lee/hpr-knowledge-base
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Initial commit: HPR Knowledge Base MCP Server

Browse Source 
- MCP server with stdio transport for local use
- Search episodes, transcripts, hosts, and series
- 4,511 episodes with metadata and transcripts
- Data loader with in-memory JSON storage

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
This commit is contained in:
Lee Hanken
2025-10-26 10:54:13 +00:00
commit 7c8efd2228
4494 changed files with 1705541 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

290
hpr_transcripts/hpr2356.txt Normal file
View File

@@ -0,0 +1,290 @@
Episode: 2356
Title: HPR2356: Safely enabling ssh in the default Raspbian Image
Source: https://hub.hackerpublicradio.org/ccdn.php?filename=/eps/hpr2356/hpr2356.mp3
Transcribed: 2025-10-19 01:39:12
---
This in HPR episode 2,356 entitled, safely enabling SSH in the default Russian image.
It is posted by Ken Fallen and in about 33 minutes long and Karima next visit flag.
The summary is, Ken walks us through the crypt to secure the most Russian pixel image.
This episode of HPR is brought to you by Ananasthos.com.
Get 15% discount on all shared hosting with the offer code HPR15, that's HPR15.
Better web hosting that's honest and fair at Ananasthos.com.
Hi everybody, my name is Ken Fallen and you're listening to another episode of Hacker Public Radio.
Today I want to talk to you about re-enabling SSH on the default Razbian image.
Recently they've disabled us because having SSH enabled by default, having a known username of Pi with a known password of Raspberry.
And having that in the hands of 11 million of these devices in the hands of people who by definition are beginners as computing is probably not a good idea.
So they've rightly disabled that.
You can enable it, of course, quite easily via the GUI preferences, interfaces SSH enabled or via Raz being config,
in face options SSH, yes, and okay, finish.
There is however a third option.
So when you're burning the image itself, you can either put it into a Pi and then log in with a keyboard video and a keyboard,
it should be my screen and mouse, and then create a file called SSH on the boot partition.
You can shortcut this by when you're burning the image you can eject us and then re-inject us, re-insert it into your computer,
open it up the boot image and then create a file called SSH in the boot directory.
It actually doesn't matter what's in there, couldn't contain text or nothing at all, and that will enable SSH by default.
Or you could actually go into the configuration and turn it on manually.
Boss, I want to do another of these things because in that amount of time that I basically want it
coming out of the box already enabled with SSH enabled in multiple images and I want to enable it as soon as it downloads.
So basically that will involve mounting the image, writing that file, but I want to do a few other things as well
with the image. I want to change the root password and change the Pi password to something very very long
and on Wheel D. And then I also want to make a few settings to the SSH server that's going to be running on the Pi image
that's the one that listens for your incoming connections when you SSH to my Pi at my Pi.
That's the one that's listening over there. And that will forbid the Roosh user from ever logging in via SSH.
And we also are going to prevent any user logging in unless they're using SSH keys.
What are SSH keys can? Well, they're a cool thing that is asynchronous.
