Skip to content

Defender's reference · MITRE T1505.003

Find the web shell before it finds your data.

A practical reference for blue teams and incident responders: how web shells get planted, where they leave traces, how to investigate them, and how to eradicate and harden for good.

Areas
6
In-depth guides
9
Glossary terms
12
case-0142 · correlated timelinehigh
  1. 10:42:07FIMw3wp.exe created /aspnet_client/.../errorpage.aspx
  2. 10:42:49LOGPOST /aspnet_client/.../errorpage.aspx · 1 client · no referer
  3. 10:42:50PROCw3wp.exe → cmd.exe /c whoami /all
  4. 10:45:13PROCw3wp.exe → cmd.exe → nltest /dclist:
w3wp.execmd.exewhoami.exenltest.exe
illustrative · T1505.003 → T1033 → T1482

Defence map

From first write to lessons learned

Six areas that follow a web shell intrusion end to end. Each one links to its hub of guides.

Where shells leave traces

Four layers, one intrusion

A web shell can hide from one control, rarely from all four. Layered visibility is the whole game.

  1. L1

    Files

    New executable files in web-served paths, changed configs, odd owners and timestomped timestamps.

    FIM · Sysmon 11 · auditd

  2. L2

    Web logs

    Rare URIs, POST-only pages, a handful of clients, empty referers and wildly varying response sizes.

    W3C · combined · KQL

  3. L3

    Process trees

    w3wp.exe, httpd, php-fpm or java spawning shells and reconnaissance tools like whoami and nltest.

    Sysmon 1 · 4688 · EDR

  4. L4

    Memory

    In-memory implants, injected assemblies, filters and modules that never touch the disk.

    memory image · modules

Guides

Latest guides

All guides
Indicators & detection

YARA and Sigma rules for web shell detection

Write, test and tune YARA rules for suspicious server-side scripts and Sigma rules for web shell behaviour in logs and process telemetry — with annotated examples and a false-positive strategy.

Defensive by design

Everything here is written for defenders: concepts, detections, forensics and hardening. We never publish web shell source code, deployable payloads or exploit code.

Browse the glossary