supply-chain-security

Typosquatting detection, install script analysis, dependency confusion prevention, and phantom dependency detection for npm/pip.

422 stars

Best use case

supply-chain-security is best used when you need a repeatable AI agent workflow instead of a one-off prompt.

Typosquatting detection, install script analysis, dependency confusion prevention, and phantom dependency detection for npm/pip.

Teams using supply-chain-security should expect a more consistent output, faster repeated execution, less prompt rewriting.

When to use this skill

  • You want a reusable workflow that can be run more than once with consistent structure.

When not to use this skill

  • You only need a quick one-off answer and do not need a reusable workflow.
  • You cannot install or maintain the underlying files, dependencies, or repository context.

Installation

Claude Code / Cursor / Codex

$curl -o ~/.claude/skills/supply-chain-security/SKILL.md --create-dirs "https://raw.githubusercontent.com/vibeeval/vibecosystem/main/skills/supply-chain-security/SKILL.md"

Manual Installation

  1. Download SKILL.md from GitHub
  2. Place it in .claude/skills/supply-chain-security/SKILL.md inside your project
  3. Restart your AI agent — it will auto-discover the skill

How supply-chain-security Compares

Feature / Agentsupply-chain-securityStandard Approach
Platform SupportNot specifiedLimited / Varies
Context Awareness High Baseline
Installation ComplexityUnknownN/A

Frequently Asked Questions

What does this skill do?

Typosquatting detection, install script analysis, dependency confusion prevention, and phantom dependency detection for npm/pip.

Where can I find the source code?

You can find the source code on GitHub using the link provided at the top of the page.

SKILL.md Source

# Supply Chain Security

Patterns for auditing third-party dependencies before they own your production environment.

## Typosquatting Detection

Common substitution patterns attackers use against popular packages:

```python
# Levenshtein distance check against known-good package list
def levenshtein(a: str, b: str) -> int:
    if len(a) < len(b):
        return levenshtein(b, a)
    if not b:
        return len(a)
    prev = list(range(len(b) + 1))
    for i, ca in enumerate(a):
        curr = [i + 1]
        for j, cb in enumerate(b):
            curr.append(min(prev[j + 1] + 1, curr[j] + 1, prev[j] + (ca != cb)))
        prev = curr
    return prev[-1]

POPULAR_PACKAGES = ['express', 'lodash', 'react', 'axios', 'moment', 'chalk']

def is_typosquat(pkg: str, threshold: int = 2) -> list[str]:
    return [p for p in POPULAR_PACKAGES if 0 < levenshtein(pkg, p) <= threshold]

# Examples of known typosquats
typosquats = {
    'lodahs': 'lodash',       # character swap
    'expres': 'express',      # missing char
    'reakt': 'react',         # phonetic substitution
    'axois': 'axios',         # transposition
    'momnet': 'moment',       # transposition
}
```

Common substitution patterns to check manually:
- Character transposition: `lodash` → `lodahs`
- Missing character: `express` → `expres`
- Extra character: `chalk` → `cchalk`
- Hyphen/underscore swap: `my-lib` → `my_lib`
- Homoglyph: `rn` instead of `m` in package name

## Install Script Audit

```bash
# List all packages with install scripts (npm)
npm query ":root > *" | \
  node -e "const d=require('/dev/stdin');
    Object.entries(d.dependencies||{}).forEach(([k,v])=>{
      const pkg = require(\`./node_modules/\${k}/package.json\`);
      if(pkg.scripts?.preinstall||pkg.scripts?.postinstall||pkg.scripts?.install)
        console.log(k, Object.keys(pkg.scripts))
    })"

# Safer: use npm pack --dry-run to see what a package would install
npm pack lodash --dry-run

# Check a specific package's install scripts before installing
npm show <package> scripts
```

```typescript
// Automated install script audit
import { execSync } from 'child_process'
import fs from 'fs'
import path from 'path'

function auditInstallScripts(nodeModulesDir: string): void {
  const packages = fs.readdirSync(nodeModulesDir)

  for (const pkg of packages) {
    const pkgJsonPath = path.join(nodeModulesDir, pkg, 'package.json')
    if (!fs.existsSync(pkgJsonPath)) continue

    const pkgJson = JSON.parse(fs.readFileSync(pkgJsonPath, 'utf-8'))
    const scripts = pkgJson.scripts ?? {}
    const suspicious = ['preinstall', 'postinstall', 'install']
      .filter(s => s in scripts)

    if (suspicious.length > 0) {
      console.warn(`[WARN] ${pkg} has install scripts: ${suspicious.join(', ')}`)
      console.warn(`       ${scripts[suspicious[0]]}`)
    }
  }
}

auditInstallScripts('./node_modules')
```