You get a public in a private key, your public key is secreting, you keep it and your private key,
you put it onto the Pi. And that means once you have unlocked your public key, you can type SSH Pi at Pi
and it will do the key exchange, confirm that you are who you say you are and allow you access
without a username and password. Surely that's more insecure. Actually, it's not. So that's what we're
going to do. For those of you who don't want to stick around for the long excruciating
explanation of what I'm doing, there is a bash script in the show notes and there are details
show notes available to explain how to do this using with step-by-step instructions. So if you
just want to have a look at that bash script, feel free to do so. I've deliberately met it so
that you need to edit it. But the steps are kind of obvious. I don't go into massive
customizations. There's loads of stuff I could do. I could change the background. I could create
other users. I could do all sorts of stuff like that. But I want to not do that because experience,
a lot of experience has shown that that's not a good thing. You should use automation tools
like Ansible Chef Puppet or CF Engine or something like that and manage your Raspberry Pi
with those tools if you've got a lot of them or if you've only got one or two then do it however
you want to do it yourself but using SSH keys. If you have a lot of Raspberry Pi's and don't like
this image thing, another approach is to use the network boot option linked in the show notes,
which will allow you to do what's similar to a pixie boot in the past where you mount an NFS mount
immediately so that the Raspberry image boots up, connects to the network and then mounts
its home directory, kernel image, everything comes across in the network. In that situation,
there's very little read and writes on the SD card, very little chance that it will get damaged
and you can probably use a smaller SD card or something like that. But then those units just become
quite easy and replaceable. So the rest of the show will basically be describing how I don't
know the image, how I verify it's intact, extracting the image, enabling SSH, changing the
username and password and securing the SSH server for the Pi. So the first thing that we need to do
is download the image from the Raspberry and Website. They're using a show once on to confirm that
the file is intact, that has actually been shown to be compromised. So this should be used in show
256 so you can expect that change coming soon, I guess. I personally use, I get the show one
sum from the Website and then I show one sum and then the image zip file name and I pipe that to
grep with the show one sum coming out and that way if grep is able to find it, then I know that
the show one has passed, which is one way to do this. I use seven zip, which is usually not installed
by default as my extraction tool of choice, largely because I use quite a lot of seven zip files,
they're the highest compression method for large XML files, taking in my experience a 1.5 gig file
down to something like 35 megabytes, 1.5 gig XML file. So good compression. Nice thing about it is
you can throw zip files at it, you can throw B, you can do files at it and it's free and opens
or software as well, available in all repositories. So I'm doing LS here and the show one sums,
which you can, which I'm going to use for later. So that, after unzipping the file, we're left with
an image file and some of you may be asking, while you hear this term, I've gotten, download this
image for this and download this image for that. So what exactly is an image file? And it's kind of
important to understand. In our case, an image file is a bit by bit copy of what the developer has
on their desk. So they have a SD card and a position one, second one, blah one, there's a number
and that's either one or zero. And after finishing this, the position, the position one will be the
same position two will be the same position three and so forth right through so that you will have
an identical copy of your SD card, including all empty space and everything will be on your SD card.
Now to make it transporter of that as efficient as possible, the first thing you usually do in
your Raspberry Pi is you extend those partitions to take in the additional space that's going to be
on yours, so you're not transporting four gigabytes of empty zeros at the end of their disk. But
what you're doing when you get an image file, be it an image file of a CD or DVD or in fact,
Raspberry an image, you're getting a bit by bit copy of their file. Okay, so you can mount that
using the Raspberry or the Linux mount command and that has supported the loop option for years
and probably lots of you have mounted CDs and DVD images using the loop option in the past.
But the first thing I would like you to do is nip over to have a look at what the boot sequence is
for for a Raspberry and pixel just desktop. What is the boot sequence and it gives you
the basics which is absolutely great infographic. I won't go through here, it tells you the names
of all the files where they're located on the boot partition of the SD card and what's loaded,
what file name is loaded and that file loads the next one and that file loads the next one.
Very good infographic, so you can have a look at that. So down to the nuts and bolts, we now have
the image and in order to enable SSH, we need to write a file into the boot directory of this SD card
and how we can do that is we can mount the image. No problem there, we've all mounted DVDs and stuff.
Well, there kind of is. Thanks to that too, coffee.
In this case, unlike a CD image, this SD card has got two partitions. One is a small boot partition
and the other is a larger Linux partition. At the beginning, there is partition information
to tell you where the boot partition is and the root partition or the Linux partition is.
So how do we find out that? So I have a link there to an Ask Ubuntu question that will show you
the answer to this and that is using fdisk with the dash dash list and the dash dash units option.
So the list shows you the list partition table and the units shows you the size in sectors or
cylinders of that. So when we run fdisk space dash dash list, space dash dash units on 2017, 06,
21 runs been jesse.image in our case. We see the second line has unit sectors of one multiple by
252 bytes, which is the important one, blah, blah, blah, blah, blah, device, boot, start, and thingy.
And then we see there are two images in there, two partitions. One of type win 95, fat 32,
LBA and the other of type Linux. One is 41.7 meg and the other is 4.3 megabit. And we also see
a start and then section. So it's the start that we're interested in. That is 8192. So essentially,
what that does is it gives you that is the number of units from the start that that partition begins.
