[ seeded red // blue cyber range ]
The range that fights back.
Red-team CTF, blue-team detection CTF, and live red-vs-blue — on one seeded engine. Every team gets a different, uncopyable, self-verified instance. They go to war live: red works the kill chain while blue detects and counters in real time. Then Hotwash writes the debrief.
// three products, one engine
One engine. Red, blue, and live red-vs-blue.
Most ranges do one. Hotwash runs all three on the same seeded core — same artifacts, same flags, same grader. Pick the fight; the engine doesn’t change.
[ red CTF ]
Seeded red-team
Per-team, uncopyable Jeopardy against sandboxed live targets. Recon to objective across the kill chain — no two boxes alike.
engine: seed → artifacts → HMAC flags
[ blue CTF ]
Blue-team detection
Write a real Sigma, YARA or Suricata-style rule; it’s graded server-side against the exact telemetry an attack produced. Catch the plant, no false positives.
engine: same seed → telemetry corpus
[ live red // blue ]
Live arena
Both cells on one board at once — red works the kill chain while blue detects, contains and keeps services up. The clock decides it.
engine: same seed → shared live network
// the integrity contract
Leaked flags are worthless. Broken challenges don’t ship.
Uncopyable flags
Every flag is derived from the team’s seed. Paste it into another team’s box and it’s just wrong — and the submission log points straight back to whoever leaked it.
flag = HMAC(seed, challenge) # different for every team
Never-broken challenges
Before any instance is served, an independent solver re-solves it headlessly. If it can’t recover the flag, the instance is regenerated or pulled. Nothing reaches a player unverified.
[verify] fresh-catch: solver recovered primary ✓
// the main event
Live attack & defend.
Red recons, exploits, harvests credentials, pivots, escalates and goes for the objective on a seeded network. Every move throws off telemetry.
A behaviour-analytics layer turns that telemetry into alerts. Blue works the queue and hits back — isolate a host, kill a session, disable an account, patch a service. Cut red’s active path and that’s a point for defense. Red adapts, or gets contained before the clock runs out.
// the headline differentiator
The live arena, on the projector.
Players use real tools on real infrastructure. Hotwash doesn’t score the commands they type — it scores what they actually achieve, checked server-side. Observers watch it unfold on the board — nothing staged, every point a fact the mediation plane can see.
// the rule
Gate outcomes, not keystrokes.
No points for an exploit that merely fires. Red scores only by recovering the seeded proof from the box it was planted on — and holding it. Blue scores by detecting, containing and keeping services answering. Running a tool proves nothing — surviving on the box, and evicting whoever is already on it, is the whole game.
Red scores on real compromise
Points exist only as seeded secrets you can’t read without genuine compromise — and capture means hold, not touch: a dwell timer makes you survive on the box while blue tries to evict.
✓ proof recovered from dc01 · dwell 90s held
Blue scores on real defense
Points for detecting the move, containing the active path, and keeping graded services available — minus the disruption cost of heavy-handed responses that take your own network down.
✓ lateral path cut · app01 isolated · SLA held
gate 01
Proof-gated scoring
Every point is a seeded secret unreadable without real compromise.
gate 02
Infra-enforced prerequisites
Segmentation and secret placement form a chain no tooling can skip.
gate 03
Capture = hold
A dwell requirement means red must survive eviction, not just land.
gate 04
Economy lives on blue
Loud activity trips detection; responses that cut your own services cost availability.
The mediation plane. Both cells act on one seeded network through a mediation layer. A data plane — a segmentation gateway every flow crosses, plus host sensors reading ground truth — feeds one correlated stream to both sides at once. A control-plane referee holds the prerequisite graph, dwell timers, proof grading and the uptime/disruption ledger over an immutable journal. The score is ground truth, not either side’s claim.
// the draft
Draft your loadout. Spend the prep window.
A match runs draft → prep → live → hotwash. Each cell spends a symmetric budget on real capabilities, then gets a timed prep window to deploy them — with the availability checker already running, so you can’t just turn everything off. Spend on an EDR agent, a deception grid, a zero-day or an evasion kit — every pick is a real trade, and nobody buys their way to all of it.
⚑ white cell Timed or ad-hoc injects hit mid-match — a phishing foothold that lands, a tip-off alert, a decommissioned host, an armed decoy. Both sides react. Deterministic when the scenario declares them.
