Pentesting

External Recon Methodology

Assets discoveries

So you were said that everything belonging to some company is inside the scope, and you want to figure out what this company actually owns.

The goal of this phase is to obtain all the companies owned by the main company and then all the assets of these companies. To do so, we are going to:

  1. Find the acquisitions of the main company, this will give us the companies inside the scope.

  2. Find the ASN (if any) of each company, this will give us the IP ranges owned by each company

  3. Use reverse whois lookups to search for other entries (organisation names, domains...) related to the first one (this can be done recursively)

  4. Use other techniques like shodan organd sslfilters to search for other assets (the ssl trick can be done recursively).

Acquisitions

First of all, we need to know which other companies are owned by the main company. One option is to visit https://www.crunchbase.com/, search for the main company, and click on "acquisitions". There you will see other companies acquired by the main one. Other option is to visit the Wikipedia page of the main company and search for acquisitions.

Ok, at this point you should know all the companies inside the scope. Lets figure out how to find their assets.

ASNs

An autonomous system number (ASN) is a unique number assigned to an autonomous system (AS) by the Internet Assigned Numbers Authority (IANA). An AS consists of blocks of IP addresses which have a distinctly defined policy for accessing external networks and are administered by a single organisation but may be made up of several operators.

It's interesting to find if the company have assigned any ASN to find its IP ranges. It will be interested to perform a vulnerability test against all the hosts inside the scope and look for domains inside these IPs. You can search by company name, by IP or by domain in https://bgp.he.net/. Depending on the region of the company this links could be useful to gather more data: AFRINIC (Africa), Arin(North America), APNIC (Asia), LACNIC (Latin America), RIPE NCC (Europe). Anyway, probably all the useful information (IP ranges and Whois) appears already in the first link.

#You can try "automate" this with amass, but it's not very recommended
amass intel -org tesla
amass intel -asn 8911,50313,394161

You can find the IP ranges of an organisation also using http://asnlookup.com/ (it has free API). You can fins the IP and ASN of a domain using http://ipv4info.com/.

Looking for vulnerabilities

At this point we known all the assets inside the scope, so if you are allowed you could launch some vulnerability scanner (Nessus, OpenVAS) over all the hosts. Also, you could launch some port scans or use services like shodan to find open ports and depending on what you find you should take a look in this book to how to pentest several possible service running. Also, It could be worth it to mention that you can also prepare some default username and passwords lists and try to bruteforce services with https://github.com/x90skysn3k/brutespray.

Domains

We know all the companies inside the scope and their assets, it's time to find the domains inside the scope.

Please, note that in the following purposed techniques you can also find subdomains and that information shouldn't be underrated.

First of all you should look for the main domain(s) of each company. For example, for Tesla Inc. is going to be tesla.com.

Reverse DNS

As you have found all the IP ranges of the domains you could try to perform reverse dns lookups on those IPs to find more domains inside the scope. Try to use some dns server of the victim or some well-known dns server (1.1.1.1, 8.8.8.8)

dnsrecon -r <DNS Range> -n <IP_DNS> #DNS reverse of all of the addresses
dnsrecon -d facebook.com -r 157.240.221.35/24 #Using facebooks dns
dnsrecon -r 157.240.221.35/24 -n 1.1.1.1 #Using cloudflares dns
dnsrecon -r 157.240.221.35/24 -n 8.8.8.8 #Using google dns

For this to work, the administrator has to enable manually the PTR. You can also use a online tool for this info: http://ptrarchive.com/

Reverse Whois (loop)

Inside a whois you can find a lot of interesting information like organisation name, address, emails, phone numbers... But which is even more interesting is that you can find more assets related to the company if you perform reverse whois lookups by any of those fields (for example other whois registries where the same email appears). You can use online tools like:

You can automate this task using DomLink (requires a whoxy API key). You can also perform some automatic reverse whois discovery with amass: amass intel -d tesla.com -whois

Note that you can use this technique to discover more domain names every time you find a new domain.

Trackers

If find the same ID of the same tracker in 2 different pages you can suppose that both pages are managed by the same team. For example, if you see the same Google Analytics ID or the same Adsense ID on several pages.

There are some pages that let you search by these trackers and more:

Favicon

Did you know that we can find related domains and sub domains to our target by looking for the same favicon icon hash? This is exactly what favihash.py tool made by @m4ll0k2 does. Here’s how to use it:

cat my_targets.txt | xargs -I %% bash -c 'echo "http://%%/favicon.ico"' > targets.txt
python3 favihash.py -f https://target/favicon.ico -t targets.txt -s
favihash - discover domains with the same favicon icon hash

Simply said, favihash will allow us to discover domains that have the same favicon icon hash as our target.

