Azure DevOps

This article shows you how to integrate ShiftLeft CORE's NG SAST into your Azure DevOps workflow to provide automated code analysis. We provide two sets of instructions based on how you define your Azure pipelines: using YAML syntax or the Classic interface.

For Azure Pipelines defined using YAML syntax

This tutorial assumes that you have an existing YAML-base Azure Pipeline defined. You will be adding the tasks required to integrate NG SAST to this file.

Step 1: Create your secret variables

You will need to create secret variables to store authentication information for ShiftLeft.

We recommend creating your secret variables using variable groups since this method balances security with ease of deployment across multiple repositories.

Add variable group

However, you could also create your secret variables using:

When creating a variable group, we recommend calling it something like shiftleft-token. You can then provide your access token as SHIFTLEFT_ACCESS_TOKEN.

When running in a production environment, we recommend that you use a CI config token as the access token. You can create your CI config token in the ShiftLeft Dashboard.

Add ShiftLeft tokens

At this point, you can refer directly to this group in the pipeline's YAML configuration file using the group property under the variables section.

variables:
- group: shiftleft-token

Step 2: Add ShiftLeft to your pipeline

You will need to include instructions in your Pipeline to download the ShiftLeft CLI so that the Pipeline can run NG SAST.

If you're running Windows, you can do so using a PowerShell task:

- task: PowerShell@2
displayName: Download ShiftLeft cli
inputs:
targetType: 'inline'
script: |
Invoke-WebRequest -Uri 'https://cdn.shiftleft.io/download/sl-latest-windows-x64.zip' -OutFile $(Agent.HomeDirectory)\sl.zip
Expand-Archive -Path $(Agent.HomeDirectory)\sl.zip -DestinationPath $(Agent.HomeDirectory)\

If you're running Linux or macOS, you can use a script task:

- task: CmdLine@2
displayName: Download ShiftLeft cli
inputs:
targetType: 'inline'
script: |
curl https://cdn.shiftleft.io/download/sl > $(Agent.HomeDirectory)/sl && chmod a+rx $(Agent.HomeDirectory)/sl

Step 3: Invoke NG SAST for code analysis

The following sections will show you how to analyze your Java or C# applications.

When invoking NG SAST, you need to refer to the SHIFTLEFT_ACCESS_TOKEN variable. As suggested earlier in this article, if you create a variable group, these variables will be available automatically to all of your Pipelines.

Analyzing a Java application

The following examples show how you can build your Java application (which is required before NG SAST can analyze your code), then use the ShiftLeft CLI to invoke NG SAST for code analysis.

Please note that the ShiftLeft CLI requires Java 8 to create the Code Property Graph (CPG) representation of your source code. This is necessary before code analysis unless you are analyzing a project written in C#. If so, you can skip this step.

To set up Java 8, use the Java Tool Installer.

- task: JavaToolInstaller@0
inputs:
versionSpec: '8'
jdkArchitectureOption: 'x64'
jdkSourceOption: 'PreInstalled'

On Windows:

- task: Maven@3
inputs:
mavenPomFile: 'pom.xml'
mavenOptions: '-Xmx3072m'
javaHomeOption: 'JDKVersion'
jdkVersionOption: '1.8'
jdkArchitectureOption: 'x64'
publishJUnitResults: false
goals: 'package'
- task: CmdLine@2
displayName: Analyze with NG SAST
inputs:
script: |
$(Agent.HomeDirectory)\sl.exe analyze --app ShiftLeftJavaAzWin --tag branch=$(Build.SourceBranchName) --java target/hello-shiftleft-0.0.1.jar
workingDirectory: '$(Build.SourcesDirectory)'
env:
SHIFTLEFT_ACCESS_TOKEN: $(SHIFTLEFT_ACCESS_TOKEN)

On Linux/macOS:

- task: CmdLine@2
displayName: Analyze with NG SAST
inputs:
script: |
$(Agent.HomeDirectory)/sl analyze --app ShiftLeftJava --tag branch=$(Build.SourceBranchName) --java target/hello-shiftleft-0.0.1.jar
workingDirectory: '$(Build.SourcesDirectory)'
env:
SHIFTLEFT_ACCESS_TOKEN: $(SHIFTLEFT_ACCESS_TOKEN)

Analyzing a C# application

The following examples show you how to build your .NET Core application, then use the ShiftLeft CLI to invoke NG SAST for code analysis.

- task: DotNetCoreCLI@2
displayName: Build console app
inputs:
command: 'build'
projects: '$(Build.SourcesDirectory)\netcoreConsole'
- task: CmdLine@2
displayName: Analyze with NG SAST
inputs:
script: |
$(Agent.HomeDirectory)\sl.exe analyze --app netcoreConsole --tag branch=$(Build.SourceBranchName) --csharp --dotnet netcoreConsole/netcoreConsole.csproj
workingDirectory: '$(Build.SourcesDirectory)'
env:
SHIFTLEFT_ACCESS_TOKEN: $(SHIFTLEFT_ACCESS_TOKEN)

For .NET applications, plass in the .sln file instead of the .csproj file:

- task: CmdLine@2
displayName: Analyze with NG SAST
inputs:
script: |
$(Agent.HomeDirectory)\sl.exe analyze --app netfwWebapi --tag branch=$(Build.SourceBranchName) --csharp netfwWebapi/netfwWebapi.sln
workingDirectory: '$(Build.SourcesDirectory)'
env:
SHIFTLEFT_ACCESS_TOKEN: $(SHIFTLEFT_ACCESS_TOKEN)

To pass additional information to Azure, you can use predefined variables that you then reference in the env block of your script.

Recursively finding and scanning your solution/project files

The following example shows you how to modify the sl analyze invocation to recursively find all .sln files and scan them with ShiftLeft:

# Recursively find all .sln files and scan them with ShiftLeft CORE
# Be sure to change the app.group value from "test-appgroup" to your preferred name
Get-ChildItem -Path . -Filter *.sln -Recurse -ErrorAction SilentlyContinue -Force | ForEach-Object {
sl.exe analyze --csharp --oss-project-dir $($_.Directory) --tag app.group=test-appgroup --app $($_.Name -replace '.sln', '') $($_.FullName)
}

Alternatively, you can recursively find and scan all .csproj files if .sln-based scans are taking too long:

# Recursively find all .csproj files and scan them with ShiftLeft CORE
# Use this when .sln based scans are taking too long
# Be sure to change the app.group value from ß"test-appgroup" to your preferred name
Get-ChildItem -Path . -Filter *.csproj -Recurse -ErrorAction SilentlyContinue -Force | ForEach-Object {
sl.exe analyze --csharp --oss-project-dir $($_.Directory) --tag app.group=test-appgroup --app $($_.Name -replace '.csproj', '') $($_.FullName)
}

Running sl check-analysis in Azure DevOps

ShiftLeft's check-analysis feature allows you to compare your analysis results against a set of build rules you've defined. To use check-analysis in your Azure DevOps workflow, you must first enable build vaidation on your branch.