Geneva, on site
Insomni'hack 2017: a guide to the Geneva CTF and its teaser
Insomni'hack is a Swiss security conference with a jeopardy CTF attached. The 2017 edition ran as an online teaser in January and an on-site final in March; this guide covers format, results and task styles.
Insomni’hack is a security conference held in Geneva, Switzerland, organized by the Swiss security company SCRT. Its CTF is the part of the program that most outside teams know. The 2017 edition followed the pattern of earlier years: a qualifying-style online teaser a couple of months before the event, and a final played in a conference venue in the city.
Dates and formats
The online teaser ran from January 21 to January 22, 2017, and CTFtime lists it as a jeopardy event. Teams solved tasks from anywhere, submitted flags, and scored by difficulty. CTFtime records 339 teams for it. The on-site final was played from Friday, March 24, 2017, 17:00 UTC until Saturday, March 25, 03:00 UTC, which is a ten-hour game that stretches into the Geneva night. It is also jeopardy style, with a fixed list of tasks and points per solve, not an attack-defense setup with services to patch.
The final’s CTFtime page describes a free competition open to all, with 400 seats available, and challenge areas including pwning, web, hardware, crypto, reverse engineering, mobile and networking. The number of teams that appear in the results is 68, a much smaller field than the online events of the same period, because physical presence in Geneva is part of the format.
Published results
The top places on CTFtime for the final read as follows.
| Place | Team | CTF points |
|---|---|---|
| 1 | Dragon Sector | 97,400 |
| 2 | Tasteless | 89,700 |
| 3 | int3pids | 79,600 |
Dragon Sector, Tasteless and int3pids also took the first three places in the teaser, in the same order, with 2,900, 2,400 and 2,400 points respectively. The tied scores in the teaser were separated by the time of the last solve, which is the usual tiebreak in jeopardy scoreboards.
Task styles
The categories on the event page cover the usual jeopardy range, with hardware and mobile listed next to the software categories. A typical team splits by specialty: one or two people on binary exploitation and reverse engineering, one on web, one on crypto, and someone who picks up whatever is left, including mobile, network captures or hardware. Web tasks at this event commonly run behind a stock server with CGI or a small custom framework, where the work is reading how the application treats cookies, uploads and paths. Binary tasks commonly come as x86 or x86-64 ELF files with a remote service, and the first hour usually goes to triage with file, strings and a decompiler.
One task of the 2017 final was named Deep Experiments. It belonged to the web category and involved a small CGI-driven site with a login form and a file upload feature; published writeups describe it as solved by only two teams. Details of the solution are left out here on purpose, since the task is better read through the published writeups once a team has tried it.
What to expect as a new team
Ten hours is short. Teams that arrive without a working split of roles lose the first hour to setup. A few practical points apply to a final of this kind.
- Bring the tools that work offline: a decompiler, a debugger with a heap-aware plugin, Python with
pwntoolsandrequests, a proxy such as Burp Suite, and a local copy of reference documentation. Conference Wi-Fi in a room with hundreds of laptops is not a dependable resource. - Check the rules on flag format and submission limits as soon as the game opens. Jeopardy scoreboards at this level are decided by a handful of solves, so one wrong-flag penalty or a missed format detail can matter.
- Pick tasks by the size of the point drop. Easy tasks lose value as more teams solve them, so solving them early is worth more than the same solve later.
- Keep notes in a shared document. Half-finished work on a task is worth little if the person who started it needs to sleep.
For orientation on the data side of such tasks, binary data visualization shows how to recognize file types and compressed regions in raw bytes, which saves time when a task hands over an unlabeled blob. The CTF team page covers how teams organize this kind of work, and the Ninth writeup shows a typical forensics-style size check in practice.
Rating and weight
CTFtime converts final standings into rating points, and the 2017 final awarded 89.680 points to the winner, 63.715 to second place and 51.592 to third. The teaser awarded 40.000, 26.552 and 23.218 to its top three. A team that places fifth in the final, as KITCTF did with 70,700 points, still collects 41.516 rating points, which shows how much a single strong on-site result moves a team in the yearly ranking. The ten-hour window and the 400-seat limit explain why the scoreboard stays compact: a field of 68 teams produces gaps of several thousand points between neighboring places near the top, so the order is settled by the hardest tasks, not by the easy ones every team solves.
The winner’s margin over second place was 7,700 points in the final and 500 points in the teaser, so the on-site game separated the leaders more clearly than the online one did.
Reading other editions
CTFtime keeps results and writeup links for every edition, and the organizers publish their own summary of the contest after each year. Comparing the teaser results with the final results for one year, as above, shows how much the on-site field narrows: 339 teams online against 68 in the room, while the same leaders appear at the top of both lists.
Questions readers ask
How was Insomni'hack 2017 split into two contests?
An online teaser ran on January 21-22, 2017 as a jeopardy game open to anyone. According to CTFtime it drew 339 teams. The on-site final took place in Geneva on March 24-25, 2017, with 68 teams listed in the results.
Who finished on top at the 2017 final?
CTFtime lists Dragon Sector first with 97,400 points, Tasteless second with 89,700 and int3pids third with 79,600. The same three teams also took the top three places in the online teaser, in the same order.
Is it a good first CTF for a new team?
The final is on-site, with limited seats, and the field consists of experienced teams. A new team gets more from the online teaser format, or from reading published writeups of earlier editions first, and then comparing its own approach with them.
New to capture the flag? Formats, teams and a first-month plan in one place.
Read the CTF guide