VYPR
Knowledge base

The VYPR taxonomy.

Six orthogonal dimensions for reasoning about a vulnerability — why it exists, what is wrong, how it is exploited, where it happens, what damage it does, and how easy it is to weaponize. This is the foundation for VYPR’s CVE intelligence, AI classification, and prioritization.

Dimension 1 · WHY it exists

Root cause

13 categories · 52 leaves

The underlying flaw class — what the developer or operator did wrong. Distinct from the symptom (weakness) and the technique used to exploit it.

Input handling

  • Missing validation
  • Improper sanitization
  • Type confusion
  • Encoding error
  • Canonicalization error

Authentication

  • Weak auth mechanism
  • Credential handling
  • Session management
  • Token handling
  • MFA bypass

Authorization

  • Missing access control
  • Broken object-level auth
  • Broken function-level auth
  • Privilege boundary failure
  • Tenant isolation failure

Memory safety

  • Out of bounds
  • Use after free
  • Double free
  • Uninitialized memory
  • Null pointer

Configuration

  • Insecure default
  • Exposed service
  • Misconfigured policy
  • Debug feature enabled
  • Overpermissive setting

Cryptography

  • Weak algorithm
  • Bad key management
  • Insecure randomness
  • Improper encryption usage
  • Certificate validation failure

Logic error

  • Business logic bypass
  • State desync
  • Race condition
  • Time-of-check / time-of-use
  • Inconsistent state

Resource management

  • Resource leak
  • Unbounded resource use
  • Improper cleanup

Trust boundary

  • Trusted data assumption
  • Cross-boundary validation failure

Parsing & serialization

  • Unsafe parser
  • Ambiguous grammar
  • Format confusion

Numeric & calculation

  • Integer overflow
  • Precision loss
  • Rounding error

Supply chain

  • Untrusted dependency
  • Tampered artifact
  • Signature validation failure

Isolation boundary

  • Sandbox failure
  • Container isolation failure
  • VM escape condition
Dimension 2 · WHAT is wrong

Weakness type

14 categories · 42 leaves

The shape of the flaw as it manifests in the code or system. Maps loosely onto CWE families but normalized for cross-language consistency.

Injection

  • SQL
  • Command
  • Template
  • LDAP
  • XPath
  • NoSQL

Memory corruption

  • Buffer overflow
  • Heap corruption
  • Stack corruption
  • Integer overflow
  • Format string

Deserialization

  • Insecure deserialization
  • Object injection

Access control

  • Broken access control
  • IDOR
  • Privilege escalation
  • Auth bypass

Data exposure

  • Sensitive data exposure
  • Information leak
  • Metadata exposure

Configuration

  • Misconfiguration
  • Exposed endpoint
  • Insecure header

Concurrency

  • Race condition
  • Deadlock
  • Thread safety

Crypto failure

  • Weak crypto
  • Plain-text storage
  • Broken TLS

Input / output

  • Path traversal
  • File inclusion
  • Uncontrolled redirect

Parsing

  • Unsafe parser
  • Format confusion

Numeric

  • Integer overflow
  • Underflow

Protocol

  • Protocol state error
  • Handshake failure

Supply chain

  • Dependency confusion
  • Package tampering

Isolation

  • Sandbox escape
  • Container escape
Dimension 3 · HOW exploited

Attack technique

12 categories · 37 leaves

The technique an attacker uses to weaponize the weakness. Aligned with MITRE ATT&CK / CAPEC at the leaf level where there is a clean mapping.

Injection attacks

  • SQL injection
  • Command injection
  • SSTI
  • LDAP injection
  • NoSQL injection

Scripting

  • XSS reflected
  • XSS stored
  • XSS DOM
  • Script gadget

Request manipulation

  • CSRF
  • Parameter tampering
  • HTTP request smuggling
  • Host header injection

SSRF & network

  • SSRF
  • DNS rebinding
  • Internal port scanning

Execution

  • RCE
  • Code injection
  • DLL hijacking

Auth attacks

  • Brute force
  • Credential stuffing
  • Password spraying
  • Session hijacking

Privilege abuse

  • Privilege escalation
  • Sandbox escape
  • Container escape

Data extraction

  • Data exfiltration
  • Side channel

Supply chain attacks

  • Dependency confusion
  • Malicious package

Protocol attacks

  • Protocol desync
  • Downgrade attack

Side channel

  • Timing attack
  • Cache attack
  • Speculative execution

File & path

  • Path traversal
  • File inclusion
Dimension 4 · WHERE it happens

Asset context

11 categories · 36 leaves

The component or layer that contains the flaw. Used to focus prioritization on what is actually deployed in a given environment.

Application

  • Web app
  • Mobile app
  • Desktop app
  • CLI app

API layer

  • REST API
  • GraphQL
  • gRPC
  • Internal API

Data layer

  • Database
  • Cache
  • Search engine
  • File storage

Infrastructure

  • OS
  • Container
  • VM
  • Kubernetes

Cloud

  • IAM
  • Storage bucket
  • Serverless
  • Managed DB
  • Message queue

Network

  • Load balancer
  • Proxy
  • Firewall
  • DNS

Hardware

  • CPU
  • Firmware
  • IoT device

Identity

  • Auth provider
  • Federation

CI / CD

  • Build pipeline
  • Artifact registry

Observability

  • Logging system
  • Metrics system

Edge

  • CDN
  • API gateway
Dimension 5 · WHAT damage

Impact

10 categories · 25 leaves

The consequence if the flaw is exploited. Spans CIA, account-level outcomes, business-level damage, and post-exploitation effects.

Confidentiality

  • Data exfiltration
  • Information disclosure
  • Secret leak

Integrity

  • Data modification
  • Unauthorized action
  • Code tampering

Availability

  • Denial of service
  • Resource exhaustion
  • Service degradation

Account

  • Account takeover
  • Session compromise
  • Identity impersonation

System

  • Remote code execution
  • Full system compromise
  • Persistence

Financial

  • Fraud
  • Transaction manipulation

Reputation

  • Trust loss
  • Brand damage

Lateral movement

  • Internal spread
  • Pivoting

Defense evasion

  • Log tampering
  • Detection bypass

Compliance

  • Regulatory violation
  • Data residency breach
Dimension 6 · HOW easy

Exploitability

6 categories · 19 leaves

How readily the flaw can be exploited in practice. Aligned with CVSS metrics for vector, complexity, privileges, and user interaction, plus exploit maturity and reachability.

Attack vector

  • Network
  • Adjacent
  • Local
  • Physical

Complexity

  • Low
  • Medium
  • High

Privileges required

  • None
  • User
  • Admin
  • System

User interaction

  • None
  • Required

Exploit maturity

  • Theoretical
  • Proof of concept
  • Weaponized
  • In the wild

Reachability

  • Direct
  • Indirect