IDOR Plus XSS: The Bug Chain That Beats Every Single Fix

# IDOR Plus XSS: The Bug Chain That Beats Every Single Fix

Vulnerable vs fixed JavaScript code side by side
Vulnerable vs fixed JavaScript code side by side

Somewhere right now a developer is fixing an IDOR by adding an ownership check. Somewhere else another developer is upgrading their HTML sanitizer. Both feel productive. Neither touches the actual problem, because on the platforms I've been testing this year, the bugs that pay aren't single vulnerabilities — they're two mediocre issues holding hands.

This is the pattern I want to burn into your hunting brain, because it changed my hit rate more than any tool or wordlist did.

The Anatomy of a Boring-But-Lethal Chain

Recent public writeups keep landing on the same skeleton, and once you see it you can't unsee it.

Take one real example from this year's Intigriti challenge. Sequential session IDs — set your cookie to 1, and you're admin. Textbook IDOR, trivially exploitable, but on its own worth maybe a low severity because the platform had no cross-user data rendering. Then a separate quirk: the app read its analytics script URL from a window global. And a third: the sanitizer version in use didn't strip `id` attributes — that protection shipped in a later release.

None of those three is impressive alone. Chained: the IDOR let the attacker plant content in the admin's session view, the surviving `id` attribute clobbered the window global, the tracker loaded whatever script URL the clobbered value pointed at, and the sanitizer waved it all through. Full XSS executed against an admin session, from three bugs that would each have been closed as informational on their own.

Individual bugs get severity scores. Chains get payouts. Triage teams know the difference, and so should you before you spend three weeks hunting the perfect standalone RCE.

Another writeup from earlier this year chained four defects into one-click account takeover: a DOM sink with no scheme validation, a WAF rule with a regex gap, the browser's cross-origin `window.name` persistence, and a first-party SDK that handed signed credentials to whatever JavaScript asked nicely. The WAF never saw the payload because it traveled in tab state, not the URL. Four individually-mediocre issues. Combined impact: victim's JWT and live AWS credentials.

Why Scanners Miss Every One of These

Automated tools test vulnerabilities in isolation by design. An IDOR scanner checks whether object access crosses ownership boundaries. An XSS scanner checks whether input reflects unsanitized. Neither asks the compositional question: can bug A put my payload where bug B renders it for a high-value user?

That's not a tooling failure exactly — composition space explodes combinatorially and nobody has cracked practical automated chaining. Which means the human hunter who thinks in chains is competing mostly against other humans who don't. That gap is your edge, and unlike tooling advantages, it doesn't commoditize.

The mental shift is small but total. Stop asking "is this endpoint vulnerable." Start asking "what could I do if this endpoint behaved slightly worse than it does." Every finding becomes raw material instead of a verdict.

Three bug symbols forged into one chain
Three bug symbols forged into one chain

Where to Point This Lens First

After enough of these, patterns emerge about which combinations tend to exist together on real platforms.

Anywhere user-generated content gets rendered back to privileged users — support ticket systems, moderation queues, admin dashboards reviewing flags — a stored payload meeting a rendering flaw is the highest-value chain in web hunting, and both halves usually look harmless in isolation. Anywhere state-changing actions lack CSRF tokens, check whether a same-origin script execution primitive exists elsewhere; the combination means full account control without ever touching a password. And anywhere an authentication SDK exposes credential methods to page JavaScript — increasingly common as apps move auth client-side — every XSS anywhere in scope escalates to credential theft, which quietly changes how you should score every reflected-input oddity you find.

I'll admit something about my own process here: I ignored chaining advice for months because it felt like an advanced technique reserved for people with bigger rep bars than mine. That was backwards. Chaining is how hunters without exotic skills beat the ones who only chase headline bug classes. My first accepted reports were all single low-severity findings. The first report that paid real money was two lows holding hands.

Your Next Session Looks Different Now

Practical change for your next hunt: when you find a medium-severity issue, before writing the report, spend thirty minutes asking what else on the platform combines with it. Check what renders user content, who sees it, and whether any global JavaScript objects are readable from injected markup. Check whether your medium sits upstream or downstream of any trust boundary.

Keep a private list of "unresolved quirks" per program — the weird behaviors that weren't exploitable alone. When you find the missing half months later, that stale note becomes the payout. Several hunters I respect treat these lists as their primary asset.

And fix-side, if you build things: the lesson isn't "fix everything" — resources are finite. It's that fixes need compositional review. Before closing a ticket as low-severity, ask what this behavior enables in combination with anything else you ship. The chains that cost companies six figures almost always had every component documented somewhere internally already.

For a worked example of chaining logic applied end-to-end, my Noones IDOR writeup shows the discovery-and-escalation thought process on a real target, including the parts that didn't work. And if you're newer to this, the audit checklist piece covers where to test first before chain-hunting makes sense.

🔎 Found this useful? Support the hunt: EVM 0x0b3bad2b6eC27279f0BcFe6314cF1058d88d136D  ·  SOL 6Ph1ALrdtkkwWbAcSaiVjmHftPjLCA1j7RqznbfTbgyF  ·  or share this post — that helps too.

Disclosure: some links here are affiliate links at no extra cost to you. I only recommend gear and services I actually use while hunting.

Admin dashboard with session data highlighted
Admin dashboard with session data highlighted

Comments

Popular Posts