// the debrief
Every arena ends in a facilitated hotwash.
The moment the clock stops, the engagement journal becomes an after-action report — a three-axis scorecard per side, a per-operator breakdown, and an owned, prioritized improvement plan. Replay the match frame by frame, see exactly where a detection was missed, and leave with fixes assigned to names — not a vague debrief.
// the slate
Dozens of engagements. Every sector.
Industry-accurate attack paths — correct components, ports and kill-chain ordering, all fictional by construction. Each is a seeded network with multiple crown jewels, multiple paths, decoys and real detection gaps. Red picks which prize to chase; blue works a limited budget and can’t lock the whole network down at once.
OPERATION IRONROOT
Active Directory
OPERATION HELMSMAN
Kubernetes escape
OPERATION CLEARWIRE
SWIFT / payments
OPERATION SUBSTATION
IT → OT crossover
OPERATION SKYFALL
Cloud takeover
OPERATION NIGHTFALL
Corporate breach
OPERATION ZERO HOUR
Ransomware race
THE QUIET HOUR
Insider threat
OPERATION TRUSTED BUILD
Supply chain
OPERATION CHARTBREAK
Healthcare PII
OPERATION KEYSTONE
MSP / SaaS pivot
OPERATION CROSSFIRE
Three crown jewels
IRON VAULT · TWIN PATHS
Pivots & branches
// run the room
One analyst, or a ranked cohort.
Spin up a single seat or a whole competition. Every team gets its own seeded, uncopyable variant of the same scenario, ranked live, and advanced prep→live in a single organizer action. Run twenty teams on one scenario and no two instances match — a leaked flag is worthless and the leaderboard stays honest.
Live leaderboard
Rank the whole cohort by offense or defense in real time. One organizer action advances every team from prep to live together — no per-team babysitting.
Persistence & replay
Matches survive restarts. Replay any finished engagement frame-by-frame on the attack map, and browse every match in a full history view.
// two ways to deploy
Browser consoles, or your own kit. Graded identically.
tier 01
Console
Zero player setup. Red, blue and observer consoles in the browser; the plane does the grading.
- Launch in minutes, nothing to install
- Built-in attack map for the room
- Ideal for classrooms and open events
✓ grades on the mediation plane
tier 02
Own-kit / digital twin
Teams work from their own tools against real, container- or VM-backed infrastructure — blue holds admin on the hosts and pulls a live telemetry feed into its own SIEM.
- Bring your real offensive and defensive stack
- Live telemetry into your SIEM
- Ideal for SOC / IR teams and MSSPs
✓ identical scoring to the console tier
// the name is the payoff
// questions
The short version.
What does “uncopyable” actually mean?
Each team’s flag is derived from its own seed, so a flag that’s correct for one team is wrong for every other team. Share it and it fails — and the server-side log ties the attempt back to the team it belongs to.
Is the live attack-and-defend real exploitation?
Both sides use real tools against real, sandboxed infrastructure, and Hotwash grades observed outcomes — a seeded proof you can’t read without genuine compromise, real detections, real eviction, real availability. Not scripted buttons. Everything is isolated per team and torn down after.
Do defenders just click our buttons?
No. In the own-kit tier blue works from its own tools against real infrastructure with a live telemetry feed into its own SIEM; the platform grades availability, flag retention, real detections and real eviction. The console is optional scaffolding.
Does it work offline / air-gapped?
Yes. Targets, consoles and scoring run air-gapped; challenge content ships as signed offline bundles, and nothing is pulled at play time.
How is a detection-engineering challenge graded?
You submit a Sigma, YARA or Suricata-style rule. It’s run server-side against the challenge’s planted malicious data and its benign noise; you pass by catching the plant with no false positives.
How is it fair across a whole cohort?
Every team gets a different seeded variant of the same scenario, and an independent solver verifies each instance before it’s served. Grading is server-side, so the answer key never reaches a player.
Can we author our own scenarios?
Yes — there’s a documented DSL and authoring guide, and the engine validates and verifies every scenario before it can be served.
// spin one up
Pick a side.
Train the attacker. Train the defender. Then put them in the same room — and let Hotwash write the debrief.