprise architectures, and software supply chain compliance frameworks.

Strict Governance & Architectural Boundaries

MGC operates as a deterministic security monitor. We adhere to rigorous operational boundaries.

Non-Speculative Evaluation

MGC does not employ black-box LLMs to guess at policy compliance. All evaluation relies on formal bounded semilattices and verified cryptographic signatures.

No Single Master Backdoor

There is no universal superuser key capable of bypassing or clearing conditions. Only hardware KMS keys matching explicit typed authority roles can discharge encumbrances.

Fail-Closed Security Posture

Unreachable lineage paths, severed derivation graphs, or unrecognized authority signatures default strictly to DENY, preventing unauthorized pipeline execution.

Enforce governance that survives the pipeline.

Deploy the canonical MGC Reference Monitor or integrate the portable specification into your existing software and model deployment pipelines.

Launch Reference Monitor Explore Specification v0.1

Governance That Follows Derivation Trees.