cd /news/ai-tools/i-left-a-linux-security-guide-half-f… · home topics ai-tools article
[ARTICLE · art-16981] src=dev.to pub= topic=ai-tools verified=true sentiment=↑ positive

I left a Linux security guide half-finished for months — Copilot helped me wrap it up

A teacher at SENAI in Taguatinga, Brazil, used GitHub Copilot to complete a long-stalled Linux security guide that had sat with half-finished modules and labs lacking terminal output examples. The developer integrated Copilot to generate a new lab on suspicious process detection, an automated hardening script, and expected command outputs for existing labs — though manual review was required to catch omissions like missing `ChallengeResponseAuthentication` in SSH hardening. The project now includes a functional hardening script and labs with visual feedback for students.

read3 min publishedMay 28, 2026

I'm a teacher at SENAI in Taguatinga, Brazil. I teach Systems Development and Fullstack. A few months ago I created the linux-security-guide — a practical Linux guide focused on security, born directly from my classroom.

The reason I started it was simple and honest: I was tired of 2-hour YouTube videos teaching things wrong, and course materials that just make students memorize commands. I had about 15 students fail job interviews because they couldn't answer "how do you secure a Linux server in practice?" — and that made me angry. So I put together everything I teach in class and use day-to-day so nobody would go through the same frustration I went through when I started.

The project grew well. Fundamentals, networking, hardening, practical labs. But at some point it stopped. It just sat there with a nice structure, half-finished modules, and labs that didn't show students what they should actually see in the terminal.

This challenge was the push I needed.

Three fronts:

1. Lab 04 — Suspicious process detection

This lab didn't exist. I knew what I wanted: something practical, with real commands (ps

, top

, lsof

, netstat

, /proc

), defensive focus, and expected output for each step. I gave Copilot the context and it put together a solid structure I wouldn't have written that quickly on my own. I still had to review every block — the technical content was correct, but some sections needed the adjustment of someone who's actually worked with this in a classroom.

2. Automated hardening script

Module 03 of the guide teaches secure SSH, auditd, suspicious crontab. But there was no script to automate all of that. Copilot generated a hardening.sh with dependency checking, automatic backup of files it modifies, logging everything to /var/log/hardening.log

, and decent error handling. Good enough to show students as a production script example.

3. Expected output in existing labs

The three existing labs had the commands but didn't show what students should see after running them. That sounds like a detail, but it isn't — beginner students get stuck exactly because they don't know if what appeared in the terminal is right or wrong. I used Copilot to review each lab and add those output examples. Repetitive work I always kept putting off.

While working on the suspicious crontab module (03-hardening/crontab-suspeito.md ), I was listing the basic commands to check scheduled tasks. Standard stuff.

Copilot suggested creating a lab where the student finds a cron job running a hidden script at /var/tmp/.cache/.systemd-fix.sh

and has to trace where it came from.

I laughed out loud. Because that's exactly the kind of dirty trick an attacker would use in the real world. A name that looks legitimate, a hidden directory inside /var/tmp

, a path that slips past a quick listing. I wouldn't have thought of that level of "dirty detail" on my own — at least not at that moment. It was smart because it helped me turn a boring "look at the crontab" lesson into a treasure hunt that students loved.

Copilot is useful, but stubborn like a student who thinks they know everything.

In the SSH hardening lab (lab-03-ssh-hardening.md ), it generated a correct command to disable password authentication — it put PasswordAuthentication no

just fine. But it forgot to mention ChallengeResponseAuthentication

. If a student followed only that, the server would still accept some older authentication methods. A guaranteed vulnerability.

I had to add three more configuration lines and an explicit warning in the text. Copilot helps, but it doesn't replace someone who's already been burned trying to fix a server at 3am.

Before: organized structure, incomplete content, labs with no visual feedback for whoever runs them.

After: new lab on process detection, functional hardening script, and existing labs with expected output at each step.

The guide still isn't perfect. But now it's in a state where I can recommend it to a student without having to apologize for any unfinished part.

github.com/MMVonnSeek/linux-security-guide Created and maintained by Professor Max — SENAI Taguatinga, Brazil.

── more in #ai-tools 4 stories · sorted by recency
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/i-left-a-linux-secur…] indexed:0 read:3min 2026-05-28 ·