Internet Relay Chat Probe (IRCP)

TRIPLE 6 SEVEN OCULOUS

A robust information gathering tool for large scale reconnaissance on Internet Relay Chat servers, made for future usage with internetrelaychat.org for public statistics on the protocol.

Meant to be used in combination with masscan checking 0.0.0.0/0 (the entire IPv4 range) for ports 6660-6669, 6697, 7000, & other common IRC ports.

The idea is to create a proof-of-concept documenting how large-scale information gathering on the IRC protocol can be malicious & invasive to privacy, while also yielding deep-dive look at the IRC protocol & it's internal statistics & commonalities.

Usage

The only required arguement to pass is a direct path to the targets list, which should be a text file containing a new-line seperated list of targets.

Targets must be a valid IPv4 or IPv6 address & can optionally be suffixed with a port.

Edit ircp.py & tweak the settings to your favor, though they rest with sane defaults.

Order of Operations

First, an attempt to connect using SSL/TLS is made, which will fall back to a standard connection if it fails. If a non-standard port was given, both standard & secure connection attempts are made on the port as-well. The RPL_ISUPPORT (005) response is checked for the SSL= option to try & locate secure ports.

Once connected, server information is gathered from ADMIN, CAP LS, COMMANDS, HELP, MODULES -all, VERSION, IRCOPS, MAP, INFO, LINKS, SERVLIST, STATS p, & LIST replies. An attempt to register a nickname is then made by trying to contact NickServ.

Lastly, every channel is joined with a WHO command sent & every new nick found gets a WHOIS sent. Registered channels & nicks are issued a NickServ/ChanServ INFO command. CTCP requests are sent to channels & nicks aswell.

Once we have finishing scanning a server, the information found is saved to a JSON file. The data in the logs are stored in categories based on numerics (001 is RPL_WELCOME, 322 is RPL_LIST, etc) & events (JOIN, MODE, KILL, etc).

Everything is done in a carefully throttled manner for stealth to avoid detection. An extensive amount research on IRC daemons, services, & common practices used by network administrators was done & has fine tuned this project to be able to evade common triggers that thwart what we are doing.

Preview

Threat Scope

While IRC is an generally unfavored chat protocol as of 2023 (roughly 7,000 networks), it still has a beating heart (over 300,000 users & channels) with potential for user growth & active development being done on IRCv3 protocol implementations.

Point is..it's is not going anywhere. With that being said, every network being on the same port leads way for a lot of potential threats:

• A new RCE is found for a very common IRC bot
• A new 0day is found for a certain IRCd version
• Old IRC daemons running versions with known CVE's
• Tracing users network/channel whereabouts
• Mass spamming attacks on every network

Mass scanning default ports of services is nothing new & though port 6667 is not a common target, running an IRCd on a non-standard port should be the standard. If we have learned anything in the last 10 years, using standard ports for anything is almost always smells like a bad idea.

Todo

• Built in identd
• Checking for IPv6 availability (SSL= in 005 responses may help verify IPv6)
• Support for IRC servers using old versions of SSL
• Support for hostnames in targets list (Attempt IPv6 & fallback to IPv4)
• Support for multiple vhost
• How do we handle the possibility of connecting to multiple servers linked to same network?
• Seperate lists for failed & banned networks.
• Learn network target-change throttles from 439 ERR_TARGETTOOFAST replies (Research IRCd defaults)
• Store last command execute to detect triggers

Opt-out

You can request to opt out of our scans by sending an email to scan@internetrelaychat.org

Mirrors

acid.vegas • GitHub • GitLab • SuperNETs