## Dependency Confusion Risk Assessment

```bash
# Check if your internal package names exist on the public registry
# If they do AND the public version is newer, npm may pull the public one

INTERNAL_PACKAGES=("@mycompany/auth" "@mycompany/utils" "internal-logger")

for pkg in "${INTERNAL_PACKAGES[@]}"; do
  result=$(npm show "$pkg" version 2>/dev/null)
  if [ -n "$result" ]; then
    echo "RISK: $pkg exists on public registry at version $result"
  else
    echo "OK:   $pkg not on public registry"
  fi
done
```

Prevention in `.npmrc`:

```ini
# Force scoped packages to your private registry
@mycompany:registry=https://npm.mycompany.com
@internal:registry=https://npm.mycompany.com

# Block public fallback for scoped packages
//npm.mycompany.com/:always-auth=true
```

## Phantom Dependency Detection

```bash
# Find imports/requires that aren't in package.json
# Node.js (using depcheck)
npx depcheck --ignores="@types/*" .

# Python (using pipreqs)
pip install pipreqs
pipreqs . --print | sort > required.txt
pip freeze | sort > installed.txt
diff required.txt installed.txt
```

```typescript
// Custom phantom dep detector for TypeScript
import { execSync } from 'child_process'
import fs from 'fs'

function detectPhantomDeps(srcDir: string): void {
  const pkgJson = JSON.parse(fs.readFileSync('package.json', 'utf-8'))
  const declared = new Set([
    ...Object.keys(pkgJson.dependencies ?? {}),
    ...Object.keys(pkgJson.devDependencies ?? {}),
  ])

  // Extract all import/require statements
  const imports = execSync(
    `rg --no-heading -oh "from ['\"]([^./][^'\"]+)['\"]" ${srcDir} | sed "s/from ['\"]//;s/['\"]//;s/\\/.*//"`
  ).toString().split('\n').filter(Boolean)

  const unique = new Set(imports)
  for (const imp of unique) {
    const root = imp.startsWith('@') ? imp.split('/').slice(0, 2).join('/') : imp.split('/')[0]
    if (!declared.has(root)) {
      console.warn(`[PHANTOM] "${root}" is imported but not in package.json`)
    }
  }
}
```

## Lock File Integrity Verification

```bash
# Verify package-lock.json integrity (npm)
npm ci --dry-run

# Detect if lock file is out of sync with package.json
npm install --package-lock-only
git diff --exit-code package-lock.json

# pip: verify requirements are pinned with hashes
pip install --require-hashes -r requirements.txt

# Generate hash-pinned requirements
pip-compile --generate-hashes requirements.in -o requirements.txt
```

## npm audit / pip-audit Integration

```bash
# npm - break build on high severity
npm audit --audit-level=high

# pip-audit
pip install pip-audit
pip-audit --requirement requirements.txt --fail-on CRITICAL,HIGH

# GitHub Actions integration
- name: Security audit
  run: |
    npm audit --audit-level=moderate
    npx better-npm-audit audit --level moderate
```

## SBOM Generation

```bash
# CycloneDX for Node.js
npm install -g @cyclonedx/cyclonedx-npm
cyclonedx-npm --output-format JSON --output-file sbom.json

# CycloneDX for Python
pip install cyclonedx-bom
cyclonedx-bom -r -o sbom.xml

# SPDX via syft
syft . -o spdx-json > sbom.spdx.json
```

## Known Malicious Package Indicators

Red flags when reviewing a new dependency:

```
- Install scripts that curl | bash external URLs
- Package published < 48 hours ago with thousands of weekly downloads
- Zero issues/stars on GitHub but high download count
- package.json "main" points to an obfuscated file
- Dependencies on packages with random hex names
- Author email is a free provider with a new account
- Version bump from 1.x to 99.x (dependency confusion attack)
```

