feat: add SelfbotGuard v1.0.0 — advanced selfbot detection & punishment

Detection heuristics:
  - Rich-embed abuse (user accounts can't send rich embeds)
  - Inhuman response speed (<300ms consistently)
  - Message timing precision analysis
  - 24/7 activity profiling
  - Automated command-response pattern matching
  - Rapid-fire burst detection

Configurable per-server:
  - Toggle on/off (default: off)
  - Actions: log, warn, mute, kick, ban
  - Sensitivity: low/medium/high + custom threshold
  - Log channel, exempt roles/users
  - Staff notifications, mute duration

Commands: [p]sbguard enable/disable/action/sensitivity/threshold/logchannel/scan/flagged/reset
Slash: /sbguard-status, /sbguard-scan

selfbotguard/__init__.py

@@ -0,0 +1,14 @@
+"""SelfbotGuard — Advanced selfbot detection & punishment for Red-DiscordBot."""
+
+from .selfbotguard import SelfbotGuard
+
+__red_end_user_data_statement__ = (
+    "This cog stores Discord user IDs and guild IDs for selfbot detection. "
+    "Per-user data includes message timestamps, response timing, activity windows, "
+    "and suspicion scores. Users may request deletion via the bot owner."
+)
+
+
+async def setup(bot):
+    cog = SelfbotGuard(bot)
+    await bot.add_cog(cog)

selfbotguard/info.json

@@ -0,0 +1,31 @@
+{
+    "author": [
+        "everestmcarthur"
+    ],
+    "install_msg": "Thanks for installing **SelfbotGuard v1.0.0** \u2014 advanced selfbot detection & punishment.\n\nGet started: `[p]sbguard enable` to turn it on, `[p]sbguard action <action>` to set punishment.\n\nDetection heuristics: rich-embed abuse, inhuman response speed, timing precision analysis, 24/7 activity profiling, and automated command-response pattern matching.\n\nCog is *disabled by default* \u2014 enable per-server.",
+    "name": "SelfbotGuard",
+    "disabled": false,
+    "short": "Advanced selfbot detection with configurable per-server punishments.",
+    "description": "SelfbotGuard v1.0.0 detects Discord selfbots using multiple heuristic layers: rich-embed abuse detection (user accounts cannot send rich embeds via the client), inhuman response-time analysis, message-timing precision scoring, 24/7 activity profiling, automated command-response pattern matching, and rapid-fire burst detection. Each heuristic contributes to a cumulative suspicion score per user. When the score exceeds a configurable threshold, the bot takes action: warn, mute/timeout, kick, ban, or log-only. Admins can configure sensitivity, exempt roles/users, set a dedicated log channel, and review flagged users. Cog is disabled by default per-server.",
+    "tags": [
+        "selfbot",
+        "detection",
+        "antibot",
+        "moderation",
+        "security",
+        "automod",
+        "protection",
+        "anticheat",
+        "heuristic",
+        "guard"
+    ],
+    "type": "COG",
+    "end_user_data_statement": "This cog stores Discord user IDs and guild IDs for selfbot detection. Per-user data includes message timestamps, response timing, activity windows, and suspicion scores. Data is used solely for detection heuristics and is periodically pruned.",
+    "min_bot_version": "3.5.0",
+    "min_python_version": [
+        3,
+        9,
+        0
+    ],
+    "requirements": []
+}