Photographed relaxing above on his couch is my cat, Plop.
This is the second of twelve planned posts. (You can read the first.) This is an experiment in public goal-setting and I am pleased to be on-track — so far. I’ll tell you briefly about what’s been on my mind this month to help you determine whether this post and the rest of the series will be of interest to you:
- technical leadership
While I am very much a “hands on the keyboard” kind of person, we are human beings. Things in the “technical leadership” category like how to embrace change and deal with uncertainty are at least as important as remote code execution vulnerabilities. Keep reading if you are interested in GraphQL, Kubernetes, infosec, cool tools and all of the human stuff that goes along with working in tech. I tend to write about what’s on my mind during the course of a given month: it won’t all be hyper-recent current events stuff. Discussion of things like ChatGPT will very likely soon show their age but there are many less time-sensitive topics here, too.
P.S. I am inclined to publish these posts via email. Buttondown seems a reasonable choice. Subscribe at buttondown.email/brie if you want to see these in your inbox once per month in 2023.
🔖 Good Reads
Take a few moments, grab a notebook and read through these articles!
- LastPass says employee’s home computer was hacked and corporate vault taken
- I want to lose every debate.
- Aaron Parecki’s life stack
- Email: explained from first principles
- The curious case of the disappearing Polish S: One keyboard bug three decades in the making!
- I would both love and hate to track down something like this!
- How to build an in-house on-call training program
- Exploring Prompt Injection Attacks
- This is more useful for fun than profit at present.
- You can’t solve AI security problems with more AI
🔒 RCE of RCE
Really Cool Example(s) of Remote Code Execution
I love a good write-up especially when remote code execution is involved! (Who doesn’t?) My favorite read this month:
Even if this kind of thing is not your typical jam, I would recommend reading this one. It’s pretty easy to follow along even if you are not already familiar with PHP (!) or Cacti.
This was a real blast from the past for me: Cacti is one of the very first monitoring tools I ever learned about. (It was an obvious choice in an RRDtool shop.)
N.B. It might not be THAT cool but I could not miss the opportunity for wordplay.
🌐 Mindmapping in Markdown
Markmap = Markdown + Mindmap
I am rather fond of
markmap for turning Markdown into pretty mindmaps in the form of interactive HTML! You can see an example at markmap.brie.dev or fork my project to get started with your own mindmap. While you can bring existing Markdown to
markmap, you might need to tweak it a bit based on how
markmap decides what to render. Take a look at the frontmatter of the example for a few ways to modify the behavior of
markmap instead of the content of your Markdown.
I tooted about this.
💨 On tooting
Before writing this, I would roll my eyes in bemusement at the usage of
toot on Mastodon. I wondered:
The answer is pretty interesting, as it turns out:
It’s no longer officially called tooting. In the merged Change the nouns “toot” and “status” to “post” PR, the lead developer of Mastodon points to the very thread from 2016 where the whole tooting thing started: a dare from YouTuber and Twitch streamer Hbomberguy. The referenced Patreon is still active.
After reading the comments in the PR , I get it and don’t mind “toot” so much but it still takes a little getting used to. It does make me smile so I’ll take it.
🎒 Noteworthy Uses of ChatGPT: Academia
In this moment, a semi-regular section on this topic seems worthwhile. If nothing else, it presents the opportunity for an interesting time capsule.
According to the friend of someone on Reddit (😂), a history professor purportedly has students correct essays written by ChatGPT:
My friend is in university and taking a history class. The professor is using ChatGPT to write essays on the history topics and the students need to mark up its essays and point out where ChatGPT is wrong and correct it.
What a cool idea!
In My class required AI. Here’s what I’ve learned so far. and the associated paper New Modes of Learning Enabled by AI Chatbots: Three Methods and Assignments we find a more credible account of ChatGPT being embraced in the classroom.
We also have this story:
- Vanderbilt University Staff Apologized For Using AI To Write An Email To Students About The Shooting At Michigan State
A follow-up statement is available but given the
https://peabody.vanderbilt.edu/statement.php URL and the link rot that seems inevitable, it’s archived for future reference. (The original email is also available in archive form.)