Other ways

Note that you can use this technique to discover more domain names every time you find a new domain.

Shodan

As you already know the name of the organisation owning the IP space. You can search by that data in shodan using: org:"Tesla, Inc." Check the found hosts for new unexpected domains in the TLS certificate.

You could access the TLS certificate of the main web page, obtain the Organisation name and then search for that name inside the TLS certificates of all the web pages known by shodan with the filter : ssl:"Tesla Motors"

Google

Go to the main page an find something that identifies the company, like the copyright ("Tesla © 2020"). Search for that in google or other browsers to find possible new domains/pages.

Assetfinder

Assetfinder is a tool that look for domains related with a main domain and subdomains of them, pretty amazing.

Looking for vulnerabilities

Check for some domain takeover. Maybe some company is using some a domain but they lost the ownership. Just register it (if cheap enough) and let know the company.

If you find any domain with an IP different from the ones you already found in the assets discovery, you should perform a basic vulnerability scan (using Nessus or OpenVAS) and some port scan with nmap/masscan/shodan. Depending on which services are running you can find in this book some tricks to "attack" them. Note that sometimes the domain is hosted inside an IP that is not controlled by the client, so it's not in the scope, be careful.

Subdomains

We know all the companies inside the scope, all the assets of each company and all the domains related to the companies.

It's time to find all the possible subdomains of each found domain.

DNS

Let's try to get subdomains from the DNS records. We should also try for Zone Transfer (If vulnerable, you should report it).

dnsrecon -a -d tesla.com

OSINT

The fastest way to obtain a lot of subdomains is search in external sources. I'm not going to discuss which sources are the bests and how to use them, but you can find here several utilities: https://pentester.land/cheatsheets/2018/11/14/subdomains-enumeration-cheatsheet.html

A really good place to search for subdomains is https://crt.sh/.

The most used tools are Amass, subfinder, findomain, OneForAll, assetfinder, Sudomy. I would recommend to start using them configuring the API keys, and then start testing other tools or possibilities.

amass enum [-active] [-ip] -d tesla.com
./subfinder-linux-amd64 -d tesla.com [-silent]
./findomain-linux -t tesla.com [--quiet]
python3 oneforall.py --target tesla.com [--dns False] [--req False] run
assetfinder --subs-only <domain>

Another possibly interesting tool is gau. It fetches known URLs from AlienVault's Open Threat Exchange, the Wayback Machine, and Common Crawl for any given domain.

This project offers for free all the subdomains related to bug-bounty programs. You can access this data also using chaospy or even access the scope used by this project https://github.com/projectdiscovery/chaos-public-program-list

You could also find subdomains scrapping the web pages and parsing them (including JS files) searching for subdomains using SubDomainizer or subscraper.

RapidDNS

Quickly find subdomains using RapidDNS API (from link):

rapiddns(){
curl -s "https://rapiddns.io/subdomain/$1?full=1" \
| grep -oP '_blank">\K[^<]*' \
| grep -v http \
| sort -u
}

Shodan

You found dev-int.bigcompanycdn.com, make a Shodan query like the following:

  • http.html:”dev-int.bigcompanycdn.com”

  • http.html:”https://dev-int-bigcompanycdn.com”

DNS Brute force

Let's try to find new subdomains brute-forcing DNS servers using possible subdomain names. The most recommended tools for this are massdns, gobuster, aiodnsbrute and shuffledns. The first one is faster but more prone to errors (you should always check for false positives) and the second one is more reliable (always use gobuster).

For this action you will need some common subdomains lists like:

For massdns you will need to pass as argument the file will all the possible well formed subdomains you want to bruteforce:

sed 's/$/.domain.com/' subdomains.txt > bf-subdomains.txt
./massdns -r resolvers.txt -w /tmp/results.txt bf-subdomains.txt
grep -E "tesla.com. [0-9]+ IN A .+" /tmp/results.txt
Gobuster bruteforcing dns
gobuster dns -d mysite.com -t 50 -w subdomains.txt

VHosts

IP VHosts

You can find some VHosts in IPs using HostHunter

Brute Force

If you suspect that some subdomain can be hidden in a web server you could try to brute force it:

gobuster vhost -u https://mysite.com -t 50 -w subdomains.txt
wfuzz -c -w /usr/share/wordlists/SecLists/Discovery/DNS/subdomains-top1million-20000.txt --hc 400,404,403 -H "Host: FUZZ.example.com" -u http://example.com -t 100
#From https://github.com/allyshka/vhostbrute
vhostbrute.py --url="example.com" --remoteip="10.1.1.15" --base="www.example.com" --vhosts="vhosts_full.list"

DNS Brute Force v2

Once you have finished looking for subdomains you can use dnsgen and altdns to generate possible permutations of the discovered subdomains and use again massdns and gobuster to search new domains.

