eve-bot - the mumble 1.2 delayed relay bot

Eve the mumble eavesdropper bot

This bot is written for online multiplayer communities playing games like TF2 or CSS, where the spectator's view of the game is sometimes delayed to prevent collusion.

Specators may wish the listen to the communication of the match team, however the delay (typically 90 seconds) between the in-game action and the mumble server normally makes this impossible. This bot was written to facilitate that.

The bot joins the match channel and listens to all comms. For every user in the channel, a mimic will join the spectate channel, and the eavesdropper bot will relay the comms to the correct mimic, with the specified delay. As the bot joins as a normal user, no special murmur server support is required.

The bot is available in 3 versions; a compiled windows package, a compiled ELF binary for Linux or BSD, and the original python script, which should work anywhere you have python, although it might require installing a couple of python modules first.

Windows and linux compiled downloads

Extract the files from the archive and run eve-bot.exe (or eve-bot for the *nix version) from a command prompt. On windows, you might need the MSVC runtime although this is likely to have been installed by one thing or other already. The windows version is heavily tested; the unix version less so. In either case, if you encounter errors, please contact me and I'll try to fix them.

Python source - prerequisites

• This bot needs python with ssl support. This either requires version 2.6 or for the ssl module to be downloaded and installed seperately from PyPi
• Mumble uses protocol buffers for control messages. To run the bot, the python protobuf modules and the protoc compiler must be installed. The protoc compiler generates the python module necessary for the bot to understand the control messages. To do this, navigate to the directory the bot will be run from and type

  protoc --python_out=. Mumble.proto

  You must also have Mumble.proto in the current directory. The latest version of this can be found in the mumble/murmur source code

Remember that if you have trouble running the script version, compiled versions are available.

Using the bot

(If you are using the windows binary, the command will be eve-bot.exe, and for the ELF binary it will be eve-bot)

In the simplest case, running the bot by typing

eve-bot.py -e "Match Channel" -r "Spectator Channel"

This will attempt to connect to a mumble server on the default port, on the same machine as the bot is running on. The channels MUST exist; they can be the same if you wish, though this is likely to cause confusion. Running the bot with no options will give a list of the available parameters.

If you are running this from a different machine, and the server is example.com at port 1234:

eve-bot.py -e "Match Channel" -r "Spectator Channel" -s "example.com" -p 1234

Automatic ban

By default, the mumble server bans for too many connection attempts in a short period of time. This can cause issues when testing, and if many people join the eavesdrop channel at once (many mimics will need to be spawned). If this happens, consider changing the autobanAttempts value upwards.

Known Issues

• Very few error messages from the server are checked. For instance, if the bot's name is not valid, no helpful error message will be returned.
• The bots don't notice if they have been moved; this won't affect mimics, but the eavesdropper bot will only attempt to listen to people who are in the correct channel, so unless it's moved to a linked channel it won't hear much.
• If the bot refuses to work properly, the -v option will decode and print all control messages from the server, which should indicate the problem. (The main one is if the bot tries to join using a name that's already in use)
• The bot should not eavesdrop in the root channel - every client which joins the server initially joins the root channel, so attempting to listen there would lead to an infinite amount of mimics joining.
• Passworded access to a server has not been checked in mumble 1.2.
• The bot has only been checked with python version 2.5.2 (with external ssl module) and 2.6.4; if it fails to work on earlier versions, PLEASE contact me and I can see about rewriting the relevant part of the script.

Planned Features

Depending on feedback, there are several features that could be added to the bot:

• Ability to change the speech delay by sending a text message to the bot
• Ability to have the mimics join a different server than the eavesdropper
  (this would basically involve disabling some "infinite-loop-of-mimics" checks)
• (Hard) Ability to turn the volume level of the mimic's speech down
  (This would require parsing and decoding the speech packets instead of just sending them on, and would increase the CPU usage and complexity of the bot)

None of these is likely to be written unless there is interest, so:
Feedback greatly appreciated
frymaster@127001.org