Types of XSS

1. Reflected (non-persistent)
   Malicious input is embedded in a URL or request and immediately reflected in the server response (e.g., search results, error message). Usually exploited via a crafted link.

2. Stored (persistent)
   Attacker input is saved server-side (comments, profile fields, message boards) and served to other users later — more dangerous because it can affect many victims.

3. DOM-based
   The vulnerability exists in client-side code: JavaScript reads user-controlled data (URL fragment, query, or window.location) and writes it into the DOM unsafely. No server-side reflection necessary.

Core defenses (high level, prioritized)

1. Output encoding/escaping — Always encode user data for the context where it’s inserted: HTML body, attribute, URL, JavaScript, or CSS contexts. Use established libraries.

   • HTML body: escape <, >, &, " and ' (as required).

   • HTML attribute: additionally encode quotes and handle attribute contexts.

   • JavaScript context: use safe APIs instead of string concatenation.

2. Use safe templating frameworks / avoid raw innerHTML
   Frameworks and templating engines that auto-escape (React, Angular, Twig, Mustache) reduce risk. When you must insert HTML, sanitize it with a well-maintained sanitizer (DOMPurify for browser).

3. Content Security Policy (CSP)
   A strong CSP (e.g., disallow inline scripts, specify script-src) mitigates impact by blocking inline/external scripts. CSP is defense-in-depth, not a replacement for escaping.

4. HTTP-only & Secure cookies
   HttpOnly prevents client JavaScript from reading cookie values; Secure ensures cookies sent only over HTTPS.

5. Input validation (as secondary)
   Validate input shape and length (whitelisting where appropriate). But do not rely on input validation alone for XSS prevention — escaping on output is primary.

6. Avoid dangerous APIs
   Avoid eval, new Function(), innerHTML with untrusted strings, document.write, setTimeout(string), etc.