Buckets Brute Force

While looking for subdomains keep an eye to see if it is pointing to any type of bucket, and in that case check the permissions. Also, as at this point you will know all the domains inside the scope, try to brute force possible bucket names and check the permissions.

Looking for vulnerabilities

Check for possible subdomain takeovers. If the subdomain is pointing to some S3 bucket, check the permissions.

If you find any subdomain with an IP different from the ones you already found in the assets discovery, you should perform a basic vulnerability scan (using Nessus or OpenVAS) and some port scan with nmap/masscan/shodan. Depending on which services are running you can find in this book some tricks to "attack" them. Note that sometimes the subdomain is hosted inside an IP that is not controlled by the client, so it's not in the scope, be careful.

Web servers hunting

We have found all the companies and their assets and we know IP ranges, domains and subdomains inside the scope. It's time to search for web servers.

In the previous steps probably you have already perform some recon to the IPs and domains discovered, so you may already found all the possible web servers. However, if you haven't we are now going to see some fast tricks to search for web servers inside the scope.

Please, note that this will be oriented to search for web apps, you should perform the vulnerability and port scanning also (if allowed by the scope).

A fast method to discover ports open related to web servers using masscan can be found here. Another friendly tool to look for web servers is httprobe and fprobe. You just pass a list of domains and it will try to connect to port 80 (http) and 443 (https). You can additional indicate to try other ports:

cat /tmp/domains.txt | httprobe #Test all domains inside the file for port 80 and 443
cat /tmp/domains.txt | httprobe -p http:8080 -p https:8443 #Check port 80, 443 and 8080 and 8443

Screenshots

Now that you have discovered all the web servers running in the scope (in IPs of the company and all the domains and subdomains) you probably don't know where to start. So, let's make it simple and start just taking screenshots of all of them. Just taking a look to the main page of all of them you could find weird endpoints more prone to be vulnerable.

To perform the proposed idea you can use EyeWitness, HttpScreenshot, Aquatone, shutter or webscreenshot.

Recapitulation 1

Congratulations! At this point you have already perform all the basic enumeration. Yes, it's basic because a lot more enumeration can be done (will see more tricks later). Do you know that the BBs experts recommends to spend only 10-15mins in this phase? But don't worry, one you have practice you will do this even faster than that.

So you have already:

  1. Found all the companies inside the scope

  2. Found all the assets belonging to the companies (and perform some vuln scan if in scope)

  3. Found all the domains belonging to the companies

  4. Found all the subdomains of the domains (any subdomain takeover?)

  5. Found all the web servers and took a screenshot of them (anything weird worth a deeper look?)

Then, it's time for the real Bug Bounty hunt! In this methodology I'm not going to talk about how to scan hosts (you can see a guide for that here), how to use tools like Nessus or OpenVas to perform a vuln scan or how to look for vulnerabilities in the services open (this book already contains tons of information about possible vulnerabilities on a lot of common services). But, don't forget that if the scope allows it, you should give it a try.

Bug hunting OSINT related information

Now that we have built the list of assets of our scope it's time to search for some OSINT low-hanging fruits.

Api keys leaks in github

Dorks: AWS_SECRET_ACCESS_KEY, API KEY, API SECRET, API TOKEN… ROOT PASSWORD, ADMIN PASSWORD, COMPANYNAME SECRET, COMPANYNAME ROOT, GCP SECRET, AWS SECRET, “username password” extension:sql, “private” extension:pgp...

More Github Dorks

  • extension:pem private

  • extension:ppk private

  • extension:sql mysql dump password

  • extension:json api.forecast.io

  • extension:json mongolab.com

  • extension:yaml mongolab.com

  • extension:ica [WFClient] Password=

  • extension:avastlic “support.avast.com”

  • extension:js jsforce conn.login

  • extension:json googleusercontent client_secret

Anyway, the majority of the vulnerabilities found by bug hunters resides inside web applications, so at this point I would like to talk about a web application testing methodology, and you can find this information here.

Recapitulation 2

Congratulations! The testing has finished! I hope you have find some vulnerabilities.

At this point you should have already read the Pentesting Web Methodology and applied it to the scope. As you can see there is a lot of different vulnerabilities to search for.

If you have find any vulnerability thanks to this book, please reference the book in your write-up.