Aegis AntiCheat

2.6-alpha — Release (alpha)

Major objectives:

• Expand detection coverage and reduce false positives.
• Provide richer diagnostics for tuning and debugging.
• Keep enforcement conservative by default; require sustained confirmations before punitive actions.

What's new:

• New detections:
  • AimAssist (heuristic using low-variance yaw/pitch changes around hits).
  • VelocitySpoof (detects missing or altered knockback behavior).
  • VClip (suspicious vertical teleport-like movement).
  • AutoBlock and AutoTool (heuristic detection for automated blocking and automatic tool switching).
  • AntiKB (checks for missing expected knockback).
  • AutoArmor, AutoClicker, FastBreak, Nuker, Reach, KillAura, Timer, Fly, Speed, Spider, Jesus, Phase, Elytra, FastPlace, FastBow, FastConsume.
• Timer detection redesigned:
  • Uses per-interval mean & standard deviation across multiple samples.
  • Requires environment checks (no gliding, not in vehicle, low horizontal speed).
  • Requires confirmation streaks before marking a violation.
• Confirmation streak system:
  • Each check must be sustained for a configurable number of confirmations (streak.*) before triggering a violation.
  • Streaks decay over time to avoid stale state.
• Diagnostics:
  • /aegis info prints current score, flags, streaks and client brand to help tuning.
• Safety-first enforcement:
  • Raised flag/kick thresholds by default.
  • Kicks and bans require multiple flags; persistent kick-counts are stored in config.
• Logging & webhook:
  • violations.log for auditing; optional Discord webhook support for admin notifications.
• Robustness:
  • Null-safety, thread-safe collections, all kicks/bans run on the main thread.
  • Config-driven weights and thresholds for easy tuning.
• Recommended flow:
  • Deploy to staging with verbose logging for 24–72 hours.
  • Use /aegis info and violations.log to tune per-check streaks and weights.
  • Gradually lower thresholds if you confirm stable detection results.

Notes:

• This is an alpha release: additional tuning will be required for specific server setups, plugins, and playstyles.
• If you need packet-level precision (to reduce false positives further), integrating a protocol-level listener or hooking into a networking library is recommended (not included here to keep compatibility).