Vulnerabilities & Fixes
Deep dives on the vulnerabilities that actually hit AI-built apps · exposed keys and secrets, XSS, CSRF, injection, CORS, missing headers · each with a concrete fix.
33 articles

Email Security Audit for DMARC Compliance: Sending Sources, Reports and Enforcement
For a DMARC compliance audit, begin by identifying every system and service that sends email for your domain. Aggregate reports reveal which senders use your domain and where alignment breaks down. Monitoring before enforcement reduces the risk of blocking legitimate email.
Read article
How to Check and Configure Security Headers in Next.js Apps
Security headers send rules with each HTTP response. Browsers enforce these rules to help protect the app. CSP controls allowed content, while HSTS enforces HTTPS. Fresh Next.js starter apps lack most app-specific security headers.
Read article
How secure are Replit environment variables and Secrets?
Replit Secrets encrypt stored values with AES-256 and protect them in transit with TLS. However, organization members without the Owner role can still read values by printing environment variables. Development and production secrets use separate stores, so editor changes do not automatically update published apps.
Read article
How do you set up website uptime monitoring with Slack alerts?
Choose a check interval based on service priority, following Uptime Monitor’s published guidance. Its recommended threshold is two or three consecutive failures for most services. For Slack alerts with blocks, include the recommended text fallback for notifications.
Read article
How should you read and act on a website quickscan?
OWASP says a test report is a guide, not a promise to cover all risks. For a website quickscan, my suggested method is to log the target, scope, and time, then review each flag. Keep what you saw apart from possible harm, assign a fix owner, and plan a retest.
Read article
How does an exposed API key detector find secrets?
Gitleaks is a secret detector that searches files and Git repos. Its rules match secret text and can check that text's Shannon entropy. For secrets it can classify, TruffleHog can log in to check validity.
Read article
How can you get a website security badge for free?
Sectigo recommends its free dynamic TrustLogo for sites with purchased SSL certificates. Its static seals lack Point to Verify functionality. Let's Encrypt staff reject seals that imply a security guarantee. Choose based on the required certificate and verification, not the graphic alone.
Read article
How do you validate request data with Joi schemas?
Define schemas for incoming request bodies, headers, and query parameters before routing payloads to downstream application logic. Compare observed validator output against agreed endpoint specifications, reject unrecognised keys across your service boundary, and map sanitized values into internal handler workflows.
Read article
How do you audit SPF, DKIM and DMARC records with an email security audit tool?
Audit your primary domain and sending hostnames using an automated email security audit tool to confirm existing records. Compare observed SPF syntax limits, published DKIM public key selectors, and designated DMARC reporting mailboxes against agreed operational delivery policies.
Read article
How to Check Your Website for Env-Var Leaks and Exposed Files
Review your build bundles, public asset directories, and exposed web paths for secret tokens and sensitive files. Record all observed responses, matched key identifiers, and client script references. Choose appropriate key revocation steps or routing adjustments whenever unexpected files or private environment variables appear in client accessible outputs.
Read article
What should a website domain health check cover?
Choose a target hostname and confirm verification boundaries across your public assets. Review root DNS entries, transport certificates, routing headers, and mail authenticity records. Record each observed state against agreed internal standards. Decide on remediation steps whenever an observed value diverges from your team baseline.
Read article
OpenAI API Security: Reducing Prompt-Injection Risk in Vibe-Coded Applications
Reduce prompt-injection risk by separating trusted instructions from untrusted data, restricting tool permissions, and validating sensitive actions outside the model. Protect secret API keys on the server and enforce usage limits. Sanitization, moderation, and scans do not prove an integration is secure.
Read article
How to Remove Secrets from Environment Variables and Git History
To remove secrets from environment variables, you must rotate compromised credentials immediately and use tools like git filter-repo or BFG Repo-Cleaner to scrub them from Git history. Simply deleting the file is insufficient. Migrate to a dedicated secret manager to prevent plain-text exposure in .env files or system process lists.
Read article
Hapi.js Injection Prevention: Secure Your Backend Against Attacks
To prevent injection in Hapi.js, you must use Joi for strict input validation and employ parameterized queries for database interactions. By enforcing schemas on payloads and query parameters, Hapi.js blocks malicious data before it reaches your logic, effectively mitigating SQL, NoSQL, and command injection risks in Node.js backends.
Read article
Is It Safe to Expose Your Supabase Anon Key? Yes | With One Condition
Yes, exposing your Supabase anon key is safe because it is designed to be public. However, this is only true if Row Level Security (RLS) is enabled on every table. Without RLS, the anon key allows anyone to read your entire database. Never expose the service_role key.
Read article
Broken Access Control Checklist | Find and Fix OWASP A01
Broken access control occurs when an application permits data access or actions outside its intended rules. Test protected operations on the server or database with the relevant users and roles. Ownership checks, role checks, and database policies should reflect the application’s actual sharing model.
Read article
security.txt: Give Researchers a Simple Way to Report Bugs (RFC 9116)
security.txt is a plain-text file at /.well-known/security.txt, defined by RFC 9116, that provides a standardized way for researchers to report vulnerabilities. It requires two fields: Contact and Expires. For AI-built apps, it is a vital safety net that ensures bugs are reported privately rather than disclosed publicly.
