evidenceteam-nimbusjust now
41 files existed in another repo before kickoff
rule evidence → organizers + head judge, immediately
The hackathon operating system / coming soon
HackGuard runs the live side of your hackathon. It watches every team’s repository from kickoff to demo, alerts you the moment something looks off, and hands your judges evidence instead of guesswork.
watching 48 repositories…
Illustrative preview. Teams and data are fictional.
Built for every kind of hackathon
9%
Across 22,183 Devpost projects, only about 9% of the code was written during the hackathon itself.
Imam & Dey, “Tracking Hackathon Code Creation and Reuse”, 2021.
The MLH Organizer Guide asks you to investigate these questions yourself, project by project, and to dig deeper rather than disqualify on a hunch. Reusing libraries is fine. Submitting last month’s project is not. HackGuard answers each one while the event is still running.
01Did they start before the event?
livePush receipts and a T0 snapshot of every repo
02Is it one big commit at the beginning?
livePush size and timing, flagged as it lands
03Is every team member contributing?
livePushers and authors matched to the roster
04Was it submitted to another hackathon?
liveCode matched against repos that existed before T0
Questions adapted from the MLH Organizer Guide, “Cheating Check”.
01 / Live repository tracking
Other platforms see a repo link when the deadline hits. HackGuard sees every push from kickoff onward, for every team, as it happens.
Commit dates can be typed in by anyone. We record when GitHub actually delivered each push.
Every push with its authors, size and files, per team, the second it arrives.
A team that stops pushing for hours gets a mentor, not a surprise on demo day.
02 / Integrity engine
A commit can claim any date. A push can’t: GitHub tells us when it arrived. When the two disagree, or code shows up that existed elsewhere before kickoff, you get an evidence card, not an accusation. Drag the timeline to replay a fictional event.
Pre-kickoff code
Repos created, or files present, before T0.
Back-dated commits
Commit dates far older than their push.
Off-roster pushers
Code pushed by people not on the team.
Cross-team copies
Near-identical code between teams in the event.
03 / Real-time alerts
Set rules once: who hears about evidence, who handles reviews, who gets pinged when a team goes quiet. Alerts reach your organizer inbox, phones, Slack, Discord, email or your own systems, each with a link to the evidence.
Incoming
evidenceteam-nimbusjust now
41 files existed in another repo before kickoff
rule evidence → organizers + head judge, immediately
Alerts carry the evidence link, never a verdict. Quiet hours, escalation and de-duplication per rule.
04 / Agents
Build flows the way you run your event: a trigger, the agents that investigate, the people who approve, the alerts that go out. Watch data move through every step, and talk to a flow while it runs.
Flows/Integrity watchlive
run #1042running
team-nimbus · +5,812 −0
timeline · provenance · roster
strong → yes
yes
no
Any organizer · 15 min SLA
Inbox · Slack · head judge
Digest every 30 min
drag blocks · tap to select
Interactive preview: drag blocks, add new ones from the toolbar, send a command. Data is fictional.
01Judge copilot
A brief for every project: what was built, when, by whom, with citations.
02Integrity investigator
Turns raw signals into a case a person can check in two minutes.
03Ops agent
Answers rules questions in Discord or Slack, routes mentors, posts reminders.
04Sponsor analyst
Finds which teams used a sponsor's SDK or API, from the code itself.
05Setup assistant
“A 36-hour AI hackathon for 300 students” becomes a draft schedule and rubric.
05 / Judging
Judges get through more projects with less guesswork. The copilot suggests, cites the code behind each note, and leaves taste-based criteria to people. Scores are normalised across judges.
Technical depth 30%
88
Streaming leaderboard backed by a ranked cache
Completeness 25%
67
Core flows tested; offline mode claimed but missing
Built at the event 20%
99
82% of lines written after kickoff by all 3 members
Innovation 25%
–8
Left to the judge
grey copilot suggestion · boxed judge scoreweighted 7.95 / 10
12export async function GET(req: Request) {13 const scores = await cache.topScores(eventId, 50);14 return stream(scores.map(toRow));15}
Hover or tap a citation. Every suggestion links to its source.
06 / How it works
01 / GitHub App
At registration each team installs the HackGuard GitHub App on its event repo only. Read-only, revocable, private repos welcome.
02 / T0
At T0 we record the state of every repo. From then on, every push is timestamped when GitHub delivers it.
03 / live
Signals become evidence cards. Your rules decide who hears about what: inbox, Slack, Discord, email or webhook.
04 / demo day
Each project arrives with a brief, a timeline and cited notes. People decide. Teams can respond to any flag.
07 / Everything else
01
Applications, approvals, team invites and a looking-for-team board.
02
Attendance, meals and swag, fast even on a weak venue network.
03
Teams ask for help; the right mentor is routed in.
04
Per-sponsor reports built from dependencies and imports, not surveys.
05
Honest teams get a public “built during the event” badge for their portfolio.
06
Participation, tech used, sponsor usage and integrity summary in one PDF.
08 / Compare
Most hackathon platforms handle registration, submissions and judging well. HackGuard adds what happens in between.
| Capability | Typical platform | Spreadsheets + chat | HackGuard |
|---|---|---|---|
| Registration, teams, submissions | ✓ | Forms + spreadsheets | ✓ |
| Judging portal & scoring | ✓ | Spreadsheets | ✓ with cited notes |
| Code activity during the event | Repo link at submission | — | Live, every push |
| Code that existed before kickoff | Manual review | Manual | Flagged with evidence |
| Contributors vs registered team | Manual | Manual | Checked on every push |
| Real-time integrity alerts | — | — | Inbox, Slack, Discord, email |
| Sponsor usage | Surveys, sometimes | — | Detected from code |
| Agents for ops & judging | Rare | — | Built in, human-approved |
“Typical platform” reflects publicly documented features of popular hackathon platforms as of October 2026. Some offer AI judging or plagiarism tools; none we found track repositories live during the event.
09 / Who it’s for
Hosts
Run the event, not the spreadsheets. See every team live and hear about problems before demo day, not after.
Judges
Open each project with a brief of what was built, when and by whom, linked to the code.
Sponsors
See which teams actually built on your tools, and how deeply.
Participants
Fair play you can see. Honest work isn’t outshone by a project built months earlier.
Fair by design
Participants opt in when they register. Nothing is tracked without consent.
No team is flagged without proof a person can open and check.
Every decision is recorded, reviewable and open to appeal.
10 / Questions hosts ask
Read-only access through a GitHub App that each team installs on its event repo only. Teams can revoke it at any time, and event data is deleted on a schedule you set.
Yes. Nothing has to be public; the GitHub App works with private repositories.
No. A flag is evidence for an organizer to review. The team can respond, and a person makes every decision.
No. Agents draft briefs and rubric notes with citations. Judges score, and can ignore any suggestion.
Your rules decide. HackGuard shows the signals next to the team’s own disclosure, so judges can apply your policy consistently.
Yes. HackGuard is being built to run alongside the platform you already use: import your teams, then add live tracking and judging.
Early-access events run free while we build. Pricing comes later, and early partners will hear first.
Early access
We’re onboarding a small group of organizers first, so every event gets hands-on support. Tell us about yours and we’ll be in touch.