🛡️ titanshield

AI-powered security scanner. Zero config. Under 2 minutes.

Scans your codebase for security vulnerabilities before you deploy — using fast static analysis + Gemini AI deep analysis. Explains every issue in plain English and shows you the exact fix.

Quick Start

# Scan your project right now (static analysis — instant, no key required)
npx titanshield scan ./src

# Full AI scan (Gemini finds what static patterns miss)
npx titanshield scan ./src --key YOUR_GEMINI_API_KEY

# Or set the key as an env var (recommended)
export GEMINI_API_KEY=AIzaSy...
npx titanshield scan ./src

Free Gemini key: aistudio.google.com/app/apikey

What it finds

| Category | Examples | |----------|---------| | 🔴 Hardcoded Secrets | API keys, passwords, tokens in source code | | 🔴 Code Injection | eval(userInput), shell command injection | | 🔴 SQL Injection | String-concatenated queries | | 🟠 Missing Auth | Unprotected routes touching user data | | 🟠 IDOR | Missing ownership checks on resource endpoints | | 🟠 Path Traversal | User input in file paths | | 🟡 Insecure Config | CORS *, weak JWT algorithms, MD5 hashing | | 🟡 XSS Sinks | innerHTML = userInput |

🏢 Enterprise Air-Gapped Hub & Local LLMs

TitanShield achieves 100% market coverage for off-grid, air-gapped defense and enterprise banking environments using Docker, NGINX, and models like DeepSeek-Coder.

1-Click Infrastructure Installer

IT Admins can rapidly deploy a managed, shared local LLM engine (Ollama + NGINX Proxy) directly to their internal servers:

npx titanshield setup-hub

This instantly generates the docker-compose.yml and nginx.conf securely.

Supported Local Engines

You can securely wire existing pipelines directly to any offline model:

# 1. Connect to any Universal Local Server (vLLM, LM Studio, Llama.cpp)
npx titanshield scan ./src --local-openai \
    --local-url "http://localhost:1234/v1" \
    --local-key "sk_optional_key" \
    --model "deepseek-coder"

# 2. Connect to pure local Ollama
npx titanshield scan ./src --local --model "qwen2.5-coder"

Options

npx titanshield scan <path> [options]

  --key, -k <key>       Gemini API key (or GEMINI_API_KEY env var)
  --output, -o <file>   Save HTML report (e.g. report.html)
  --format <type>       terminal (default) | json | html
  --fail-on <level>     Exit code 1 if findings at level+ (CI/CD)
  --no-ai               Static analysis only (faster, offline)
  --verbose, -v         Show all findings including low/info
  --no-color            Disable colors (clean CI logs)
  --help, -h            Show help

Example Output

  🔴 [CRITICAL] Hardcoded Secret Detected
     src/firebase.ts:12
     A secret key is hardcoded in your source code. Anyone who sees your code can steal it.

     Before: const apiKey = "AIzaSy_abc123_realkey"
     After:  const apiKey = process.env.FIREBASE_API_KEY

     Fix: Move secret to environment variable  (5 minutes)

CI/CD Integration

# GitHub Actions example
- name: Security Scan
  run: npx titanshield scan ./src --fail-on high
  env:
    GEMINI_API_KEY: ${{ secrets.GEMINI_API_KEY }}

Exit codes: 0 = pass, 1 = vulnerability found at fail-on level

Config File (optional)

Create titanshield.config.json in your project root:

{
  "geminiApiKey": "AIzaSy...",
  "failOn": "high",
  "skipDirs": ["vendor", "fixtures"],
  "outputFile": "security-report.html"
}

HTML Report

npx titanshield scan ./src --output report.html

Generates a beautiful, shareable security report. Send it to your CTO, embed in Confluence, or commit as a CI artifact.

Languages Supported

TypeScript · JavaScript (ESM/CJS) · Python · Go · JSX/TSX

Built by TitanShieldAI · Dashboard