threat-modeling

Threat Modeling Agent Skill

Purpose: Automated threat modeling from code analysis using STRIDE/DREAD methodologies

When to use: Security assessment, architecture review, compliance documentation, attack scenario planning

---

Workflow

1. Scope Determination

Determine analysis scope:

• Full repository threat model
• Feature-specific threat model
• Component-level threat model
• Architecture diagram analysis

Check prerequisites:

• Repository is indexed in knowledge graph
• Business context available (optional)
• Architecture diagrams uploaded (optional)
• CVE intelligence data available (optional)

2. Architecture Extraction

Extract system architecture from:

• Knowledge graph clusters (components)
• Call chains (data flows)
• Entry points (public APIs)
• Trust boundaries (zone crossings)
• Technology stack
• External dependencies

Enrich with:

• Architecture diagram analysis (AI vision)
• Business context (industry, compliance)
• Deployment context (cloud infrastructure)

3. STRIDE Threat Analysis

Generate threats for each category:

• Spoofing: Identity and authentication threats
• Tampering: Data and code integrity threats
• Repudiation: Audit and accountability threats
• Information Disclosure: Data leakage threats
• Denial of Service: Availability threats
• Elevation of Privilege: Access control threats

For each threat, identify:

• Affected components
• Affected data flows
• Trust boundary crossings
• Attack vectors
• Prerequisites
• Detection methods

4. DREAD Risk Assessment

Score each threat (1-10):

• Damage Potential
• Reproducibility
• Exploitability
• Affected Users
• Discoverability

Calculate:

• Risk Score = Average of 5 dimensions
• Risk Level = Critical/High/Medium/Low

5. Phoenix Intelligence Enrichment

Enrich with Phoenix data (IP-protected):

• Related CVEs
• Phoenix proprietary scores (high-level only)
• Threat actor data
• MITRE ATT&CK techniques
• CAPEC attack patterns