← Blog

Guides

How to Stop Teams Submitting Pre-Built Projects at a Hackathon

Abhinavtej ReddyNovember 9, 20267 min read

The most reliable way to stop pre-built projects is to define the rule in plain words, ask every team for evidence of when its work began, and judge in rounds so that recycled work cannot win on polish. Written rules, a timeline check and a scoring rubric each close a different gap. Use all three, not the rulebook alone.

Start with a rule teams can actually check

Most hackathon rules fail at the same point. They say "no pre-existing work" without saying what counts as pre-existing. MLH's guide recommends framing the rule around what was built or changed during the weekend, as opposed to what was merely used, and it says organisers should make the no-pre-event-work rule explicit [3]. That distinction matters, because teams legitimately bring starter kits, libraries and open-source code.

Devfolio shows how a clear list works. It allows starter kits, boilerplates, open-source libraries and specs written in advance. It bars submitting a modified pre-existing project and bars porting older work unless that work is pre-approved, which makes the team ineligible for prizes [1]. Hack the Valley X uses a similar split. Public libraries, APIs and open-source code are allowed. No prior project or pre-event work can be submitted as-is, and all work must happen during the hackathon weekend [4].

A rule like this should state four things: what teams may use, what they may not submit, when the clock starts, and what happens if a project breaks the rule. IceHacks disqualifies plagiarised projects and projects built before the event. It also requires disclosure of pre-existing code and significant AI-generated contributions [5]. Disclosure rules help because they turn a vague suspicion into a question that teams must answer in writing.

Be specific about the start time. Devfolio says teams start when the event begins unless organisers specify otherwise, so no team gets a head start [1]. If you run a 24-hour or 36-hour format, state the exact start and end times in the rulebook. Our guide to the 24-hour versus 36-hour format covers how length affects what teams can realistically build.

Ask for evidence at submission

A rule without evidence is hard to enforce. Devfolio asks for source code at submission, such as a Git repository or a Figma file, to confirm that the work did not exist before the event [1]. It also requires version control throughout the event and recommends live demos for finalists [1]. Asking for source code is the step that turns a claim into something you can check.

The repository is the strongest evidence you have. A commit history that starts on the first day, with steady activity through the weekend, is hard to fake. A single large commit timed before the event is the opposite. Devfolio flags submissions with a single commit or suspicious repo activity as possible disqualifications, unless the team verifies them [1]. The Hackathon Integrity Playbook, a vendor guide, lists pre-built projects committed before the event and disguised as new work among six cheating categories. It recommends Git-based timeline analysis [7].

MLH's standard rules require public code, ideally in a Git repository, and say code must stay public after the event for prize eligibility [10]. Public code also lets others look at the history after the event, which discourages teams from submitting old work.

If your event does not use Git, ask for equivalent evidence: dated design files, a changelog, or screenshots of work in progress at fixed checkpoints. You want proof that the build happened across the event, not a single upload at the end.

Use a filed intent document as a baseline

A public GitHub rules gist proposes a written intent document, filed before the event, that states the goal and the planned functionality [2]. When judges later see the finished project, they can compare it with what the team said it would build, and see how far the team got during the event.

This is cheap to run, which suits student hackathons. Ask each team to file a short statement before the start time giving its goal and planned features. Keep the document dated and locked. It does not prove a team is honest, but it makes a late swap to an old project much easier to spot.

Judge in rounds so recycled work scores lower

Rules stop the worst cases, but scoring decides who wins. Devfolio suggests judging originality. Hack the Valley X scores "Authenticity" on whether an idea stands apart from previously seen projects, which gives judges a direct lever against recycled work [1][4].

Rounds help here. A first round can check that the team's work matches its intent document and that the repository shows activity during the event. A second round, with live demos for shortlisted teams, lets judges ask how a feature works and why a design choice was made. IceHacks says participants may be asked to explain or demonstrate their projects to judges [5]. A team that built the project can usually answer those questions. A team that submitted an old project often cannot. For a small panel, our guide on how many judges a hackathon needs suggests three to five judges per project, plus standbys.

Unifesto Sprint is built around this flow. It has in-app submissions, rounds with weighted rubrics, judge panels, rooms with a help desk and a live leaderboard, and it is a Pro feature. Put the originality criterion in the rubric from the start, so teams see it before they build. Our guide on how to organise a college hackathon covers setting the rubric in advance. Our guide on how to write good problem statements explains how a specific problem pushes teams toward something new instead of a project they already have.

