tibet-anticheat 0.1.1

[DEPRECATED] Anti-cheat through provenance, not surveillance — every game action is a TIBET token, no kernel driver needed

tibet-anticheat

Anti-cheat through provenance, not surveillance — every game action is a TIBET token, no kernel driver needed.

Part of the TIBET protocol suite by Humotica AI Lab.

The Problem

Current anti-cheat is broken:

• Kernel drivers (EAC, Vanguard) — ring-0 access, security risk, privacy nightmare
• Signature scanning — always behind, cheat devs update faster
• Always-online DRM — punishes legitimate players
• Server-side only — can't work for P2P/indie games

The Solution

Every game action is a TIBET token. Math catches cheaters, not surveillance.

from tibet_anticheat import AntiCheatEngine, ActionType, Vec3

engine = AntiCheatEngine()
session = engine.start_session("player-42")

# Game server records actions (each = TIBET token)
session.record_action(ActionType.MOVE, position=Vec3(10, 0, 5))
session.record_action(ActionType.SHOOT, position=Vec3(10, 0, 5),
                      target_position=Vec3(50, 2, 30))

# End session — all detectors run automatically
report = engine.end_session(session.session_id)
print(report["verdict"])  # CLEAN / SUSPICIOUS / GUILTY

Install

pip install tibet-anticheat

What It Detects

Cheat	How TIBET Catches It
Aimbot	Aim snap analysis: instant corrections, zero jitter, perfect tracking
Speedhack	Position/time ratio exceeds game physics
Teleport	Position jumps with no movement chain
Triggerbot	Reaction times below human limits (<100ms consistently)
Macro	Unnatural timing regularity (coefficient of variation)
Item dupe	No TIBET token for how you acquired that item
Chain tamper	Action chain hash broken — someone modified the sequence

How It Works

Player Input → Game Action → TIBET Token → Chain → Analysis
                                ↓
                         ERIN: what happened
                         ERAAN: previous action (hash-linked)
                         EROMHEEN: input sample, tick, position
                         ERACHTER: intent

Human players have natural patterns:

• Aim has jitter (hand tremor), acceleration curves
• Movement has inertia and reaction delays
• Reaction time: 150-250ms (pros: 130-180ms)
• Accuracy varies with distance and stress

Cheaters break these patterns:

• Aimbot: instant corrections, zero jitter, constant speed
• Speedhack: impossible velocity for game physics
• Triggerbot: consistent sub-100ms reactions

Player Trust (FIR/A)

Players get a trust score that evolves over time:

• Fair Play (30%): clean sessions / total sessions
• Integrity (30%): action chain integrity
• Recency (15%): recent clean play recovers trust
• Anomaly (25%): severe cheats reduce trust faster

High-trust players → less scrutiny. Low-trust → more analysis. Zero-trust → auto-ban.

CLI

# Full demo with fair players + cheaters
tibet-anticheat demo

# Only fair players
tibet-anticheat demo --scenario fair

# Only cheaters
tibet-anticheat demo --scenario cheater

For Indie Developers

# That's it. No SDK integration. No kernel driver.
# No monthly fees. No account system. MIT licensed.

from tibet_anticheat import AntiCheatEngine

engine = AntiCheatEngine(max_speed=10.0, max_sprint_speed=18.0)

P2P Games (with tibet-mesh)

from tibet_mesh import MeshNode
from tibet_anticheat import AntiCheatEngine

# No central server — each node validates actions
node = MeshNode(device_id="game-host")
engine = AntiCheatEngine()

# Players connect via mesh, actions verified by all peers
# Cheater's action chain doesn't match → consensus rejects

Also Works As DRM

Every game copy can get a TIBET token at purchase. The provenance chain IS the proof of ownership. No always-online server needed.

IETF Drafts

• TIBET Provenance
• JIS Identity

License

MIT — Humotica AI Lab 2025-2026

---

Enterprise

For private hub hosting, SLA support, custom integrations, or compliance guidance:

Enterprise	[e-mail]
Support	[e-mail]
Security	[e-mail]

See ENTERPRISE.md for details.