Threat model

Mobius Systems Threat Model v1.0¶

Status: Draft
Version: 1.0.0
Date: 2025-11-09

Executive Summary¶

This document defines the threat model for Mobius Systems, identifying assets, adversaries, threats, and controls.

Asset Classification¶

Tier 1: Critical Assets¶

Tier 2: Important Assets¶

Adversary Models¶

Adversary 1: External Attacker¶

Capabilities: - Can send network requests to public APIs - Can attempt to exploit vulnerabilities - Cannot access private keys or internal networks - Limited to public attack surface

Motivation: Disrupt service, steal data, manipulate integrity scores

Threat Level: Medium

Adversary 2: Compromised Sentinel¶

Capabilities: - Controls one of four sentinel cortices (AUREA, HERMES, EVE, JADE) - Can provide malicious responses in consensus - Cannot access other cortices or private keys - Limited to 1-of-4 influence

Motivation: Manipulate consensus decisions, approve malicious changes

Threat Level: High

Mitigation: 3-of-4 consensus threshold prevents single compromised cortex from controlling decisions

Adversary 3: Insider Threat¶

Capabilities: - Has legitimate access to codebase - Can submit PRs and merge code - May have access to some credentials - Cannot access HSM-protected keys

Motivation: Introduce backdoors, manipulate MII calculations, bypass guardrails

Threat Level: High

Mitigation: Anti-nuke guardrails, code review requirements, MII gates

Adversary 4: Nation-State Actor¶

Capabilities: - Advanced persistent threat (APT) - Can compromise multiple systems - May have quantum computing resources (future threat) - Long-term operational security

Motivation: Espionage, sabotage, cryptographic compromise

Threat Level: Low (for current deployment scale)

Mitigation: Key rotation, multi-party key escrow, post-quantum crypto preparation

Threat Scenarios¶

T1: MII Signature Forgery¶

Scenario: Attacker attempts to forge MII signature to bypass integrity gates

Attack Vector: - Brute force Ed25519 private key (computationally infeasible) - Replay old valid signatures - Exploit signature verification bugs

Impact: Critical - Allows deployment of low-integrity code

Controls: - Ed25519 signatures (128-bit security) - Timestamp-based replay prevention - Signature verification in multiple layers

Residual Risk: Low

T2: Sentinel Consensus Manipulation¶

Scenario: Compromised cortex provides malicious consensus responses

Attack Vector: - Compromise one cortex (e.g., AUREA API key stolen) - Provide false "approve" responses - Attempt to reach 3-of-4 threshold with malicious votes

Impact: High - Malicious code could be approved

Controls: - 3-of-4 consensus threshold (requires 2 additional honest cortices) - Cortex response signatures (non-repudiation) - Deliberation Proof audit trail

Residual Risk: Medium (if 2 cortices compromised simultaneously)

T3: Ledger State Corruption¶

Scenario: Attacker attempts to modify historical attestations

Attack Vector: - Exploit ledger database vulnerabilities - Compromise ledger service - Replay attacks on ledger API

Impact: Critical - Breaks immutability guarantees

Controls: - Immutable ledger design (append-only) - Cryptographic hashing of ledger state - Regular ledger state verification - Merkle tree proofs

Residual Risk: Low

T4: Anti-Nuke Bypass¶

Scenario: Attacker attempts to delete critical files through multiple PRs

Attack Vector: - Submit multiple PRs each deleting <5 files - Coordinate deletions across time window - Exploit edge cases in deletion detection

Impact: High - Critical code could be deleted

Controls: - Per-PR deletion limits (5 files, 15% ratio) - Protected path restrictions - Time-window aggregation (future enhancement) - Human review for large changesets

Residual Risk: Medium (sophisticated multi-PR attack)

T5: MII Calculation Manipulation¶

Scenario: Attacker manipulates MII calculation to bypass gates

Attack Vector: - Inject false attestations - Manipulate component scores - Exploit MII calculation bugs

Impact: Critical - Low-integrity code deployed

Controls: - Deterministic MII calculation - Attestation verification - MII signature validation - Regular MII audits

Residual Risk: Low

Control Effectiveness Metrics¶

Cryptographic Controls¶

Consensus Controls¶

Operational Controls¶

Security Architecture¶

Defense in Depth¶

┌─────────────────────────────────────────┐
│  Layer 1: Network Perimeter            │
│  - Firewall, DDoS protection           │
└──────────────┬──────────────────────────┘
               │
┌──────────────▼──────────────────────────┐
│  Layer 2: API Gateway                  │
│  - Rate limiting, authentication        │
└──────────────┬──────────────────────────┘
               │
┌──────────────▼──────────────────────────┐
│  Layer 3: Service Mesh                  │
│  - mTLS, service authentication        │
└──────────────┬──────────────────────────┘
               │
┌──────────────▼──────────────────────────┐
│  Layer 4: Application Logic            │
│  - MII gates, consensus checks         │
└──────────────┬──────────────────────────┘
               │
┌──────────────▼──────────────────────────┐
│  Layer 5: Data Layer                    │
│  - Encryption at rest, ledger immutability│
└─────────────────────────────────────────┘

Incident Response¶

Severity Levels¶

1. Critical: MII signature compromise, ledger corruption
2. High: Sentinel compromise, consensus manipulation
3. Medium: API abuse, service degradation
4. Low: Configuration errors, non-critical bugs

Response Procedures¶

1. Detection: Automated monitoring + manual reports
2. Containment: Isolate affected services, revoke credentials
3. Eradication: Remove threat, patch vulnerabilities
4. Recovery: Restore services, verify integrity
5. Lessons Learned: Post-mortem, update threat model

Compliance & Standards¶

• NIST Cybersecurity Framework: Identify, Protect, Detect, Respond, Recover
• OWASP Top 10: Web application security
• ISO 27001: Information security management

Status: Draft v1.0
Next Review: 2025-12-15
Owner: Security Working Group