Our view: a filed intent document plus a round-by-round rubric with an originality criterion catches more recycled work than any single rule in the rulebook. Teams that have to explain their build in round two rarely risk a project they did not write.

Check timelines, then have a human review flags

Automated checks can help, but they should flag, not decide. One platform builder describes a scanner that flags repositories with one commit, or with over 90 per cent of commits before the event start. The flags go to a human reviewer rather than triggering automatic disqualification, because a one-commit repo can be legitimate [8]. Someone who built a small project in one sitting, or who committed a clean scaffold on the first morning, should not lose a prize to a rule that cannot see context.

MLH's guide also says organisers should be able to point to specific written rules when disqualifying cheaters [3]. That is the real reason to keep the rulebook specific. If a team is disqualified, the reviewer should be able to quote the clause, show the evidence and explain the decision.

Rulebooks also disagree, and it is worth knowing where. Hack at Home allows teams to use an idea they had before the event, and judges only features added during the event. It also says that open-sourcing code just to reuse it during the event goes against the spirit of the rules [9]. MLH's standard rules say teams can use an idea from before the event. MLH's digital-event rules say a project containing prior work cannot be submitted [10][3]. There is no single standard, so decide your position and state it.

Where this approach falls short

None of these checks is proof. A determined team can build a project in private, commit it on day one and present it as new. A review can miss this, and a timeline check can flag honest teams. The evidence here comes from event rulebooks and a vendor guide. It is not a controlled test of how often each method works. A Quora answer on the topic, about 4,200 days old, is weak evidence, though it suggests organisers mainly check for finished solutions for copyright and licence reasons [11].

Unifesto does not check repositories or timelines for you. Its in-app submissions and rounds help you collect and judge the work, but the originality check still depends on your rulebook and your judges.

FAQ

Can teams use open-source libraries and starter kits?

Yes, in most rulebooks. Devfolio allows starter kits, boilerplates, open-source libraries and specs written in advance, but bars submitting a modified pre-existing project [1]. Name what is allowed in your rules so teams do not have to guess.

Is a single-commit repository always cheating?

No. A flag is a reason to ask questions, not proof. One platform sends these flags to a human reviewer, because a one-commit repo can be legitimate [8].

Should I ask teams to file their plan before the event?

Yes. A dated intent document stating the goal and planned features lets judges compare the finished project with what the team said it would build [2].

What if a team says its idea existed before the event?

An idea can usually be reused, but a project cannot. Hack at Home allows a pre-event idea but judges only features added during the event [9]. Decide your rule and state it in the rulebook.

Sources

  1. Judging: https://guide.devfolio.co/docs/guide/participating-in-hackathons/judging-1
  2. General rules for the Hackathon · GitHub: https://gist.github.com/durandj/7880d4d63df1b4b98dbf
  3. Rules for Your Hackathon: https://guide.mlh.com/general-information/judging-and-submissions/rules-for-your-hackathon
  4. Hack the Valley X: Celebrating our 10th iteration of innovation, impact, and imagination as Hack the Valley X returns. Are you ready to shape the future? - Devpost: https://hack-the-valley-x.devpost.com/rules
  5. IceHacks: A week online for high schoolers to build something that actually runs, software, hardware, or the weird idea in between. - Devpost: https://icehacks.devpost.com/
  6. Hackathon Rules: https://www.lpu.in/events/hack-iot/hackathon-rules.php
  7. Hackathon Integrity Playbook: How to Prevent & Detect Cheating: https://hackhq.io/integrity
  8. We Built a Hackathon Platform Where Every Rule Lives at the Layer That Can't Be Bypassed - DEV Community: https://dev.to/maheshwarianmol/we-built-a-hackathon-platform-where-every-rule-lives-at-the-layer-that-cant-be-bypassed-4546
  9. Hack at Home: Hack at Home! The first MLH Summer League Hackathon - Devpost: https://hack-at-home.devpost.com/rules
  10. mlh-policies/standard-hackathon-rules.md at main · MLH/mlh-policies · GitHub: https://github.com/MLH/mlh-policies/blob/main/standard-hackathon-rules.md
  11. How do hackathon organizers ensure that teams haven't started their projects before the hackathon starts? - Quora: https://www.quora.com/How-do-hackathon-organizers-ensure-that-teams-havent-started-their-projects-before-the-hackathon-starts