xibeguard
v1.0.0
Published
AI-powered dynamic penetration testing — auto-starts your app and actually hacks it
Readme
☠️ XibeGuard v3.0
AI-Powered Dynamic Penetration Testing
Detects your stack → Starts your app → Crawls all routes → Fires real exploits → AI analysis
XibeGuard v3 actually runs your application and attacks it. No more "theoretical vulnerabilities" — these are real confirmed exploits with evidence from your live app.
⚡ How It Works
1. 🔎 Detect stack — Node/Python/PHP/Go/Ruby/Rust/Java auto-detected
2. 📦 Install deps — npm install / pip install / composer install / etc.
3. 🚀 Start app — Boots your app on a free local port
4. 🕷️ Crawl routes — Spiders the app + parses source for all endpoints
5. ⚔️ Attack suite — Fires real SQL injection, XSS, auth bypass payloads
6. 🧠 AI analysis — Static code review for what attacks missed
7. 📄 Report — Confirmed exploits + evidence + fixes🚀 Quick Start
npm install && npm run build && npm link
xibeguard config # setup API key + provider
cd /your/project
xibeguard pentest # full attack
xibeguard hack # alias⚔️ Attack Modules
| Attack | What It Does | |--------|-------------| | 💉 SQL Injection | Fires 15+ payloads on every route/param. Checks for DB errors, timing attacks, UNION bypass | | 🕸️ XSS | Reflected + template injection. Tests query params, body, all GET routes | | 🔐 Auth Tests | Default credentials, brute force (rate limit check), auth bypass on protected routes | | 🔄 IDOR | Enumerates IDs on all parameterized routes, checks for unauthorized access | | ⚙️ Security Headers | Checks CSP, X-Frame-Options, HSTS, CORS misconfiguration | | 🌐 Open Redirect | Tests 6 common params with evil.com payloads on all routes | | 🔍 Info Disclosure | Probes .env, .git, stack traces, debug endpoints, backup files | | 🌐 SSRF | AWS metadata (169.254.169.254), internal services, file:// via URL params | | 💣 DoS / Rate Limit | 30-request flood test, 10MB payload test on sensitive endpoints |
🤖 Stack Support
XibeGuard auto-detects and auto-starts:
| Language | Frameworks | |----------|-----------| | Node.js | Express, Fastify, Koa, NestJS, Next.js, generic | | Python | Flask, Django, FastAPI, generic | | PHP | Laravel, generic (php -S) | | Ruby | Rails, Sinatra | | Go | Any (go run .) | | Rust | Actix, Axum (cargo run) | | Java | Spring Boot (mvn/gradle) |
📖 Commands
# Full dynamic pentest (auto-start + crawl + attack + AI)
xibeguard pentest [directory]
xibeguard hack [directory] # alias
# Options
--output ./reports # report output dir
--stream # stream AI analysis live
--skip-static # dynamic attacks only, no AI
--skip-dynamic # AI static analysis only
--no-reality-check # disable brutal commentary🔧 Configuration
xibeguard config # interactive wizard
# Anthropic
xibeguard config --provider anthropic --anthropic-key sk-ant-... --anthropic-model claude-opus-4-6
xibeguard config --anthropic-url https://custom-proxy.com/v1 # support proxies/relays
# OpenAI
xibeguard config --provider openai --openai-key sk-... --openai-model gpt-4o
# Groq (fast + cheap)
xibeguard config --provider openai --openai-key gsk_... \
--openai-model llama3-70b-8192 \
--openai-url https://api.groq.com/openai/v1
# Ollama (free, local)
xibeguard config --provider openai --openai-key ollama \
--openai-model llama3 \
--openai-url http://localhost:11434/v1Env vars also work: ANTHROPIC_API_KEY, OPENAI_API_KEY, OPENAI_BASE_URL
📊 Sample Output
☠️ XibeGuard Dynamic Penetration Test
🔎 Detecting tech stack... → Node.js / Express
📦 Installing dependencies... → npm install complete
🚀 Starting app on port 3001... → App running at http://localhost:3001
🕷️ Crawling routes... → Found 23 routes (47 URLs crawled)
⚔️ Launching attack suite...
💉 Testing SQL Injection...
⚡ CONFIRMED: SQL Injection [POST /api/login]
Evidence: You have an error in your SQL syntax near ''' at line 1
🕸️ Testing XSS...
⚡ CONFIRMED: Reflected XSS [GET /api/search]
Evidence: <script>alert(document.domain)</script> in response
🔐 Testing Authentication...
⚡ CONFIRMED: Default Credentials [POST /api/login]
Evidence: Login succeeded with admin:admin
⚡ CONFIRMED: No Rate Limiting [POST /api/login]
Evidence: 20 requests, no 429 response
🔍 Checking information disclosure...
⚡ CONFIRMED: Sensitive File Exposed [GET /.env]
Evidence: DB_PASSWORD=supersecret123
📊 Pentest Summary
Stack: Node.js / Express
Target URL: http://localhost:3001
Routes: 23 discovered
Attacks: 8 vulnerabilities CONFIRMED ← REAL, LIVE CONFIRMED
██ 3 CRITICAL
4 HIGH🔗 XibeCode + XibeGuard Loop
# 1. Find everything
xibeguard hack ./my-app
# 2. Auto-fix everything
xibecode run "Fix all vulnerabilities in ./xibeguard-reports/pentest-report-*.md"
# 3. Verify fixes
xibeguard hack ./my-app⚠️ Legal
For use only on code you own or have written authorization to test. Unauthorized use is illegal.
XibeGuard v3.0 — Built with TypeScript, Anthropic SDK, OpenAI SDK
