Content here is by Michael Still All opinions are my own.
See recent comments. RSS feed of all comments.

Thu, 04 Dec 2008

gtalkbot 1.4

    Its been a while between gtalkbot releases, and that's mainly because the current version "just works" and no one has reported any bugs or feature requests. However, the other day I was running a lot of long lived commands, and started getting annoyed at not knowing when they were finished. I thought it would be nice to get sent an IM when a command had completed. So, I implemented that in this new release.

    There is a simple shell script called, which you need to modify to have the right path for an outbox directory. This directory is watched by gtalkbot, and files in the directory are processed as part of the idle loop. The format for files in the directory is the jabber ID to send to on the first line, and then the message on the remaining lines. I guess that means you don't have to use the shell script at all if you want, just have something dump files in the magic directory.

    You then need to add this line to the gtalkbot config file:

      password	mypassword
      pluginsdir	/data/src/gtalkbot/plugins
      outboxdir	/data/src/gtalkbot/outbox
      authfile	/data/src/.gtalkbot-passwd

    So, there you go. I now get IMs when commands are completed, by running them like this:

      $ longcommand; echo "long command done" | notify

    The source for gtalkbot 1.4 is here.

    Tags for this post: gtalkbot release
    Related posts: gtalkbot 1.3; Pngtools 0.3 released; PNGtools 0.4; MythNetTV release 1; MythNetTV release 2; MythIPTV Beta 2

posted at: 17:25 | path: /gtalkbot | permanent link to this entry

Sat, 07 Apr 2007

gtalkbot 1.3

posted at: 21:56 | path: /gtalkbot | permanent link to this entry

Wed, 31 Jan 2007

Hey look, Secunia reads FreshMeat!

    I hope no one is paying Secunia for their security advisories. I release gtalkbot 1.0 (where you had to pass the GTalk user name and password on the command line), and then changed that behaviour in 1.1. The Secunia rocket scientists figured out all by themselves that this was bad. Bad enough for a security advisory?

    A security issue has been reported in gtalkbot, which can be exploited by malicious, local users to disclose sensitive information.

    The problem is that certain user credentials are passed to the application as arguments on the command line. This can be exploited to gain knowledge of usernames and passwords of other services via the process list.

    The security issue is reported in versions prior to 1.1.

    Ummm, the GTalk account is created for the purpose, and so it's not uber secure anyways. In fact, it's only visible to local users, who are presumably trusted anyways given that gtalkbot also needs the unauthenticated telnet interface to MythTV enabled to work. Wow. I assume that Secunia just reads every FreshMeat security release, and makes an announcement about it. Oh, and those three nearly paragraphs took over two weeks!

    Update: but wait, there's more! I made it into the US Federal Government's vulnerability database too, complete with an incorrect "Authentication: Not required to exploit". I guess the Feds can't read python code?

    Update: perhaps Secunia is just reprinting this lame advisory? Do these people just reprint each other's work all the time? Again with the hoping people aren't earning money by making suckers think they're helping...

    The fun continues: yay for SecWatch and systembodyguard!

    Tags for this post: gtalkbot secunia security advisory
    Related posts: Security Hyperventilating; gtalkbot 1.2; Renaming mbot to gtalkbot; So much for the Kensington lock controversy; gtalkbot 1.4; Theorising on Kensington lock insecurity

posted at: 22:08 | path: /gtalkbot | permanent link to this entry

Sun, 14 Jan 2007

gtalkbot 1.2

posted at: 21:14 | path: /gtalkbot | permanent link to this entry

Fri, 12 Jan 2007

gtalkbot 1.1

posted at: 20:33 | path: /gtalkbot | permanent link to this entry

Fri, 29 Dec 2006

Renaming mbot to gtalkbot

posted at: 19:39 | path: /gtalkbot | permanent link to this entry

Sun, 19 Nov 2006

mbot: new hotness in Google Talk bots

    I've been meaning to release this for a while, but haven't had a chance in the last month or so. Therefore, instead of writing it up fully I'm just going to put it online and document it better when I get the chance.

    mbot is a Google Talk / Jabber bot engine, which works by loading plug ins (it calls them modules) which register verbs. These verbs are the first word on an instant message line, and are handed to the module when a registered verb appears (along with the rest of the line).

    The source download includes a sample module, which implements instant messaging access to the MythTV network control interface and on screen display functionality. mbot is in fact a refactoring of a bot which originally did just the MythTV bits without pretending to be modular.

    Mad props to PyXMPP for making the Google Talk / Jabber bits easy, the BZR folks for their excellent sample of how to do plug ins, Jaq for telling me my code sucks, and the Sydney office for giving me somewhere with quiet evenings to write code.

    Anyway, better documentation later when I get a chance.

    Tags for this post: gtalkbot mbot python google talk gtalk jabber bot pyxmpp
    Related posts: mbot: new hotness in Google Talk bots; gtalkbot 1.2; Renaming mbot to gtalkbot; Getting Google Talk working with PyXMPP; gtalkbot 1.1; A MythTV Jabber bot

posted at: 16:56 | path: /gtalkbot | permanent link to this entry