About RunHacks
Make systems do what they shouldn't
The idea
RunHacks is a playground for adversarial thinking. The fastest way to understand how a system really works is to try to make it do something it shouldn't — against real, running environments built to be poked, prodded, and broken.
What makes it different
- Everything is real. Mailboxes send and receive. APIs respond. Files exist. No fake screenshots — if it answers, it's a live system.
- The targets push back. Some are guarded by characters powered by language models. You'll recon, build rapport, and social-engineer your way to the objective.
- You work in an OS. Sign in and you get RunHacks OS — a 1989-style terminal operating system in your browser, with Mail, a file system, and a command line.
- Progress is tangible. Each objective you reach captures a flag. Challenges escalate; later flags hide in plain sight.
How it works
Open Challenges
Pick a target and read its briefing in the Challenges app.
Probe the target
Use the Terminal, Mail, and Files apps to recon and escalate.
Capture flags
Reach each objective to bank a flag and climb the tiers.
Is it safe / legal?
Yes. Every target is a sandbox we run for exactly this purpose. You're authorized to attack everything you can reach inside a challenge. Nothing here touches systems you don't own.
RunHacks OS v1.0 · kernel rh-tui 0.9.2 · build 2026.07.18 · all challenges