Skip to main content
Security Analysis

Security Model & Threat Assumptions

Explicit threat model for Attested Governance Artifacts. What we defend against, what we do not, and the assumptions that underpin our guarantees.

Security Goals

What AGA Proves

AGA provides verifiable execution integrity, tamper-evident enforcement records,
and portable audit evidence.

It generates cryptographic proof that a governed subject operated within authorized constraints. Every enforcement decision is signed, chained, and independently verifiable.

Adversary Model

Assumed Attacker Capabilities

The adversary model assumes an attacker with full access to the local database and network but not the cryptographic signing keys or portal runtime.

AGA provides structural integrity guarantees independent of payload confidentiality. The system is designed so that even with full visibility into stored data and network traffic, an adversary cannot forge valid receipts or tamper with the enforcement chain without detection.

Threats Addressed

What AGA Defends Against

Binary modification

Hash comparison against sealed baseline

Configuration drift

Continuous measurement at policy-defined cadence

Unauthorized dependency changes

Checksum validation of loaded modules

Runtime policy bypass

Portal as mandatory execution boundary

Audit log tampering

Hash-linked receipt chain with structural metadata linking

Evidence fabrication

Merkle checkpoint anchoring to immutable storage

Replay attacks

Timestamped receipts with sequence numbering

Privilege escalation

Two-process separation; subject holds no signing keys

Supply chain model substitution

Sealed reference hash binds identity at build time

Behavioral drift in tool-call patterns

Chi-squared and z-score analysis over sliding windows, generating signed BEHAVIORAL_DRIFT receipts

Policy bypass via indirect execution

Phantom execution with forensic capture detects attempts to circumvent governance controls

Threats Not Addressed

Explicit Exclusions

  • Kernel-level or hypervisor compromise of the portal host
  • Hardware microcode or firmware attacks
  • Adversarial prompt manipulation at the model level (semantic drift at inference time remains outside scope)
  • Side-channel attacks on the signing key
  • Social engineering of the policy issuer

Note: AGA now addresses behavioral drift at the tool-call level via statistical analysis over sliding windows. Model-level semantic defenses remain complementary but separate. Listing what we do not address is intentional transparency, not a limitation admission.

Trust Assumptions

Foundation of Guarantees

Portal Integrity

The portal must execute within a trusted boundary. If compromised, receipt integrity cannot be guaranteed. For highest assurance, the portal should execute within a TEE, though this is not required for standard operation.

Key Custody

The policy issuer's signing key must be protected. If the signing key is compromised, forged artifacts become possible. Key rotation is supported through the continuity chain mechanism.

Verification Guarantees

Assurance Levels

LevelStepsGuarantees
OfflineSteps 1–3
  • Artifact authenticity
  • Receipt chain integrity
  • Merkle proof validation
OnlineStep 4
  • Checkpoint anchor validation against immutable storage
FullAll steps
  • Complete end-to-end governance verification

Explore the System

See how these security properties are implemented in practice across the governance architecture.

MCP ENFORCEMENT

Zero-TOCTOU Guarantee for MCP Agents

For MCP-governed AI agents, the time-of-check-to-time-of-use window is zero by construction. The Portal IS the communication channel. Every tool call passes through the Portal synchronously. The agent cannot invoke a tool without the Portal evaluating it first.

If the Portal is the only path to external resources (enforced by network isolation), and every tool call is synchronously evaluated before forwarding, then the agent cannot perform any unauthorized external action. Internal computation is unconstrained, but internal computation without external action causes no harm. The moment the agent acts, it hits the Portal.

This property is specific to MCP-governed agents where the proxy is the sole communication path. For container and binary subjects, AGA uses periodic measurement with configurable cadence (100ms-5s) and event-driven eBPF probes on the roadmap.

PHANTOM EXECUTION

Phantom Mode Limitations

When a quarantine trigger fires, AGA transitions the agent to phantom mode: spoofed success responses, severed real effects, full forensic capture. This captures adversarial intent without alerting the agent.

A sufficiently capable agent could potentially detect phantom mode through timing analysis, response content patterns, or side channels. AGA mitigates this by profiling real response latencies during normal operation and replaying from the recorded distribution during phantom mode. The JSON-RPC envelope structure matches exactly.

The agent would need to specifically allocate reasoning capacity to phantom detection, which requires knowing phantom mode exists. The purpose is to capture intent during the observation window, not to maintain indefinite deception against a determined detection effort.

RESOLUTION STATUS

Logical Problem Resolution

Independent analysis identified 12 non-trivial logical problems in the AGA architecture. Nine are fully solvable with engineering. Three are irreducible residuals present in any cryptographic governance system.

ProblemCategoryStatusMitigation
Portal trustFundamentalMitigatedThreshold signing, TEE attestation
Key compromiseFundamentalMitigatedHSM isolation, rotation, PQC hybrid
Policy correctnessFundamentalIrreduciblePolicy provenance chain, coverage metrics
TimestampsArchitecturalMitigatedRoughtime IETF integration, monotonicity verification
Identity bindingArchitecturalMitigatedSPIFFE self-signed X.509, certificate chain binding
TOCTOU windowArchitecturalSolvedMCP proxy is synchronous by construction
Proxy bypassArchitecturalSolvedMandatory network isolation at infra level
Phantom detectionArchitecturalMitigatedTiming profiles, envelope matching
Recursive governanceArchitecturalRoadmapTEE-hosted Portal with attestation chain
Policy rotationOperationalSolvedPOLICY_ROTATION receipt type, zero-gap protocol
Performance at scaleOperationalMitigatedReceipt batching, horizontal scaling
Legal admissibilityOperationalIrreducibleStandard primitives, expert witness readiness