/breach

Threat Modeling Reference

Purpose: Detailed threat modeling methodologies for Breach. Read when: Building or reviewing a threat model for a target system.

2026 framing addition. Classical STRIDE / PASTA still apply, but the highest-impact 2026 threats live at two layers STRIDE was not built for. Add these as a sweep after the STRIDE pass before declaring the model complete:

• Agentic action surface (OWASP ASI 2026) — when an AI agent takes real actions (file writes, API calls, MCP tools, browser automation), model the goal-hijack / tool-misuse / memory-poisoning / cascading-failure paths from ai-red-teaming.md. Anthropic's published ~78.6% breach rate at 200 prompt-injection attempts against an unprotected GUI agent is the calibration point — assume the agent boundary is porous and the human-in-the-loop is load-bearing.
• Supply-chain pipeline (Mini Shai-Hulud family, 2026) — model the build / publish / OIDC / SLSA-attestation chain as a first-class trust boundary. The 2026 in-the-wild worms abuse SLSA L3 attestations from hijacked pull_request_target workflows, so a model that trusts the attestation without verifying signer identity + branch lineage is incomplete. See supply-chain-attack-design.md.

1. STRIDE (Default)

Microsoft's threat classification framework. Best for component-level analysis.

Process

1. Diagram — Create a data flow diagram (DFD) of the system
2. Enumerate — Walk each component/flow and apply STRIDE categories
3. Prioritize — Rate each threat by likelihood × impact
4. Mitigate — Map mitigations to each identified threat

STRIDE Template

## Threat Model: [System Name]

### System Overview
[Brief description and architecture diagram reference]

### Data Flow Diagram
[ASCII or reference to Canvas-generated diagram]

### Threats

#### [Component/Flow Name]

| # | Category | Threat | Likelihood | Impact | Risk | Mitigation |
|---|----------|--------|------------|--------|------|------------|
| T-01 | Spoofing | [Description] | H/M/L | H/M/L | [Score] | [Mitigation] |
| T-02 | Tampering | [Description] | H/M/L | H/M/L | [Score] | [Mitigation] |
| T-03 | Repudiation | [Description] | H/M/L | H/M/L | [Score] | [Mitigation] |
| T-04 | Info Disclosure | [Description] | H/M/L | H/M/L | [Score] | [Mitigation] |
| T-05 | Denial of Service | [Description] | H/M/L | H/M/L | [Score] | [Mitigation] |
| T-06 | Elevation | [Description] | H/M/L | H/M/L | [Score] | [Mitigation] |

STRIDE Per-Element Mapping

2. PASTA (Process for Attack Simulation and Threat Analysis)

Risk-centric, 7-stage methodology. Best for business-aligned assessments.

Seven Stages

PASTA Template

## PASTA Threat Model: [System Name]

### Stage 1: Business Objectives
- Primary business function: [___]
- Regulatory requirements: [___]
- Risk tolerance: [High/Medium/Low]

### Stage 2: Technical Scope
- Components in scope: [___]
- External dependencies: [___]
- Data classification: [___]

### Stage 3: Decomposition
[DFD with trust boundaries]

### Stage 4: Threat Agents
| Agent | Motivation | Capability | Likelihood |
|-------|-----------|------------|------------|
| [External attacker] | [Financial] | [High] | [Medium] |
| [Insider] | [Disgruntlement] | [Medium] | [Low] |

### Stage 5: Vulnerability Correlation
| Component | Known Vulnerabilities | CVE/CWE | Exploitability |
|-----------|----------------------|---------|----------------|
| [___] | [___] | [___] | [___] |

### Stage 6: Attack Scenarios
[Attack trees or kill chain diagrams]

### Stage 7: Risk Ranking
| Finding | Likelihood | Impact | Risk Score | Priority |
|---------|-----------|--------|------------|----------|
| [___] | [___] | [___] | [___] | [___] |

3. MITRE ATT&CK Mapping

Map adversary behaviors to established technique taxonomy.

Key Matrices

ATT&CK Mapping Template

## ATT&CK Map: [System Name]

### Relevant Techniques

| Tactic | Technique ID | Technique Name | Applicability | Test Priority |
|--------|-------------|----------------|---------------|---------------|
| Initial Access | T1190 | Exploit Public-Facing Application | [H/M/L] | [H/M/L] |
| Execution | T1059 | Command and Scripting Interpreter | [H/M/L] | [H/M/L] |
| Persistence | T1078 | Valid Accounts | [H/M/L] | [H/M/L] |
| Privilege Escalation | T1068 | Exploitation for Privilege Escalation | [H/M/L] | [H/M/L] |
| Defense Evasion | T1027 | Obfuscated Files or Information | [H/M/L] | [H/M/L] |
| Credential Access | T1110 | Brute Force | [H/M/L] | [H/M/L] |
| Lateral Movement | T1021 | Remote Services | [H/M/L] | [H/M/L] |
| Exfiltration | T1041 | Exfiltration Over C2 Channel | [H/M/L] | [H/M/L] |

4. Attack Trees

Goal-oriented decomposition. Best for analyzing specific attack objectives.

Structure

[Root Goal: What the attacker wants to achieve]
├── [Sub-goal 1: Attack path A]
│   ├── [Leaf: Specific technique]  [Cost: L] [Difficulty: M]
│   └── [Leaf: Specific technique]  [Cost: M] [Difficulty: H]
├── [Sub-goal 2: Attack path B]
│   ├── [AND node: Requires both]
│   │   ├── [Leaf: Prerequisite 1]
│   │   └── [Leaf: Prerequisite 2]
│   └── [Leaf: Alternative technique]
└── [Sub-goal 3: Attack path C]
    └── [Leaf: Specific technique]  [Cost: H] [Difficulty: L]

Annotations

Framework Selection Guide

INTERACTION_TRIGGERS Question Templates (SKILL.md excerpt)

SCOPE_DEFINITION

questions:
  - question: "What is the scope of this red team assessment?"
    header: "Scope"
    options:
      - label: "Application layer (Recommended)"
        description: "Web/API endpoints, business logic, authentication, authorization, input handling"
      - label: "AI/LLM system"
        description: "Prompt injection, jailbreak, data poisoning, agentic risks, guardrail bypass"
      - label: "Full stack"
        description: "Application + infrastructure + CI/CD + supply chain"
      - label: "Purple Team exercise"
        description: "Collaborative Red/Blue with detection validation and SIEM rule tuning"
    multiSelect: false

FRAMEWORK_SELECTION

questions:
  - question: "Which threat modeling approach should be applied?"
    header: "Framework"
    options:
      - label: "STRIDE (Recommended)"
        description: "Categorize threats by Spoofing/Tampering/Repudiation/Info Disclosure/DoS/Elevation"
      - label: "PASTA"
        description: "Risk-centric 7-step process aligned to business objectives"
      - label: "MITRE ATT&CK mapping"
        description: "Map attack techniques to known adversary TTPs"
      - label: "Attack Trees"
        description: "Goal-oriented tree decomposition of attack paths"
    multiSelect: false

SEVERITY_DISPUTE

questions:
  - question: "How should this finding's severity be classified?"
    header: "Severity"
    options:
      - label: "Critical"
        description: "Remote code execution, auth bypass, or data exfiltration with no user interaction"
      - label: "High"
        description: "Significant impact requiring minimal attacker effort or privilege"
      - label: "Medium"
        description: "Moderate impact requiring specific conditions or elevated access"
      - label: "Low"
        description: "Limited impact, difficult to exploit, or defense-in-depth already mitigates"
    multiSelect: false