I will turn to an interesting side question that this raised for me. At the end of the letter, we see:
(Paraphrase from OpenAI’s ChatGPT AI language model, personal communication, February 15, 2023).
This was published by a university, an institution unlikely to take the matter of a citation unseriously.
- Q: How does one cite ChatGPT (and friends)?
Unsurprisingly, a lot has been written on this topic!
Bond University notes that responses from ChatGPT are considered non-recoverable sources. Other examples of non-recoverable sources include personal correspondence. Writers are encouraged to treat responses from ChatGPT and the like as personal communication or correspondence. It seems very reasonable to find and use something that already exists for this purpose. This will be an interesting space to observe over time.
Beyond the details of how to cite ChatGPT, the Implications in academia section of the ChatGPT article on Wikipedia raises a few other interesting questions.
🔍 GraphQL Security - in brief
If we have worked together or you know me IRL, we have probably talked about GraphQL!
Compared to REST, GraphQL is the grassy road wanting for wear. A good ELI5 on the rough edges in GraphQL for pentesters, defenders and the curious can be found in the GraphQL - Security Overview and Testing Tips post. Observing the 2018 timestamp, I think that the age of the post is an advantage: a simpler description of GraphQL is provided and we can test whether the advice stood the test of time. Beyond that:
- OWASP’s Web Security Testing Guide has a section on testing GraphQL that may be of interest
- the HackTricks book has a section on GraphQL that’s great if you have a basic overview of what GraphQL is and want to get started now.
- If you are lucky, you will find that GraphQL introspection is enabled but clairvoyance offers a way to obtain a GraphQL API schema even if introspection is not enabled. (Introspection is a way of asking the GraphQL API to give you the schema. If the schema is public: this may not be super helpful but it could be worthwhile to compare the published schema with what introspection tells you. You just might find something interesting.)
However you get the schema: the
graphql-path-enum tool described in GraphQL path enumeration for better permission testing looks promising!
😼 Deploying the HTTP Cats to Kubernetes, Part 1
Last year, I launched the HTTP Status Cats. There’s more about it here but in short: it’s a Flask app that serves cute cat pictures to remind you what different HTTP status codes stand for. (Check out httpcat.us/404 or httpcat.us/502 to see what I mean.) I am extending it for use in a Kubernetes cluster simply for the exercise. (There’s no real need for these cats to be on a cluster except to make it cuter.)
I am not a cloud-native software engineer and this is just for fun and experimentation so I’m taking the “can I make this thing work?” approach.
First things first: I need to write some YAML that I can pass to
kubectl apply to have it tell the Kubernetes cluster to spin up the cats.
- Input: a Flask app | source
- Output: Kubernetes manifest
After some research, trial, error, trial and success, I decided on kompose to solve this problem. I wrote a bit more about using
kompose to generate a k8s manifest for a Flask app on my TIL (today I learned) site. Since
kompose works best with Docker Compose and I was just using
Dockerfile, I assembled this
# docker-compose.yml version: '3' services: httpmeow: build: '.' image: registry.gitlab.com/brie/httpmeow:latest ports: - '5000:8008'
git clone https://github.com/kubernetes/kompose.git cd kompose && docker build -t kompose:v0 . cd .. git clone https://gitlab.com/brie/httpmeow.git cd httpmeow docker run --rm -it -v $PWD:/opt kompose:v0 sh -c "cd /opt && kompose convert"
In part 2, we’ll look at different options for deployment and use that to guide what we do with the network policy and the service. (If we use something like Okteto, we’ll have to work within what’s available to us.)
❤️ From the Heart
In January, I wrote that moments of happiness should be savored when they are encountered. It is possible to create and foster conditions that are likely lead to moments of happiness. Other times, the moments of happiness appear by surprise and these unexpected moments of joy on tough days are a Giladean balm. (Is that the right demonym?)
On a very bad day, I found comfort, wonder and joy in this flower-filled lovely gentle introduction to Minecraft:
I was rather ignorant of Minecraft before this: I knew there were blocks, you could build worlds and different people ran Minecraft servers. That’s about it.
While the creator is a good friend, I happened upon the video organically and was really moved by the care and passion. Enjoy!
— Brie 🦄🌈