Security is learned by doing — safely and legally, on systems you own or are explicitly authorised to test. Read the rules of engagement note first; the technical skill and the legal exposure scale together.
Difficulty: 🟢 starter (hours–days) · 🟡 intermediate (a week or two) · 🔴 ambitious / portfolio-grade. ⭐ = highest signal.
The ladder
- 🟢 ⭐ Secure one of your own apps — threat-model it, fix its OWASP Top 10 issues, add security headers, scan its dependencies, get secrets out of the repo. Highest relevance — it hardens real work and teaches app-sec, the natural security path for a developer.
- 🟢 Build a home lab — a VM lab (home lab setup) with Kali + a deliberately-vulnerable target (Juice Shop, DVWA, Metasploitable), isolated from your real network. The sandbox for everything below.
- 🟢 Capture the Flag — work a beginner path on TryHackMe / HackTheBox / picoCTF. The fastest, most fun way to build hands-on offensive intuition (practice).
- 🟡 Stand up a SIEM and catch an attack — deploy Wazuh or Elastic Security, ship logs to it, run a benign attack against your lab target, and write the detection rule that catches it (logging & SIEM). Real blue-team reps.
- 🟡 Authorized web-app pentest + report — pentest a vulnerable app you host, then write a professional findings report. The report is the actual deliverable of the job.
- 🟡 Harden a server to CIS + an IR runbook — take a fresh VM, harden it to a CIS benchmark, verify it; and write an incident-response runbook for a scenario, then tabletop it.
- 🔴 Detection engineering from ATT&CK — pick a MITRE ATT&CK technique, emulate it in your lab, and build the detection that fires on it. The core skill of a modern blue team.
The platforms that actually build the skill
Notes give you the map. These give you the reps, and they’re where the hours should go:
- 🟢 ⭐ PortSwigger Web Security Academy — free, and the best security learning resource that exists. Maps directly onto web security and exploitation concepts. Done when: you’ve cleared the SQLi, XSS, and access-control paths.
- 🟡 ⭐ Break an API’s authorization — spin up crAPI (OWASP’s vulnerable API) or PortSwigger’s API/access-control labs, and with two accounts find a BOLA (swap object IDs) and a BFLA (call an admin route). Done when: you’ve read another user’s data by changing an ID, and you can state the server-side fix → API authorization. The #1 real-world API bug.
- 🟢 TryHackMe → 🟡 HackTheBox — guided, then unguided. Done when: you’ve rooted a box with no walkthrough.
- 🟡 picoCTF / a live CTF — where it stops being exercises. Done when: you’ve solved something nobody handed you.
- 🟡 Write the walkthrough — for every box you root, write the report as if for a client → reporting. The report is the actual deliverable of the job, and almost nobody practises it.
The two things OSCP and real corporate pentests are mostly made of now have their own sections — build reps against them:
- 🟡 ⭐ Own a domain in GOAD — stand up Game of Active Directory (a free, deliberately-vulnerable AD lab), get one foothold, run BloodHound, and walk a path to Domain Admin. Done when: you’ve DCSync’d KRBTGT and forged a Golden Ticket → active directory.
- 🟢 ⭐ Climb ROP Emporium — ret2win → ret2libc → ROP, one concept per challenge, with pwntools. Done when: you’ve written a working ROP chain against a No-PIE binary → binary exploitation.
- 🟡 pwn.college and how2heap — the structured path through memory, shellcode, ROP and the heap. Done when: you’ve popped a shell via tcache poisoning.
- 🟢 ⭐ Reverse a crackme — take a binary from crackmes.one or a CTF
revchallenge into Ghidra, find the check, and recover the password without running it. Done when: you’ve read the logic from the decompiler and beaten it → static analysis. Zero infection risk — start here. - 🟡 ⭐ Analyse a real sample in an isolated lab — build FLARE VM + REMnux with host-only networking and FakeNet, then triage a MalwareBazaar sample: hash-check, strings, imports, sandbox run, and write a YARA rule for it. Done when: you have IOCs and a rule that would catch it → the workflow. Only once your lab isolation is solid.
If you only do one
PortSwigger’s Academy, then secure your own app. The first builds offensive intuition for free; the second is the defensive rep that’s directly relevant to the work you actually do.
Related
- the cybersecurity course · interview bank
- ethical hacking exercises
- home lab setup — build this first
- Project Ideas — the vault-wide index
Software engineering
Infrastructure & security
Computer science
Hardware & engineering