In short
ezXSS is an open-source PHP tool for testing blind cross-site scripting (XSS). The tester plants a payload that loads a script from their own ezXSS server. When the payload fires in a victim's browser, ezXSS records the page URL, cookies, storage, DOM and a screenshot, then sends an alert.
Key takeaways
- What it is
- An MIT-licensed PHP application for penetration testers and bug bounty hunters that collects the results of blind XSS payloads.
- How it works
- A planted payload loads a collector script from the ezXSS server. When it runs, the script sends page data back and ezXSS sends an alert by email, Telegram, Slack or Discord.
- Why it matters
- Blind XSS often fires in admin panels and support tools. A hostile operator can collect non-HttpOnly cookies and page content from staff browsers.
- What Odin Research saw
- 127 web servers labeled ezXSS in Monograph 13 were ordinary PHP 8.5 sites. Only two servers showed real ezXSS panel markers.
- What to do first
- Do not block on a PHP version header. Look for the ezXSS login path, page title or collector script before acting.
What is ezXSS?
ezXSS is an open-source PHP application that collects the results of blind cross-site scripting (XSS) payloads. A tester plants a payload that loads a script from their ezXSS server. When it runs in someone's browser, ezXSS records what that page showed and sends an alert. Testers and attackers both use it.
The project lives on GitHub under the MIT license. Its README describes it as "an easy way for penetration testers and bug bounty hunters to test (blind) Cross Site Scripting."
Cross-site scripting lets an attacker run script in another user's browser through a vulnerable site. Blind XSS is the case where the attacker never sees the page where the payload runs. A support ticket, contact form or log viewer may only render it later, often for staff. A collector such as ezXSS is how the tester finds out it worked.
How ezXSS works
- Set up the server. The operator installs ezXSS on Apache, NGINX or Docker. The README asks for PHP 7.1 or later, a domain name, and a TLS certificate for testing HTTPS sites.
- Plant the payload. The tester puts a short script tag that points to the ezXSS server into fields the target application stores, such as names, comments or headers.
- Wait for a trigger. When a staff member or another user opens the stored data, their browser loads the collector script.
- Collect and alert. The script sends data back to ezXSS. The README lists the page URL, IP address, referer, user agent, non-HttpOnly cookies, local and session storage, the full HTML DOM, origin, timing and a screenshot. It can also fetch extra paths the operator configures.
The operator then gets an alert by email, Telegram, Slack, Discord or a custom callback URL, and reviews the report in a web panel.
ezXSS belongs to a wider group of out-of-band testing tools. Others, such as Interactsh and Burp Collaborator, catch out-of-band network callbacks rather than browser data. See our OAST page for the bigger picture.
ezXSS requirements and popularity
ezXSS is a small, self-hosted web application. It needs only PHP, a web server, a domain and a certificate, so it fits on a cheap VPS. The README points to the project wiki for install steps on Apache, NGINX and Docker.
The repository had about 2,400 stars and 382 forks on GitHub on 21 November 2026. That makes it one of the better-known blind XSS collectors.
Is ezXSS legal? Yes, for testing applications you own or have permission to test, including bug bounty programs within their rules. Planting payloads in other sites is illegal in most countries.
How attackers use ezXSS
The same features that help a bug bounty hunter help an attacker. Blind XSS payloads tend to fire in the most privileged places: help desk consoles, admin dashboards, CRM views and log viewers.
- Session theft. Cookies without the HttpOnly flag can be read by script. A stolen session cookie can let an attacker act as the staff member, which is a form of session hijacking.
- Internal reconnaissance. The full DOM and screenshot show what an internal tool looks like, which data it holds and which links it exposes.
- Follow-up requests. Extra paths fetched from the victim's browser can reach pages only staff can see.
Because ezXSS only needs a stored input field somewhere, the entry point is often a public form. The impact lands inside the organization.
What Odin Research observed
- A label that mostly missed. In Monograph 13, a weak ezXSS rule in an upstream feed fired on 127 endpoints across 101 hosts. They were commercial websites, event agencies, regional portals and newsletter servers on ports 80, 443 and 8080.
- Runtime version drift. Those sites had upgraded to PHP 8.5.10. They then sent the same header mix the report associates with Docker ezXSS installs: a PHP version header, a
PHPSESSIDcookie and a strict Content-Security-Policy. A header-only rule labeled each of them. - A 71.8% false-positive rate. Of 177 raw actionable ezXSS rows in that feed, 127 were collisions. Only 2 ezXSS servers held up after checks.
- What the real servers showed. The two confirmed servers exposed the operator login at
/manage/login, a page title starting with ezXSS and the collector script/assets/js/ezxss.js, which harvestsdocument.cookieandlocalStorage. - Next to callback servers. Monograph 17 found one server in France that ran a five-port Interactsh listener and an ezXSS blind XSS payload manager side by side. It was one of eight corroborated co-locations of callback listeners with offensive tools.
