Understanding Layers

The Code Property Graph (CPG) is a multi-layered representation of each unique code version of your application. The most important layers are the base layer, the two default layers semanticcpg and tagging, and the securityprofile layer.

You work with layers in your workspace. Layers are frequently referred to as "overlays".

Additional information on CPG layers is provided in the API.

base Layer

For each language supported by ShiftLeft Ocular, the language frontend component translates code written in that language into a CPG through the base layer. For example, the JavaLanguageFrontend translates a JAR into a CPG.

The base layer is the foundation of the CPG, but is not sufficiently expressive for the purpose of examining data flows and vulnerabilities.

Default Layers

The default layers are automatically created and loaded into memory when you create a CPG, or load an existing CPG.

semanticcpg Layer

The semanticcp layer links together the methods and type declaration. Using the base layer and semantic information, the semanticcp layer

  • introduces additional information based on framework specific passes.

  • adds data flow semantics to the CPG as specified by Policies.

tagging Layer

The tagging layer identifies possible attacker-controlled data sources and interesting sinks as specified by Policies.

The securityprofile Layer

The securityprofile layer is dervied from Policies, and summarizes the vulnerabilities and data leaks present in your code. Specifically, it describes:

  • Read operations

  • Write operations

  • Transformations

  • Security-relevant data flows

  • Information flows

  • Findings

Unlike the base and default layers, because the Security Profile is so expensive at runtime, it must be intentially created and loaded.