So the units, kids are home, next door. So we can multiply 8192 by 512 and we will get an answer
for exactly how many bytes from the very first byte we need to start in order to access and mount
this partition. And we can get that by running that command and using Oc and using this slash slash
parameter with the character sign units showing sort of grip as in filter out the units,
something beginning with units. And then I want you to print me the second to last
piece of information in which case in our case, that's 512, give us backs with the number of
units. And we can do the same thing for the start value. And we're in that case, we're
awking and we're filtering for win 95, fat 32 and we're printing the second value, which is
the start value 8192 comes back. That's wonderful. And because we've all spent time listening to Dave
Morris, we know that we can multiply two numbers together by using the dollar, echo, dollar,
open bracket, open bracket, one number multiplied by the other number, close bracket, close bracket.
In our case, 8192 by 512 will give you 4194304. And if we use that with mount space dash O, space loop
comma offset 4194304 and then the image and then slash mount SD card, which is where we're going to
mount it on our local machine, then we will find that our disk mounts. And then in the show notes,
I give you instructions on how to do all of that in the script. So if you do LS slash MNT slash SD
card, you get a whole go of, you basically get the boot device where you can see config.text
and I don't know, kernel.image and license broadcom license article blah blah blah and all the files for
they and for the mounting the rasbin boot image. Now you could use a text editor at this point,
we're doing a script. So I'm just going to touch the command using create the file using the touch
command slash mount slash SD card slash SSH and that creates a zero byte file in the in that location.
So what we're going to do now is on mount the slash mount SD card. So we're finished with the
root partition because we don't want to do we will now enable SSH. You could
take the image, put it into your Raspberry Pi and boot it up and you're back into the exact
same position you're listening, your computer will be listening on an open port 22 and the
username Raspberry, username Pi and Rasp, username Pi password Raspberry and you'll be able to log in.
That's not really what we want to do. What we want to do is generate some SSH keys here.
If you are not familiar with SSH keys and give me a link in the show notes to a good overview of
SSH keys, why, what are the advantages? Your password is never sent over the network,
eliminates brutes, risks of brute force attack and that's a lot more convenient,
couldn't agree more. If you've never generated SSH keys, the best SSH key to use nowadays is
ED25519 and you can generate those using SSH-keygen space-t which will determine the type ED25519
space-f which will specify the file name and my guess I'm just going to pick the default in the
look wherever it happened to be, ID ED25519 and dash c and put it in Raspberry Pi keys for Ansible
and that will create two keys for your public key which has got a dot pub and a private key which
has got which is slightly larger file 111 for the dot pub and 419 bytes for the private file.
Your private file cannot stress enough, you will be asked to put in a username and password,
you should do that, cannot stress enough that you keep the private key very, very safe,
make sure you've got copies of it because you will need to be able to get access to it again.
The public key, you don't really care what it is. Now if you don't understand the reason why
there is an excellent video released by the CAN Academy and it's released under CC by SA.
I believe NC perhaps and it's been stitched together the various different sections of a
by Brick Crew, it's originally a Brick Crews video and goes to an absolutely excellent description
of how where SSH keys come from, why they were needed, why they were developed and the advantage
of them, essentially it boils down to mixing two different colors of paint. So you share the same
color so you're going to mix it with red and you have one color green and the other person has
some other color and by mixing the two colors with red you pass the the outcome of that to them
and they pass the outcome of theirs to you and you mix your color, secret color with their mix
colors that will give their other color, very, very good. If that doesn't make a lot of sense,
have a look at the video, it's well worth the watch. So I show out here what they, what's the
contents of the two key files, obviously, you keep your private key very private and you can
copy it into your own SSH directory and that will allow you to, that's where your system will be
expecting to find us. So this is the SSH directory and your computer not on the pie unless you
happen to be, your computer is a pie in this case. So the one that you are going to be connecting from,
that's where you put your public key in. You can, by the way, use the dash i command to specify
a different location but that's outside the confines of this. Anyway, onto the next step,
what we want to do is do the same thing as we did with the mount in the boot directory. We want
mount the root directory which is Linux. So we're going to do exactly the same commands that we did
before. We're going to mount the image. We're going to use och to look for units and we're instead
of having och look for win 30 win 95. It's going to look for Linux. It's going to multiply the sizes
out and it's going to mount them in an SD card. So then at the end of this, we will have the SD card
slash mount slash SD card will mount the root directory. So for example, slash home slash pie
will be available. So the eventual home slash home slash pie directory where you would be when
your pie boots up is now available on your other master computer under slash mount slash SD card
slash home slash pie. Clear, clear as well. Okay, so now what we want to do is we want to copy over
your public key to the image so that when the pie boots up, it knows where to find a list of
keys that are authorized to log into the it's computer. And those are kept in a hidden directory.