## Dependency Pinning Strategy

```json
// package.json: exact pinning for production deps
{
  "dependencies": {
    "express": "4.18.2",      // exact - no caret/tilde
    "lodash": "4.17.21"
  },
  "devDependencies": {
    "typescript": "^5.3.0"   // range OK for dev tools
  }
}
```

```bash
# Convert all ranges to exact versions
npm shrinkwrap
# or use: npm-shrinkwrap.json over package-lock.json for published packages

# Renovate bot for automated safe updates with PR-per-dep
# renovate.json
{
  "extends": ["config:base"],
  "automerge": false,
  "reviewers": ["security-team"]
}
```

## Dependency Risk Scoring

Score each dependency on a 0-100 risk scale before adoption:

| Signal | Low Risk (0-2) | Medium Risk (3-5) | High Risk (6-10) |
|--------|---------------|-------------------|------------------|
| Age | > 3 years | 1-3 years | < 1 year |
| Maintainers | 3+ active | 1-2 active | 1 inactive |
| Downloads | > 1M/week | 100K-1M/week | < 100K/week |
| GitHub stars | > 5K | 500-5K | < 500 |
| Last commit | < 3 months | 3-12 months | > 12 months |
| Open issues | < 50 | 50-200 | > 200 unresolved |
| Install scripts | None | postinstall (build) | preinstall + network |
| Dependencies | < 5 | 5-20 | > 20 transitive |
| CVE history | 0 unfixed | Fixed promptly | Unfixed > 30 days |
| License | MIT/Apache/BSD | LGPL/MPL | GPL/AGPL/Unknown |

### Risk Assessment Template

```
Package: <name>@<version>
Purpose: <why we need it>
Alternatives: <what else we considered>

Signals:
  Age: X years (score: N)
  Maintainers: X active (score: N)
  Downloads: X/week (score: N)
  Last commit: X days ago (score: N)
  Dependencies: X transitive (score: N)
  CVEs: X unfixed (score: N)

Total Risk Score: XX/100
Decision: [ADOPT | REVIEW | REJECT]
Reviewed by: <name>
Date: <date>
```

### Automated Risk Check

```bash
# npm: check package metadata
npm show <package> time modified maintainers repository

# Check download stats
curl -s "https://api.npmjs.org/downloads/point/last-week/<package>" | jq .downloads

# Check for known vulnerabilities
npm audit --json | jq '.vulnerabilities["<package>"]'

# Python: check PyPI metadata
pip show <package>
pip-audit -r requirements.txt --format json
```

**Rule**: Never `npm install <unknown-package>` in production without running `npm audit`, checking install scripts, and verifying the package name against known typosquats.

Dependency risk scoring adapted from [Trail of Bits](https://github.com/trailofbits/skills) supply-chain-risk-auditor plugin.

Related Skills

springboot-security

422
from vibeeval/vibecosystem

Spring Security best practices for authn/authz, validation, CSRF, secrets, headers, rate limiting, and dependency security in Java Spring Boot services.

security

422
from vibeeval/vibecosystem

Security audit workflow - vulnerability scan → verification

security-review

422
from vibeeval/vibecosystem

Use this skill when adding authentication, handling user input, working with secrets, creating API endpoints, or implementing payment/sensitive features. Provides comprehensive security checklist and patterns.

mcp-chaining

422
from vibeeval/vibecosystem

Research-to-implement pipeline chaining 5 MCP tools with graceful degradation

django-security

422
from vibeeval/vibecosystem

Django security best practices, authentication, authorization, CSRF protection, SQL injection prevention, XSS prevention, and secure deployment configurations.

config-security-scan

422
from vibeeval/vibecosystem

Scan .claude/ directory for security misconfigurations, exposed secrets, unsafe permissions

concurrency-security

422
from vibeeval/vibecosystem

TOCTOU prevention, distributed locking, idempotency keys, race condition detection for Node.js and serverless environments.

workflow-router

422
from vibeeval/vibecosystem

Goal-based workflow orchestration - routes tasks to specialist agents based on user goals

wiring

422
from vibeeval/vibecosystem

Wiring Verification

websocket-patterns

422
from vibeeval/vibecosystem

Connection management, room patterns, reconnection strategies, message buffering, and binary protocol design.

visual-verdict

422
from vibeeval/vibecosystem

Screenshot comparison QA for frontend development. Takes a screenshot of the current implementation, scores it across multiple visual dimensions, and returns a structured PASS/REVISE/FAIL verdict with concrete fixes. Use when implementing UI from a design reference or verifying visual correctness.

verification-loop

422
from vibeeval/vibecosystem

Comprehensive verification system covering build, types, lint, tests, security, and diff review before a PR.