Security Scanner
YourAI-generatedcode
hasvulnerabilities.
ShipSafe scans your GitHub repos for exposed API keys, SQL injection, insecure dependencies, and 50+ other patterns. Results in under 2 minutes.
The Problem
AIwritescodefast.
Itdoesn'twriteitsafe.
LLMs are trained on what code should do, not what it shouldn't. They'll hardcode your API key, skip authentication, and trust every input — then tell you the code is production-ready.
of AI-generated codebases expose API keys or secrets in their source code.
have API routes with no authentication — anyone with the URL can access them.
contain SQL injection or command injection vulnerabilities in user-facing code.
Based on scanning hundreds of AI-generated repositories. Read the full research →
Threat Detection
What'shidinginyour
vibe-codedapp?
Exposed API Keys
Your OpenAI key is in the frontend bundle. Someone just ran up a $15K bill on your account.
Hardcoded Secrets
Database passwords, JWT secrets, and tokens committed to source code. Bots scrape GitHub for these.
SQL Injection
Your queries trust user input. One crafted string downloads your entire database.
Command Injection
User input passed to exec() or eval(). Attackers can run any command on your server.
No Authentication
Your admin panel has no login. Anyone with the URL can delete all your data.
Missing CSRF Protection
No CSRF tokens on forms. A malicious link can make actions on behalf of your logged-in users.
Insecure CORS
Access-Control-Allow-Origin set to *. Any website can make authenticated requests to your API.
Known CVEs
Your dependencies have published exploits. Attackers can run code on your servers.
Outdated Dependencies
Packages months or years behind. Known vulnerabilities with public proof-of-concept exploits.
Weak Cryptography
Using Math.random() for tokens or MD5 for passwords. Both are trivially crackable.
Information Leakage
Console.log(token), verbose error messages, or stack traces exposed to users.
Insecure HTTP
API calls over plain HTTP. Anyone on the network can read passwords and tokens in transit.
How It Works
Threestepsto
shipwithconfidence
Connect GitHub
One-click OAuth. Read-only access. We never store your source code.
Automatic Scan
50+ patterns analyzed in under 2 minutes. No config needed.
Fix in Plain English
Every issue explained. Copy-paste fixes included. No security expertise required.
Real Findings
Whatascanactually
lookslike
These are real patterns we find in AI-generated code every day. Each one comes with a plain-English explanation and a copy-paste fix.
Exposed OpenAI API Key
const openai = new OpenAI({ apiKey: "sk-proj-xK9m..." })Move to .env.local and use process.env.OPENAI_API_KEY. Rotate the exposed key immediately.
SQL Injection in User Query
const result = await db.query(`SELECT * FROM users WHERE email = '${email}'`)Use parameterized queries: db.query("SELECT * FROM users WHERE email = $1", [email])
Missing Authentication on Admin Route
export async function DELETE(req) { await db.delete(users)... }Add auth middleware to verify the user is authenticated and has admin role before processing.
Security Playbook
24ruleseveryAIapp
shouldfollow
We compiled the most common security mistakes in AI-generated code into a free, actionable checklist. ShipSafe checks all 24 automatically.
Auth & Sessions
Secure API Dev
API & Access
Data & Infra
Other Rules
Built For
IfyoucodewithAI,
thisisforyou
Vibe Coders
You built your app with ChatGPT, Cursor, or Copilot. It works great. But you have no idea if it's secure.
Indie Hackers
You're shipping fast and alone. You don't have time to learn security, but you can't afford a breach either.
Non-Technical Founders
Your developer used AI to build your MVP. You need to know if your users' data is actually safe.
Freelancers
You're delivering client projects built with AI tools. A ShipSafe badge shows you take quality seriously.
Bootcamp Grads
You learned to code recently and use AI heavily. Security wasn't in the curriculum, but it matters.
Small Teams
Your 2-3 person team ships fast with AI assistance. No one has security expertise. ShipSafe is your safety net.
Pricing
Simple,transparentpricing
Start free. Upgrade when you need full details.
Free
- ✓ Unlimited scans & repos
- ✓ Safety Score (A–F)
- ✓ Finding counts & categories
- ✓ First scan with full details
- ✕ File paths & line numbers
- ✕ AI explanations & fixes
Pro
Popular- ✓ Everything in Free
- ✓ Full finding details
- ✓ AI explanations
- ✓ Copy-paste fixes
- ✓ 10 repositories
- ✓ README badge
Team
- ✓ Everything in Pro
- ✓ Unlimited repos
- ✓ Scan on every push
- ✓ Slack & email alerts
- ✓ CI/CD integration
- ✓ 5 team seats
FAQ
Commonquestions
Do you store my source code?
No. We stream your files through the scanner and discard them immediately. Nothing is saved to disk or database. Only the scan results (findings, scores, file paths) are stored.
What GitHub permissions do you need?
Read-only access to your repository contents. We use GitHub OAuth with the minimum required scope. We cannot push code, create issues, or modify your repos in any way.
How is the Safety Score calculated?
You start at 100. Each finding deducts points based on severity: critical (-25), high (-15), medium (-8), low (-3). The score maps to a letter grade: A (90-100), B (75-89), C (60-74), F (below 60).
What does the free plan include?
Unlimited scans with Safety Scores and finding counts. Your very first scan includes full details (file paths, AI explanations, copy-paste fixes). After that, full details require Pro or a $9 single-scan unlock.
Can I scan private repositories?
Yes. ShipSafe works with both public and private repos. Your code is accessed via your GitHub OAuth token with read-only permissions.
How long does a scan take?
Most scans complete in under 2 minutes. Larger repositories with many files may take slightly longer. You can watch the progress in real time.
Is ShipSafe itself secure?
We scan our own codebase with ShipSafe. Check our live security badge on this page. We also detail our data practices on our Terms & Data page.
Stopshipping
insecurecode
Scan your repo now. Get results in 2 minutes. Your first scan includes full details — free.