vulnhawk 0.2.0

AI-powered code security scanner that finds vulnerabilities Semgrep and CodeQL miss.

Quick Start • GitHub Action • Comparison • Languages • FAQ

The Problem

Traditional SAST tools rely on pattern matching and AST rules. They excel at catching known vulnerability patterns, but they fundamentally cannot reason about intent.

If your API has 20 endpoints and 19 of them verify authorization before acting on a resource, Semgrep has no way to flag the one that doesn't - because there is no pattern to match against. The vulnerability is the absence of a pattern.

The Solution

VulnHawk analyzes code with AI, and for every piece of code it examines, it includes related code from elsewhere in your codebase as context. This enrichment step lets the AI compare how similar components handle security - and spot the one that doesn't.

Quick Start

pip install vulnhawk

Choose a backend:

# Claude Code CLI - FREE for subscribers (recommended)
vulnhawk scan ./src -b claude-code

# Codex CLI - FREE for ChatGPT Pro/Plus subscribers
vulnhawk scan ./src -b codex

# Claude API
export ANTHROPIC_API_KEY=sk-ant-...
vulnhawk scan ./src

# OpenAI API
vulnhawk scan ./src -b openai -m gpt-4o

# Ollama - free, local, fully private
vulnhawk scan ./src -b ollama -m llama3.1

No config files. No rules to write. No database to build.

Claude Code and Codex backends are free for users with existing subscriptions. VulnHawk pipes prompts through your local CLI, so there are no additional API costs.

VulnHawk vs Other SAST Tools

_{*Free with Claude Code, Codex CLI, or Ollama. API backends cost ~$0.50-$2.00 per scan.}

What VulnHawk finds that others cannot

Recommended tool combination

VulnHawk is designed as a complementary layer, not a replacement:

Usage

Scan modes

vulnhawk scan ./src                      # Full scan (default)
vulnhawk scan ./src --mode auth          # Auth bypass, missing checks, session flaws
vulnhawk scan ./src --mode injection     # SQLi, command injection, SSTI, XSS
vulnhawk scan ./src --mode secrets       # Hardcoded keys, tokens, passwords
vulnhawk scan ./src --mode config        # Debug mode, permissive CORS, insecure cookies
vulnhawk scan ./src --mode crypto        # Weak hashing, hardcoded keys, bad RNG

Output formats

vulnhawk scan ./src -o json -f results.json        # JSON
vulnhawk scan ./src -o sarif -f results.sarif       # SARIF (GitHub Code Scanning)
vulnhawk scan ./src -o markdown -f report.md        # Markdown report

Severity filter

vulnhawk scan ./src --severity high      # Critical + High only
vulnhawk scan ./src --severity info      # Everything

SARIF input - chain with other tools

Feed Semgrep, CodeQL, or any SARIF-producing tool's output into VulnHawk. It uses those findings as additional context to validate, expand, and chain them into deeper vulnerabilities.

# Run Semgrep first, then enrich with VulnHawk
semgrep --config auto ./src -o semgrep.sarif --sarif
vulnhawk scan ./src --sarif-input semgrep.sarif

What this enables:

• Validates whether other tools' findings are real or false positives
• Discovers related vulnerabilities near flagged locations
• Builds multi-step attack chains connecting findings across tools
• Checks whether suggested fixes address the actual root cause

Dry run

vulnhawk info ./src    # Preview files, chunks, and language breakdown

GitHub Action

VulnHawk runs as a baseline scan on your default branch and incrementally on every pull request.

Recommended setup

name: VulnHawk Security Scan
on:
  push:
    branches: [main, master]
  pull_request:

permissions:
  security-events: write
  contents: read

jobs:
  vulnhawk:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: momenbasel/vulnhawk@main
        with:
          target: '.'
          backend: 'claude-code'
          claude-code-oauth-token: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
          severity: 'medium'
          fail-on-findings: 'true'

Findings are automatically uploaded to GitHub's Security > Code Scanning tab via SARIF.

Backend options

backend: 'claude-code'
claude-code-oauth-token: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}

backend: 'codex'

api-key: ${{ secrets.ANTHROPIC_API_KEY }}

backend: 'openai'
api-key: ${{ secrets.OPENAI_API_KEY }}

Chaining with Semgrep in CI

steps:
  - uses: actions/checkout@v4

  - name: Semgrep (fast pattern scan)
    uses: returntocorp/semgrep-action@v1
    with:
      config: auto
      generateSarif: true

  - name: VulnHawk (deep AI analysis)
    uses: momenbasel/vulnhawk@main
    with:
      target: '.'
      backend: 'claude-code'
      claude-code-oauth-token: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
      sarif-input: 'semgrep.sarif'

Supported Languages

How It Works

Codebase ──> Discover ──> Chunk ──> Enrich ──> Analyze ──> Validate ──> Report
                │            │          │           │           │
           Respects      Functions  Related     LLM with    Dedup +
           .gitignore   Classes    code from   security    confidence
           .vulnhawk-   Routes     same dir +  prompts     scoring
           ignore       Modules    auth patterns

The enrichment step is the core differentiator. For each code chunk, VulnHawk includes:

• Other functions/routes from the same directory
• Auth middleware and guard patterns from across the codebase

This gives the AI the context it needs to identify inconsistencies.

Configuration

.vulnhawkignore

Exclude paths from scanning (gitignore syntax):

generated/
vendor/
third_party/
*.gen.go

Environment variables

Cost

FAQ

Contributing

See CONTRIBUTING.md.

git clone https://github.com/momenbasel/vulnhawk.git
cd vulnhawk
uv venv .venv && source .venv/bin/activate
uv pip install -e ".[dev]"
pytest

Support the Project

If VulnHawk is useful to you, consider sponsoring the project to support continued development.

License

VulnHawk is source-available under the VulnHawk License.

Free for everyone - individuals, teams, startups, and enterprises may use VulnHawk at no cost for internal security scanning, provided it is installed from an official distribution channel:

• GitHub Marketplace
• PyPI
• This repository

You may not sell the Software, offer it as a competing service, or redistribute forks as a product. Forks are permitted solely for submitting pull requests back to this repository. See LICENSE for full terms.