47 AI agents as RED TEAM (adversary), BLUE TEAM (UK/NATO), WHITE CELL (governance). Wargame any country. Simulate any scenario. Auto-generate MOD-ready whitepapers. All Ed25519-signed. JSP 936 compliant.
Load any nation's tech stack from OSINT. China, Russia, Iran, North Korea. Real equipment. Real doctrine. Real terrain via Cesium.
Vary terrain, force mix, weather (live Met Office), ROE, alliance structure. Run overnight. Wake up to results.
3D globe with real elevation. Yorkshire, Taiwan Strait, Baltic, South China Sea. Any location on Earth.
Counter-drone swarm defense. 100+ drones. Reinforcement learning. Test EW, kinetic, cyber options.
| Scenario Type | Agents | Runtime | Output |
|---|---|---|---|
| China-Taiwan 2027 Drone Swarm Defense | 47 | 8 hours | Whitepaper + Cesium replay |
| UK Homeland Counter-Drone | 33 | 4 hours | Gap analysis + MCP sensor map |
| NATO Article 5 Baltic | 47 | 12 hours | Force posture recommendations |
| Cyber Defense Tabletop | 25 | 2 hours | Attack graph + mitigation plan |
Any country's tech stack. Palantir AIP, Anduril Lattice, Helsing Strike reverse-engineered. Generates realistic adversary behavior from OSINT doctrine.
UK/NATO doctrine. DEFONEOS stack: FreeTAKServer, Mava, PX4, Cesium. Test defenses against adversary AI. Find gaps before real conflict.
JSP 936 compliance. BFT Council votes on escalation. 30 crosswalks validated. Watchdog Certs on every run. Ed25519 chain of custody.
Agents can switch teams mid-simulation. Test "what if our AI was compromised?" Run blue-on-blue scenarios. True adversarial testing.
Auto-generated whitepapers from simulation results. Executive summary + charts + gap analysis + recommendations. Formatted for MOD submission.
Force ratio charts. Loss-exchange ratios. Sensor coverage maps. Timeline visualisation. All auto-generated from simulation data.
Every whitepaper is Ed25519-signed. Full chain of custody. Who ran it, when, on what data, with what results. Immutable audit trail.
Red-team against Palantir AIP, Anduril Lattice, Helsing Strike. Find their weak points. Document your competitive advantage.
| Check | Standard | Method |
|---|---|---|
| Rules of Engagement | UK ROE (JSP 398) | BFT Council vote before escalation |
| AI Safety | JSP 936 | 30 crosswalks validated per decision |
| Audit Trail | Article 14 (EU AI Act) | Ed25519 Sigil chain on every run |
| Adversarial Testing | OWASP ASI Top 10 | PyRIT + Caldera + Garak red-team |
| Supply Chain | CSOAI Watchdog Certs | Every open-source component verified |
| Transparency | C2PA 2.0 | Content provenance on all outputs |
5 tools: init_swarm, deploy_drones, run_engagement, get_status, generate_tactics. Mava + PX4 integration. 10/10 tests.
GitHub5 tools for 3D COP. Real-time entity streaming. TAK-compatible. 9/9 tests pass.
5 tools: audit_model, check_compliance, validate_supply_chain, generate_certificate, query_crosswalk. 14/14 tests.
PyPI5 tools: detect_drone, classify_threat, recommend_countermeasure, track_swarm, assess_vulnerability. 14/14 tests.