PHP2Shell Attack Attempt vs. Real Breach: What to Look for on Your Server
Author : kanna dhasan | Published On : 20 Nov 2026
Quick answer: An attack attempt is a request that tried to exploit PHP2Shell (CVE-2026-109618 or CVE-2026-109619). A breach is when the attempt worked and the attacker did something afterward. The difference shows up in what happens next on the server: new files, shell commands from the web server, odd outbound connections or new logins. An IDS or WAF alert tells you about the attempt. Only host, network and identity evidence tells you about the breach.
If you run security for a business, you've probably seen the PHP2Shell alerts roll in since October 21, 2026. Your job isn't to react to every one. It's to know which few deserve a human right now.

What is a PHP2Shell attack attempt?
An attempt is any request that tries to use the flaws. Exploit traffic for PHP2Shell has been associated with multipart POST requests that use PHP's session upload-progress feature. Internet-wide scanners fire these at every server they can find, patched or not.
Most attempts fail. The server may already be patched, the PHP setup may not be exposed in the way the attacker expects, or the exploit may only crash a worker. Attempts are a fact of life, like port scans.
What is a real breach?
A breach means the attacker gained control and used it. In this case, that usually means a web shell (a script planted on the server for remote control), followed by exploring files, reading credentials, running commands or moving to other systems.
The keyword is used. A vulnerable server that was probed is not the same as a compromised one, and treating the two the same wastes your team's time, or worse, hides the real incident in the noise.
Why is it so hard to tell them apart?
Because each security tool sees one chapter of the story:
- Vulnerability scanners show which servers have the unlocked door. They can't say whether anyone walked through.
- IDS and network monitoring show someone rattling the handle. An alert alone doesn't prove the exploit worked.
- Endpoint, identity and cloud logs show what happened inside afterward.
An analyst with only the IDS alert might dismiss it. One with only the scan result might schedule the patch for next month. Meanwhile, the endpoint logs a new web shell that nobody links to either. That's the gap behind most missed incidents, and it's the core point of the DESSS guide to PHP2Shell exposure scanning.
Attempt vs. breach: side-by-side
| Attack attempt | Real breach | |
|---|---|---|
| Where it shows up | WAF/IDS alerts, access-log bursts, PHP worker crashes | New files, spawned processes, outbound traffic, new logins |
| Volume | Often high and repetitive | Usually low and quiet |
| Proof of success | None | At least one after-effect on the host |
| Server state | Unchanged | Altered (files, accounts, schedules) |
| Urgency | Patch and monitor | Contain, investigate and rebuild |
| Typical owner | Ops or security monitoring | Incident response |
The evidence ladder: how far did it get?
Think of an attack as a ladder. Each rung up means more urgency.
- Rung 1, probing. Alerts or WAF blocks for the PHP2Shell pattern. Action: confirm the server is patched.
- Rung 2, impact on the service. PHP worker crashes or bursts of 500/502 errors that line up with the alerts. Action: patch immediately and keep watching.
- Rung 3, foothold. An unexplained new PHP file in a web folder, or the web server launching a shell. Action: isolate and investigate.
- Rung 4, activity. Outbound connections to unfamiliar addresses, new logins or new scheduled tasks. Action: full incident response.
The moment you find evidence at rung 3 or 4, you stop treating it as a vulnerability-management ticket and start treating it as an incident.
How should a security team triage PHP2Shell alerts?
A simple rule keeps analysts out of trouble:
- Check patch status first. If the target is on a fixed release (PHP 8.5.12, 8.4.27, 8.3.36 or 8.2.35, or a vendor build with the backport), an alert is informational.
- If unpatched, look for after-effects on that host. Files, processes, network, identity. Don't stop at the alert.
- Escalate on any unexplained after-effect. One unexplained trace outweighs a hundred routine alerts.
- Close the loop on every escalation. Write down what you checked and what you found, even when the answer is "nothing."
Fast response matters because attackers moved first. F5 Labs reported scanning that started more than five hours before the CVEs and patches were published, and thousands of attempts within hours of disclosure.
What if your tools don't connect?
Then the process described above becomes manual: exporting from one console, pasting into another, discussing over email. That's too slow for an exploit that was being scanned for before the patch existed.
What works better is a shared view where scanner findings, IDS alerts, network flows and endpoint activity for each server sit on one timeline. The goal isn't more alerts. It's fewer, more complete ones.
What to do after you've decided
- If it was an attempt only: patch using your platform guide (Linux or Windows Server with IIS and Apache), test your database connections using these notes on PHP upgrades and database connectivity, and keep the alert rules on.
- If it was a breach: isolate the server, preserve disk and log evidence, rotate every credential the server could read, rebuild from a trusted image on a patched PHP, and review neighbouring systems for lateral movement.
Frequently asked questions
What is the difference between an exploit attempt and a breach? An attempt is a request that tried to exploit the flaw. A breach is when it worked and the attacker acted on the server, such as planting files or running commands.
Does an IDS alert prove my server was compromised? No. An alert shows an attempt. You need host, network or identity evidence to show it succeeded.
Do I need to investigate every PHP2Shell alert? Not at equal depth. Check patch status first. Investigate deeply only on unpatched, exposed hosts, and escalate anything with unexplained after-effects.
What is the strongest sign of a real breach? Unexplained new PHP files in web folders combined with the web server starting shell commands or making unusual outbound connections.
Can a patched server still be compromised? Yes, if it was breached before patching. Patching doesn't remove a web shell that's already there.
How long should I keep monitoring? Keep PHP2Shell detection rules on indefinitely. Scanning for known PHP flaws continues long after the patch release.
Is it worth getting outside help? If you lack process or network logging, or you've found a suspicious trace you can't explain, yes. An external review is cheaper than a missed breach.
The bottom line
Alerts tell you someone knocked. Evidence on the server tells you whether anyone came in. Keep the two separate, escalate on after-effects and fix the visibility gap. If you want help doing that, DESSS runs PHP2Shell vulnerability scans and compromise checks for businesses across the US.
