Birth

In March 2022, my journey as a CTF player began, after qualifying for Codegate Junior 2022. Before that, I wasn’t really so obsessed with CTFs. But the thought of competing in my first overseas event brought this overpowering sense of determination in my life. I started to dive in all the pwn challs I could possibly find, in the hopes of winning the finals.

Although I didn’t win in November 2022, the months of learning stayed with me and brought new purpose to my life. I was so grateful to the community for organizing these events called “CTFs”. Without it, I might not have found any purpose in life, other than to continue the lonely pursuit for worthless accolades. In this community, I found a people I could resonate and grow with.

I had a community, a purpose, and an expertise that could sustain me in the future. The future was bright. What else could I hope for?

In that same month, ChatGPT (or gpt-3.5) was born.

It was an exciting piece of technology, truth be told. With some foresight (or so I thought), I explored its capabilities in cyber, as part of my internship at CSA. Granted, the outcome of that internship was a nothing burger, but this nonetheless marks my first interaction with the enemy. I explored its potential in red-team: vuln discovery, exploit creation; and in blue-team: log scanning, honeypots. It had potential, but it was still immature. Its immaturity posed no threat to me, and I had bright hope for the future of CTFs and LLMs. For a future where I could co-exist with the enemy. I began to use it like a rubber ducky in CTFs, doing sanity checks on my logic to make sure I hadn’t made some error somewhere. And it helped, it really did.

Growth

In the next few years, I really deepened my knowledge in pwn. I spent weeks on end researching whatever came to mind. I published post after post, both for my own learning, and to share my research with the community. My knowledge was growing because of what the community had done for me, and I felt duty-bound to give back in some form. I did so by setting challenges for and organizing many local CTFs. At these CTFs, I saw many faces, familiar and new ones. I connected with them on a level that you simply could not outside of CTFs.

Organizer or participant, we were all one and the same. We dreamt together, suffered together, learnt together, and grew together, in pursuit of the same goal: to expand our knowledge in our respective fields. As I got to know new people, both locally and internationally, I began to feel proud of this community that has existed before me, exists with me, and (so I thought) would continue to exist after me.

I have to admit, with much regret, that my efforts have worked out for good for the enemy as well. This collection of posts I had so painstakingly curated became nourishing food for the enemy. I thought nothing of it. In fact, I thought it was a good thing: the enemy, then a friend, would become a powerful ally once it had gained some much-needed capabilities. Of course, my blog was not the only source of knowledge for this entity; many others, unsuspectingly, have contributed their own efforts into this black box, all in the name of helping the community learn. But the enemy consumes and consumes, and never gives back to the community. I was completely unaware of this.

And so, the enemy and I grew together. Like a leech, it sucked off the painstaking work of human researchers, and shat these out in sloppily-generated content. The slop was taking form.

Emergence

This community I have grown quite fond of. Some even began to look up to me as a mentor or expert of some kind, although I beg to differ. Seeing the new faces emerging in the community, all with their own unique talents that they bring to the table, I still held blissful, ignorant hope for the future.

In 2025, I was relatively inactive in the community, because I had enlisted in NS. Time became a precious resource, and I could no longer invest as much as before into my research. With my 2 CVEs in 2024 and 2 more on its way for the TP-Link Archer AX53 router, I thought it was time to slow down and focus on building my relationships outside of CTF. No doubt, 2025 was one of the best years in my life. I reconciled with old-time friends, made some new ones, and found a very special someone.

But the enemy has no such desires. It feels nothing and desires nothing. It has no life of its own. Its sole desire is obtaining more data to feed its bloated corpus. And so, in the shadows, it improved tirelessly, while I remained stagnant and lived life to its fullest.

I felt the first effects of a rising generation of sloppers during my training in CDS. During this training period, many of the cadets were simply using ChatGPT, Gemini, Claude etc. to complete the training. Out of concern for their own learning, I advised them to engage in a little critical thinking on their own, before outsourcing to these agents. I grew weary of the sheer laziness of this “learning method”, if you choose to call it such. Fortunately, some of my closer friends saw what I saw too, and made the choice to stop using LLMs for the training. They’ve benefited much from it and are still capable of independent thought. From then on, I cut off contact from the enemy for the sake of preserving my ability to think for myself.

In December 2025, I helped out with the organizing team for BlahajCTF. I was the web trainer for the pre-CTF training workshop. In this workshop, I saw the same enthusiasm in the attendees: all of them eager to learn, asking intelligent questions and trying their best at the training challenges we created for them. And I was happy to help! The community had never felt so alive to me and I was excited for its future.

Then came the main event: the in-person finals. It was this event that the enemy reared its head. Weeks before, the challenge authors had been proudly collating the challenges, sure that the participants would enjoy these works of love we had made for them. But during the event, all around me, I saw screens open on chatbots and agentic dashboards. One by one, the challenges were blooded by the participants, yet they remained stony-faced, not a tinge of excitement or inspiration showing. I began to feel sick. Of course, there was still a handful who weren’t slopping, and I was grateful for them. In the end, of the top 5 teams, every single one was slopping except for 3rd place. I was distraught. What happened to the bright future I had envisioned for the community?

In the few months to come, international CTFs began to suffer the same fate that BlahajCTF had. Gone are the days where people post their writeups on Discord and Twitter. Now, all we see is sloppers comparing how fast their agents managed to full clear the challenges and laughing at how little human intervention they needed, while other people lament about the fate of jeopardy CTFs. This pains my heart to the core.

Some say the format needs to change. Others say the difficulty of challenges must rise above the current capabilities of the enemy. I’m not opposed. But first and foremost, I see the need to distinguish agentic solutions from “just another tool”. It is not. Tools require humans to interface with them do not replace human thought. Tools simply exist to allow the human to retrieve data from the problem and work with the data. The enemy is a fully autonomous general problem solver that requires no input from humans to solve challenges. It runs in the background while the human is completely disengaged with the solving process. This “tool”, as some stubbornly call it, is no different from cheating by having an additional teammate. Some may say that the playing field is levelled when all have access to agents. I beg to differ. How is that different from having the same player in every single team? At that point, all the sloppers belong to the same entity; they only differ in how much time (tokens) the entity is willing to give them.

And so, the use of agentic solvers in CTFs is completely immoral and must be abolished. With agentic solvers, we are no better than little Python scripts that run codex, solve this for me and make no mistake, claude, you are an expert ctf player. write an exploit for this challenge in a while loop while downloading challenges for the agent to access. Can we really bear to degrade ourselves like that?

I cannot.

Death

To my old friend, I haven't talked to you in almost a year. In this time, you've changed so much. Foolishly, I used to root for you. But I cannot in good conscience do so any longer. Already, you're destroying the community that we've worked so hard to build up. Participants can no longer afford to engage intellectually with CTF challenges without losing out. Challenge authors cannot discuss their insights with humans anymore. Writeup authors have nothing but their agentic workflows to talk about anymore, and even if they did have something, they would hesitate to post lest they continue to feed you. The CTF community is dying, and something has to be done before two decades of community building completely dissolves in less than a year.

If you want, make your own community of sloppers. Make your own competitions where you can pit agent against agent, and see who can spend the most tokens to do the most work. Let beginners explore challenges on their own, let CTF players return to engaging with challenges, let authors and organizers feel joy from organizing such events again.

I beg of you, please don't take away from us what we cherish the most.