How to detect ezXSS
- Look for the panel, not the PHP version. Monograph 13 publishes Suricata SID 9513005 for the collector script with
document.cookieandlocalStorage, and SID 9513006 for the ezXSS title with/manage/login. Both are safe against PHP 8.5 header drift. - Clean your feeds. The same report's Sentinel query drops low-confidence ezXSS indicators that carry only a PHP 8.5 header. Its YARA file adds a rule for ezXSS content.
- Watch your own pages. Content-Security-Policy violation reports that show scripts from unknown domains in admin tools are an early sign of a blind XSS test or attack.
- Hunt stored inputs. Search ticket, comment and log stores for script tags that point to external hosts.
All rules are on the report pages and in the Defense Kit.
How to defend against blind XSS
- Encode output by context. The OWASP XSS Prevention Cheat Sheet puts context-aware output encoding and safe DOM APIs first.
- Use your framework's escaping. Avoid escape hatches that insert raw HTML, and sanitize any user-authored HTML with a maintained library.
- Add a Content-Security-Policy to internal tools as defense in depth. OWASP is clear that CSP is not the primary control.
- Harden cookies. Set HttpOnly on session cookies so a collector cannot read them, and keep sessions short in admin tools.
- Do not rely on a WAF alone. OWASP notes a WAF cannot reliably stop all XSS.
Then test your own stored inputs with an authorized collector, as part of regular penetration testing.
ezXSS compared with other callback tools
| Tool | What it catches | Odin sightings | Real-server marker |
|---|---|---|---|
| ezXSS | Browser data from fired XSS payloads | 2 confirmed of 252 feed rows (Monograph 13) | /manage/login and ezxss.js |
| Interactsh | Out-of-band callbacks over DNS, HTTP or mail | 638 endpoints (Monograph 17) | Reflects test IDs |
| Burp Collaborator | Out-of-band callbacks | 109 private endpoints (Monograph 17) | Private collaborator listeners |
| DNSLog | DNS lookups | 8 listeners (Monograph 17) | Often co-hosted with other tools |
Frequently asked questions about ezXSS
What is ezXSS?
ezXSS is an open-source PHP application for testing blind cross-site scripting. It hosts a collector script and records page data, cookies, storage and a screenshot when a planted payload runs in a browser.
Is ezXSS malware?
No. It is an MIT-licensed testing tool for penetration testers and bug bounty hunters. Attackers can run it too, so an ezXSS server is dual-use.
Is ezXSS legal?
Yes, when you test applications you own or are allowed to test. Planting payloads on other sites without permission is illegal in most countries.
What is blind XSS?
Blind XSS is cross-site scripting where the payload runs in a page the attacker cannot see, such as an admin panel or support console. A collector like ezXSS reports when it fires.
What data does ezXSS collect?
According to its README: the page URL, IP address, referer, user agent, non-HttpOnly cookies, local and session storage, the HTML DOM, origin, timing and a screenshot.
How do I detect an ezXSS server?
Look for the /manage/login operator page, a title starting with ezXSS and the /assets/js/ezxss.js collector script. Odin Research publishes Suricata rules for these markers in Monograph 13.
Why do threat feeds flag normal websites as ezXSS?
Odin Research found 127 PHP 8.5.10 websites labeled ezXSS because a feed rule matched only on headers. A legitimate PHP upgrade moved them into the rule's range.
ezXSS vs Interactsh: what is the difference?
ezXSS collects browser data when an XSS payload runs. Interactsh catches out-of-band network callbacks, for example over DNS or HTTP. Odin Research has seen both on the same server.
Sources
- ssl, ezXSS repository, GitHub. github.com/ssl/ezXSS (checked 21 November 2026)
- OWASP Cheat Sheet Series, Cross Site Scripting Prevention Cheat Sheet. cheatsheetseries.owasp.org/cheatsheets/Cross_Site_Scripting_Prevention_Cheat_Sheet.html (checked 21 November 2026)
Related tools and terms
What is Interactsh?
Interactsh is ProjectDiscovery's open-source OAST server and client for catching DNS, HTTP and other callbacks.
OAST serviceWhat is Burp Collaborator?
Burp Collaborator is PortSwigger's callback service for Burp Suite, used to detect out-of-band vulnerabilities.
OAST serviceWhat is DNSLog?
DNSLog is a family of open-source DNS callback collectors used to confirm blind injection and SSRF flaws.
Interception proxyWhat is mitmproxy?
mitmproxy is an open-source interactive HTTPS proxy. It installs its own root certificate to read encrypted traffic.
Security testing suiteWhat is Yakit?
Yakit is a security testing suite with a desktop interface and a remote engine that can intercept traffic, fuzz web parameters and run exploit checks.
Reconnaissance platformWhat is ScopeSentry?
ScopeSentry is an open-source asset discovery and scanning platform that maps subdomains, ports and web services.
Odin Research
Get the rules, not just the definition
Every Odin Research report ships with verified indicators and detection rules you can load today.