Read article
DNSSEC and CAA: Protect DNS Integrity and Limit Certificate Issuance
DNSSEC lets validating resolvers authenticate signed DNS data through a chain of trust. CAA records specify which certificate authorities may issue certificates. Configure both with your DNS host and registrar, then verify the results; neither protects a compromised registrar account or guarantees application security.
Read article
DNS Records Explained: A, CNAME, MX, TXT, NS and CAA for App Builders
DNS records tell the internet where your app loads, where your email lands, and who may issue certificates for your domain. A and AAAA map names to addresses, CNAME aliases them, MX routes mail, TXT carries SPF, DKIM and DMARC, NS delegates control, and CAA restricts certificate issuance · audit all six regularly.
Read article
Content Security Policy for Vibe-Coded Apps: A Practical CSP Guide
Content Security Policy (CSP) is a browser-enforced allowlist that blocks unauthorized scripts, providing the strongest defense against XSS. Many vibe-coded apps ship with no CSP or use 'unsafe-inline', which negates protection. This guide explains how to implement strict policies using nonces, hashes, and report-only mode to secure AI-built applications.
Read article
AI API Security | Protecting LLM-Powered Apps, Keys, and Endpoints
AI API security involves protecting LLM keys (OpenAI, Anthropic) and hardening the endpoints you build. Keep provider secrets server-side, authenticate and authorize proxy requests, and apply rate limits. Treat model output as untrusted and test input and output handling for the relevant attacks.
Read article
Webhook Signature Verification: Stop Trusting Unsigned Payloads
Verify every webhook by computing the HMAC-SHA256 of the raw request body using your shared secret and comparing it to the provider's signature header with a timing-safe function. Add a five-minute timestamp tolerance and event-ID deduplication to prevent replay attacks. Without this, anyone can forge events to your endpoint.
Read article
JWT Security: How to Read a Token and Catch the Red Flags
Signed JWTs (JWS) have headers and payloads that are readable without a key. JWTs can also be encrypted (JWE). This guide covers signed tokens: decoding reveals their claims, while secure use requires server-side validation, protected keys, and authorization checks.
Read article
How to Find Exposed Secrets in Your Code Before They Ship
Exposed secrets hide in three predictable places: frontend bundles, git history, and misprefixed .env files. Grep your code and built output for known prefixes like sk_live_, AKIA, and ghp_, then use an automated secret scanner to catch high-entropy leaks. If you find one, rotate the key immediately to end the exposure.
Read article
CORS Misconfigurations That Leak User Data (and How to Test For Them)
A CORS misconfiguration that reflects any request origin while allowing credentials lets malicious sites read your users' private API data. AI generators often create these permissive defaults to silence errors. To stay secure, use a hard-coded origin allow-list and test your API server-side with crafted Origin headers.
Read article
CVE-2025-48757 Explained: How to Check If Your Lovable App Is Affected
CVE-2025-48757 is a disputed record concerning insufficient database authorization in Lovable-generated sites. Frontend login screens cannot enforce database permissions. Audit your app’s table privileges, Row-Level Security policies, and server-side authorization, then test anonymous access and access between separate test accounts.
Read article
OWASP Top 10 for AI-Built Apps: What Vibe Coders Need to Know
Use the OWASP Top 10 to organize a review of your app's access controls, input handling, dependencies, configuration, and failure behavior. Generated code needs these checks too. The appropriate controls depend on the architecture; a scan or one database setting does not cover every category.
Read article
Environment Variables Security: Stop Leaking Secrets to Production
Environment variables configure an application; they do not encrypt secrets or guarantee confidentiality. Public prefixes such as VITE_, NEXT_PUBLIC_, and REACT_APP_ can expose values to browser code. Keep privileged credentials server-side, review the actual build configuration, and inspect bundles, responses, logs, and repository history for leaks.
Read article
XSS Prevention Guide: Protect Your AI-Built App from Cross-Site Scripting
Prevent XSS in AI-built apps by avoiding dangerous sinks like dangerouslySetInnerHTML and innerHTML. Use DOMPurify for sanitization, prefer textContent for plain text, and implement a strict Content-Security-Policy (CSP). Review generated code and deployment configuration independently of whether the visible features work.
Read article
CSRF Protection & Security Headers: The Missing Layer in AI-Built Apps
Review CSRF defenses separately from browser security headers. For cookie-authenticated actions, test the application’s token or origin checks and cookie settings. Response policies such as CSP, HSTS, and framing restrictions provide additional browser protections, but their presence does not establish that every attack is blocked.
Read article
How to Fix Exposed API Keys
If a secret API key leaks, revoke it through the provider, investigate its use, and repair the exposure path. Store replacements server-side and test authorized integrations. Deleting a file does not invalidate copied credentials or undo prior access.
Read article
Code Injection Prevention: SQL Injection, eval(), and Command Injection in AI Apps
Reduce injection risks by keeping untrusted values separate from query and command structure. Use parameterized database APIs, validate expected input types, and review any interpreter that accepts user-controlled expressions. Dedicated parsers and direct process execution still need limits appropriate to their use.
Read article
AI Security Risks: Prompt Injection, LLM Abuse, and API Key Exposure
Review prompt injection, abuse of AI endpoints, and exposure of provider secrets. Keep secret keys server-side, authorize requests, and bound resource use. Treat model instructions and output filters as fallible controls, with tool permissions enforced by application code.
Read article