attack-tree-generation

name	Attack Tree Generation
description	This skill should be used when building attack trees, when the user mentions "attack tree", "attack path", "AND/OR tree", "threat tree", "multi-step attack", or needs to construct structured attack decompositions.
version	1.0.0

Attack Tree Construction

Purpose

Build structured AND/OR attack trees that decompose high-level threat goals into concrete, actionable attack steps aligned with the VerSprite model.

When to Use

• During PASTA Stage 4 (draft attack trees)
• During PASTA Stage 6 (complete attack trees)
• When analyzing multi-step attack scenarios
• When visualizing attack paths for stakeholders

Tree Structure

Node Types

Type	Symbol	Meaning	Probability Rule
AND	All children required	Attacker must complete ALL sub-steps	P = P(child1) * P(child2) * ...
OR	Any child sufficient	Attacker can use ANY sub-path	P = max(P(child1), P(child2), ...)
LEAF	Terminal step	Concrete technique, no decomposition	P from 5-factor model

Node Roles (VerSprite Alignment)

Level	Role	Description	Example
0	threat-motive	Root: attacker's goal	"Steal payment card data"
1	threat-agent	Who performs the attack	"External attacker (financially motivated)"
2	target	What is being attacked	"Payment API, card database"
3	attack-vector	Delivery mechanism	"Network access, social engineering"
4	attack-pattern	Specific CAPEC pattern	"CAPEC-66: SQL Injection"

Construction Process

Step 1: Define the root (threat-motive)

Start with the threat scenario goal from STRIDE analysis.

Step 2: Identify threat agents (level 1)

Who would pursue this goal? Multiple agents = OR node (any agent can attack).

Step 3: Map targets (level 2)

What components must be compromised? Multiple sequential targets = AND node.

Step 4: Enumerate attack vectors (level 3)

How can each target be reached? Multiple vectors = OR node (alternative paths).

Step 5: Decompose to attack patterns (level 4)

What specific technique is used? Reference ATT&CK T-codes and CAPEC patterns.

Mermaid Diagram Format

graph TD
    root["<b>Steal User Data</b><br/>threat-motive<br/>P=0.52"]

    root -->|OR| path1["SQL Injection Path<br/>P=0.52"]
    root -->|OR| path2["Credential Theft Path<br/>P=0.38"]

    path1 -->|AND| step1a["Discover injection point<br/>T1190<br/>P=0.8"]
    path1 -->|AND| step1b["Extract data<br/>T1005<br/>P=0.65"]

    path2 -->|AND| step2a["Brute-force login<br/>T1110<br/>P=0.6"]
    path2 -->|AND| step2b["Escalate to admin<br/>T1068<br/>P=0.4"]
    path2 -->|AND| step2c["Export user table<br/>T1567<br/>P=0.7"]

    style root fill:#ff6b6b
    style path1 fill:#ffa07a
    style path2 fill:#ffa07a

Probability Calculation

For the SQL Injection path (AND):

• P = 0.8 * 0.65 = 0.52

For the Credential Theft path (AND):

• P = 0.6 * 0.4 * 0.7 = 0.168 ≈ 0.17

Root (OR): P = max(0.52, 0.17) = 0.52

Depth Limits

• Maximum 10 levels (prevent infinite decomposition)
• Each LEAF must reference a specific ATT&CK technique or CAPEC pattern
• Each LEAF should have a code evidence anchor where possible

Quality Checklist

• Root node clearly states the attacker's goal
• AND/OR logic is correct at each level
• Probability propagation follows the rules
• Every LEAF has an ATT&CK technique reference
• Critical paths (highest probability) are highlighted
• Detection opportunities are noted at each step