graphize-appsec

README

Graphize-AppSec

Security reachability analysis using graphize code knowledge graphs.

Answers the question: "Is this vulnerability actually exploitable in my deployment?"

Overview

Graphize-AppSec performs reachability analysis to reduce vulnerability noise by 90%+. Instead of alerting on every CVE in your dependency tree, it determines which vulnerabilities are actually reachable from your code's entry points.

This is whitebox analysis - it requires source code access to build precise call graphs and trace execution paths. Unlike blackbox scanning (which only sees external behavior), whitebox analysis can definitively prove when vulnerable code is unreachable.

SBOM Scanner Output     graphize-appsec          Actionable Results
┌──────────────────┐    ┌──────────────────┐    ┌──────────────────┐
│ 847 CVEs found   │ -> │ Reachability     │ -> │ 12 exploitable   │
│ (97% false pos)  │    │ Analysis         │    │ 835 not affected │
└──────────────────┘    └──────────────────┘    └──────────────────┘

Features

• 🔬 16 Reachability Tests - Systematic assessment across Reachable, Exploitable, and Damage categories
• 📋 VEX Generation - CycloneDX VEX statements for non-exploitable vulnerabilities
• 📦 SBOM Enrichment - Enrich existing SBOMs with exploitability context
• 🕸️ Graph-Based Analysis - Leverages graphize's code knowledge graph for path finding
• 📊 Structured Reports - Machine-readable reports via structured-evaluation

Installation

go install github.com/plexusone/graphize-appsec/cmd/graphize-appsec@latest

Quick Start

# 1. Build code graph with graphize
graphize init
graphize add .
graphize analyze

# 2. Generate SBOM (using Syft, Trivy, or similar)
syft . -o cyclonedx-json > sbom.json

# 3. Get vulnerability list (using Grype or similar)
grype sbom:sbom.json -o json > vulns.json

# 4. Enrich SBOM with reachability analysis
graphize-appsec vex enrich \
  --sbom sbom.json \
  --vulns vulns.json \
  --output sbom-vex.json

Reachability Tests

Category: Reachable (7 tests)

ID	Test	Question
REACH-001	Dependency Imported	Is vulnerable package in dependency graph?
REACH-002	Dependency Used	Is vulnerable code actually called?
REACH-003	Exposed by API	Is vuln reachable from public API?
REACH-004	Direct Dependency	Is this direct (not transitive)?
REACH-005	Public Repository	Is vuln in public repo?
REACH-006	Application Layer	Is vuln in app layer (not infra)?
REACH-007	Cloud Deployed	Is container with vuln running?

Category: Exploitable (6 tests)

ID	Test	Question
EXPLOIT-001	Weak Cryptography	Does path involve weak crypto?
EXPLOIT-002	Community Buzz	Active exploitation discussion?
EXPLOIT-003	Extensive Patching	Multiple patch iterations?
EXPLOIT-004	Multiple Public Exploits	Public exploits available?
EXPLOIT-005	EPSS Low Risk	EPSS score < 0.1?
EXPLOIT-006	AI Unexploitable	AI analysis says unexploitable?

Category: Damage (3 tests)

ID	Test	Question
DAMAGE-001	Critical Business Priority	Affects critical systems?
DAMAGE-002	Login Management	Affects auth/login?
DAMAGE-003	CVSS High Severity	CVSS >= 7.0?

VEX Output

For vulnerabilities determined to be non-exploitable, graphize-appsec generates VEX statements:

{
  "vulnerabilities": [
    {
      "id": "CVE-2021-44228",
      "analysis": {
        "state": "not_affected",
        "justification": "code_not_reachable",
        "detail": "No call path from public API to vulnerable JNDI lookup function.",
        "response": ["will_not_fix"]
      }
    }
  ]
}

Architecture

┌─────────────────────────────────────────────────────┐
│                   graphize-appsec                   │
│  ┌─────────────┐  ┌─────────────┐  ┌─────────────┐  │
│  │ Reachable   │  │ Exploitable │  │   Damage    │  │
│  │  (7 tests)  │  │  (6 tests)  │  │  (3 tests)  │  │
│  └──────┬──────┘  └──────┬──────┘  └──────┬──────┘  │
│         └────────────────┼────────────────┘         │
│                          ▼                          │
│                  ┌───────────────┐                  │
│                  │  VEX/Report   │                  │
│                  │   Generator   │                  │
│                  └───────────────┘                  │
└─────────────────────────┬───────────────────────────┘
                          │
            ┌─────────────┴─────────────┐
            ▼                           ▼
    ┌───────────────┐           ┌───────────────┐
    │   graphize    │           │    graphfs    │
    │ (code graph)  │           │  (traversal)  │
    └───────────────┘           └───────────────┘

CLI Commands

# Check environment prerequisites
graphize-appsec doctor

# Assess a specific vulnerability
graphize-appsec assess CVE-2021-44228

# List available reachability tests
graphize-appsec test list

# Generate VEX from SBOM + vulnerabilities
graphize-appsec vex enrich --sbom sbom.json --vulns vulns.json

# Generate standalone VEX document
graphize-appsec vex generate --vulns vulns.json

Dependencies

Package	Purpose
graphize	Produces code knowledge graph
graphfs	Graph storage and traversal
structured-evaluation	Report generation
cyclonedx-go	VEX/SBOM format

Documentation

Full Documentation | Getting Started | CLI Reference

• Getting Started Guide - Installation and quick workflow
• CLI Reference - All commands, flags, and examples
• Reachability Tests Reference - All 16 tests documented
• VEX Output Reference - CycloneDX VEX format
• SBOM Governance Best Practices
• Parser Landscape
• Design Documents

Related Projects

• graphize - Code knowledge graph builder
• graphize-grafana - Example implementation

License

MIT