Hidden is denoted by a any directory starting with a period of full stop followed by in this case
SSH. So in a dot SSH directory in your home directory, if there's a file called authorized
underscore keys, that will contain a list of public keys. They're in the in the files, all the
public keys that are allowed. So you your friends, whoever needs taxes, this pie, each of you have
got your own public key for your own private key. So for example, those three of you and you would
have a text file with three keys one after the other underneath there. And that's in your slash
home slash pie dot SSH slash authorized underscore keys file, which of course is mounted on slash mount
slash SD card slash home slash pie slash dot SSH slash authorized keys. Yeah, so all we need to do is
cat our ID underscore ED25519 dot pop into that file link in the show notes. And as you can
see in the show notes, we copy that file over. Unfortunately, it's not as simple as that before
we do that, we actually need to make the directory because the dust SSH directory does not exist.
So we can we use the make your command. Now SSH is absolutely paranoid about any changes on
this. If the on these directories and these files, if the files are not formatted correctly, if they
don't have the right permissions, if they're in the wrong location, it will not allow somebody to
log in, which is right and correct. So we have a series of commands that we need to do three.
In fact, first one is we need to make a directory. The second one is we need to
change
mod on that, change the mod on it. So it's 700 allowing the
user to read, write and execute and the execute from the directory being the ability to change into
the directory. And then we change the ownership to 1000, 1000, which I'll get to in a minute.
So SSH is very paranoid about that. So where does the 1000, 1000 come from for the permissions?
Well, if you are on this computer here, for example, that isn't a pie. It isn't going to have a
pie username created. It's going to have your username, Joe or Bob or whatever, BLZBub.
And that will have a user ID. So normally, when I wanted to change permissions, I can go change
owner, can, call on, can. So it's the group. It's a user called can and the group called can. But as
this doesn't exist on this image, we need to find out what that is. So how we do that is we
check GREP for the lines beginning with pi on slash mount slash SD card slash ETC slash pass WD,
P-A-S-S-W-D. And that will show pi, call on X, call on 1000, call on 1000. And then come, come,
come, come, come. And then home, pi, call on. And that basically is the line that tells the system
what user ID number, what the user name is and what the ID is of the user. And we do the same
thing. EGREP lines beginning with pi, a charat symbol, like a hat or shift six on the US keyboard.
slash mount slash SD card ETC in groups. And we see that the pi group has got an ID of 1000.
So in both cases, we can use change on 1000, call on 1000. Because even though 1000, 1000 is
probably the user that you are using on your own laptop right at the minute, because you're
more than likely the first and only user create there, that's going to be the user of the pi over
over there. So we change the ownership of the authorized key file to be what will be the
pi user when we boot that machine up. And we change the octop permissions on that file to 600,
which is read and write. We don't need execute on this. Execute on the file means this file is
allowed to load runtime code into the processor. As in, it is a program. So that prevents some
malicious attacker getting access to this file and using it as a way to run code on the machine.
So that's the that's the authorized keys part done. Excellent. Now what we need to do is restrict
the route and require keys for the pi's SSH server. So in order to do that, we need to
tell us they file SSHD underscore config, which is kept in the slash ETC file of the pi, which in
our case would be slash mount slash SD card slash ETC and then SSHD underscore config. And there are
two parameters that we want to change. One is permit route login. If this option is set to know,
route is not allowed to login. Just for clarity, I put in via SSH. And then password authentication
specifies whether password authentication is allowed. The default is yes. And what we're going to do
is we're going to change password authentication to know and permit route login to know. And I'm not
going to tell you how to do that. It's using said, if you want to know more about said, look
update morse show, which is linked in the show notes. So the next thing we're going to do is we need
to change the route and pi user password. Now this one is a little bit more complicated. So you can
check for I'm using eGrip called route pi called slash mount slash SD card slash ETC slash password.
