ВЕРСИИ (4)

Added: - Private messages and other chat commands (/msg, /me, /mail and more) now go through the same filters as normal chat - Muted players can no longer talk through private messages or other chat commands - Sign, book and anvil rename filtering - Chat slowmode, a global chat lock, and a first-join chat delay that stops bot raids - Raid detection for the same message arriving from many accounts at once - Mutes and unmutes now work on offline players - Violation points, warnings and history are now kept across restarts - Per-filter bypass permissions, so you can exempt someone from one filter instead of all of them - Mute-exempt and mute-override permissions so staff cannot mute each other by accident - New subcommands: check, clear, mutelist, slowmode, lock and unlock - Tab completion on every command - Built-in update checker that tells you when a new version is out - Optional bStats metrics - Discord alerts for manual mutes, unmutes and warnings - The mute reason is now shown to the muted player - Sexual content is now detected and reported separately from slurs - Automatic log file rotation Fixed: - Words hidden with symbols no longer pass straight through the filter - Letter-spaced words are now caught - Words with trailing punctuation could slip past detection - Staff alert sounds threw a console error on 1.21.3 and newer, and no sound played at all - Players with the bypass permission could talk while muted - Muted players could talk freely in any world where filtering was disabled - Messages could be delivered uncensored while the player was told they had been censored - Shadow-flagged messages could still be relayed to Discord by chat bridge plugins - Ordinary chat was flagged as advertising, such as "that was fun.me and you" or "see you.tv" - Custom blocked domains were still enforced after URL blocking was turned off - A whitelisted word could be used to smuggle a blocked word past the filter beside it - The word "crackers" was flagged as a slur - A blank entry in a blocked phrase list silently blocked every message on the server - A custom phrase like "nega" also matched "negative"; plain words now match as whole words only - Reloading could throw an error if someone was chatting at that moment - Certain censor characters threw an error and dropped the message - All active mutes were silently wiped on reload when mute saving was turned off - Ampersands in messages were mangled, turning "Fish & Chips" into broken text - The flood window was capped at 10 seconds no matter what you configured - Duplicate message detection compared messages inconsistently and under-detected repeats - A single profane word could count as two or three against the per-message limit - The configurable no-permission message was never actually shown to players - The statistics permission was documented but never checked - Testing a message polluted the tester's own spam tracking and could trip their own filter - An @everyone inside a flagged message pinged the entire Discord server - Mute settings were being read but were missing from the config file entirely - A mute duration without a unit silently became a permanent mute Improved: - Now runs on Java 17, restoring support for 1.20.x servers that run Java 17 - PlaceholderAPI support expanded to 15 placeholders - Discord alerts are queued and rate-limit aware instead of one request per message - The violation log is buffered and rotated instead of being rewritten on every message - Settings are cached instead of being re-read for every single message - Advertising detection now sees through spelled-out addresses like "example (dot) com" - History and message testing now work for offline players and from the console - Unmute tab completion suggests players who are actually muted - Staff alerts can be mirrored to console, and the quoted message length is configurable - Config file rewritten with clearer comments and no leftover example values - Existing config files keep working after updating, since every new option has a default - Full documentation refresh

DiscordSRV leak — fixed. SHADOW action now cancels the AsyncPlayerChatEvent (after manually delivering the formatted message to just the sender, preserving the shadow-ban illusion). Since DiscordSRV skips cancelled chat events by default, shadow-blocked and blocked messages will no longer reach a linked Discord channel. BLOCK was already safe since it cancels the event; CENSOR was already safe since ChatCop rewrites the message before DiscordSRV's listener runs.

## ChatCop v1.0.2 — Accuracy & Stability Update ### 🐛 Bug Fixes - **Fixed false flags on normal words.** Words like *night, nightmare, neighbor, knight, midnight, overnight, Nigeria, snigger* and similar are no longer mistakenly detected as slurs. Slur detection is now properly word-boundary anchored. - **Fixed whitelist bypass.** A whitelisted word now only exempts itself — it can no longer be used to sneak banned content through the rest of a message. - **Fixed /chatcop reload** not applying changes to punishment thresholds, point decay, or cooldowns. - **Mutes no longer lost on crash.** Mutes are now saved instantly when applied or removed (previously only on clean shutdown). - **Stats no longer lost on crash.** Statistics now auto-save every 5 minutes. - **Fixed /ccmute with an invalid duration** (e.g. *0m*) silently issuing a **permanent** mute — it now shows an error instead. - **Fixed Discord webhook alerts** occasionally failing on messages containing special/control characters. ### ⚙️ Improvements - The advertising filter's **block-ips** and **block-urls** options now actually work — you can toggle IP and URL/domain blocking independently. - Cleaned up the config: removed an unused/misleading option and corrected command usage text. - Fixed a memory leak in the profanity filter that occurred on reload, plus minor spam-detection performance improvements.

## ChatCop 1.0.0 - Initial Release ### Features - Spam filter (flood, duplicates, caps, repeated chars) - Advertising filter (IPs, URLs, Discord links) - Toxicity filter (slurs, hate speech, smart bypass detection) - Threat filter (death threats, dox, DDoS threats) - Profanity filter (censor mode) - Per-filter custom punishment commands - Discord webhook with rich embeds - PlaceholderAPI support - Staff in-game alerts - Persistent mute system - Full log file support