And I see that there's a route user and a pi user in there and both have X in the second column,
which means that their encrypted password is installed in the slash ETC shadow file. And if we
look in their slash mount SD card ETC, we see that the route has got an asterisk, meaning there is
no password session. It's not allowed to log in. And the pi has a password and it begins with dollar
six. And a dollar six is tells a system that is a show 512 hashed password. And normally if you're
doing this trick of generating passwords, you get a dollar five, which is an empty five password,
which is not secure anymore. So you're looking for a show six. Now the best option I could come
across to do that was a link for how to over its server fault.com. And the answer was given by
Davy, how do I create a show 25512 hashed password for shadow. And here she uses a inline python
three script to import crypt and then print crypt dot crypt and then correct hard spot tree stable
and then add some salt to the end of that file. And basically creating a root password
in a root password variable. And I do the same thing for the pi user. And then I use said again
to replace to find the root user and then change their password and find the bio user and change
the password. And we can confirm that those are changed. So those are all three things that we need
to do. So what we can do now is on mount the SD card using you mount space slash mount slash SD card.
And I'm going to rename the files. So I just rename them dash SSH enabled. And then when we do an
LS of the images, we see that the images that we have just created are I've got the same file size
as the images that we downloaded from the Raspberry and website, which is kind of strange.
And but if we do a show one some, we find that the images the check sums have in fact changed.
So there you go. That is because the file size did not change because we were just changing
ones and zeros that were already there. Two different ones and zeros. But they show some will look
at those order of ones and zeros and say, are they the same or not? And we'll give you a hash of those.
So then we have an image that we're ready to burn. And burn is a traditional term coming from
taking an ISO image and putting a seat on to a CD. And the laser literally burns with a coating on
the CD or DVD cover. And we're going to use the normal way of doing this, which is DD space BS equals
for M. So for Meg, I think by sector, don't know what that sensor didn't look it up. And there's a
cool new option in the recent versions of DD to give you a status equal progress. And that
you know, prints out some progress. Don't know how useful it is, but at least you know that
their program is doing something. And then the IF file infile equals the image file with SSH enabled.
And the OF is slash dev slash MMCBLK zero in my case. How you can check to see if that's
the same for your case is first extract the SD card. Do an LSBLK
and see the list. Insert the card again and do an LSBLK. And then you will see the differences
between the two outputs will more than likely be yours. Just to be sure, I tend to eject all other
things, my phone, my sensor tip, everything so that I know 100% sure that what I'm burning the
images on to is what I'm burning the image on to. You need to be very, very careful. Be awake,
engage brain folks, engage brain. Okay, so now what we have is a regular old Raspberry image,
which during the course of researching this and writing it up, they actually changed and
updated the image. So I am able to confirm that it still works. And when I put it into my pie,
I was able to do SSH space dash i and then the ID underscore ed25519 space pie at and then the
IP address in my pie and then boom, I'm logged in without using the password. If I try to log in
in as root, it says reject it. If I try and log in with pie without a key, it says reject it.
But that's not really very handy because every time I do that, I need to add the password.
And now you might be tempted to put in the password like that we've generated before
in the script area on, but that is not the password that's been asked for. You've been asked for
the password of your SSH key. So what I have in the morning is first thing I do, a boot off my machine,
I grab my cup of coffee and I type SSH dash agent. Actually, we all know I don't do that. I type
a shorter key sequence that does a load of other things, one of which is running SSH dash agent.
I enter my very long password and I do mean very long password and I press enter and then for
the rest of the day, I can then SSH into any server, any server that I should make use on it,
any server that has my SSH public key on it in the authorized key file, I will be able to log
into that without being challenged for using it or password. And that is it. That is how you enable SSH
securely on a Raspbian image. I think I have managed to make it a lot more complicated than
it actually is. Download the script, have a look at it, modify it for your needs, and you can even like,
I don't know, change the fault using name and password to some random strings or whatever,
because you probably won't be needing them. All right, tune in tomorrow for another exciting
episode of Hacker Public Radio.
You've been listening to Hacker Public Radio at Hacker Public Radio.org. We are a community
podcast network that releases shows every weekday Monday through Friday. Today's show, like all our
shows, was contributed by an HBR listener like yourself. If you ever thought of recording a
podcast, then click on our contributing to find out how easy it really is. Hacker Public Radio was
founded by the digital dog pound and the infonomicon computer club and is part of the binary revolution
at binrev.com. If you have comments on today's show, please email the host directly, leave a comment
on the website or record a follow-up episode yourself. Unless otherwise status, today's show is
released on the creative comments, attribution, share a life, 3.0 license.