SparkWord

<img src="https://raw.githubusercontent.com/SparkCry/SparkWord/refs/heads/main/assets/banner.png" style="zoom: 80%;" /> <p align="center"> SparkWord works by automatically monitoring chat, commands, books, signs, and player input to detect spam, abuse, or banned content in real time. It offers advanced word filtering, anti-spam protection, moderation tools, logs, and player mutes. </p><br> # 🔵 **REQUERIMENTS** - Java-21 - Paper/Purpur-1.21.x --- # 🔵 **FEATURES** - Configurable word blocking and filtering via external text-based dictionaries with hot-reload support. - Built-in persistence layer using SQLite by default, with optional MySQL or MariaDB backends for scalable environments. - Advanced anti-spam system covering IP-based checks, domain detection, message flooding, excessive capitalization, numeric abuse, character repetition, and repeated word patterns. - Injection prevention layer designed to mitigate malformed or malicious input affecting vulnerable plugins. - Detection and sanitization of dangerous Unicode sequences, including Zalgo and invisible character abuse. - Content limitations and validation applied to writable books to prevent exploits and abuse. - Command input writing limits to reduce spam, injection vectors, and abuse through command execution. - Automated mute mechanisms triggered by configurable violation thresholds across multiple detection systems. - Comprehensive moderation logs tracking mutes, warnings, and staff command usage for auditing purposes. - Suggestion and correction system to guide players toward acceptable message formats. - Preventive staff notifications with contextual information before escalation actions occur. - Unified filtering and anti-spam enforcement across chat, signs, anvils, books, and other player-input sources. --- # 🔵 **USAGE** - **Word Filtering:** The plugin does not rely on traditional regex-based filtering, which was intentionally avoided due to performance issues and server lag. Instead, it uses an Aho–Corasick pattern-matching engine optimized for high-throughput text analysis. Filtering is driven by plain-text wordlists categorized by purpose (Normal, Strong, WriteCommand), each providing a distinct level of enforcement. - Normal: Performs partial word blocking, matching standalone or contextual terms (e.g., blocking "cat" without affecting "category" or "education"). - Strong: Applies aggressive matching, blocking terms regardless of whether they appear as substrings within larger words. - WriteCommand: Designed to prevent the use of administrator or staff names within writable commands such as /rename, /tag, /nickname, and similar features. - **Anti-Spam:** This module is fully configurable via moderation.yml, allowing fine-grained control over thresholds, cooldowns, and enforcement behavior. - **Suggestions:** The suggestion command `sw-sg` allows players to submit word suggestions, feedback, or staff reports. All entries are stored persistently and can be reviewed using `sw list sg`. By default, suggestion logs are automatically purged every 3 days; this interval is configurable via `config.yml` under `suggestion.purge-hours`. - **commands.yml:** Provides centralized control for redefining or aliasing moderation commands exposed to helpers or regular users, enabling permission-based command abstraction. - **Notifications:** Staff notifications are delivered using hover-based messages to minimize chat clutter and maintain readability on high-population servers. --- <details> <summary>Commands</summary> ``````bash Arguments: - ID: Each suggestion or entry is identified by a unique ID. - Preset: Predefined warning or moderation presets for quick application. - Word: The word to be added, removed, or suggested. - List: - n: Normal list - s: Strong list - wc: WriteCommand list - sg: Suggestions list - Logs: - all: All available logs - b: Blocked words (only available when filter mode is disabled) - m: Applied mutes - w: Applied warnings Command | Arguments | Permission | Usage /sw info | none | sparkword.info | Display available commands and their usage based on assigned permissions. /sw list | List | sparkword.list | View word lists or pending suggestions. /sw accept | ID/List | sparkword.accept | Accept suggested words into a target list. /sw deny | ID/List | sparkword.deny | Deny or reject suggested words. /sw add | List/Word | sparkword.add | Add words to a specific filtering list. /sw remove | List/Word | sparkword.remove | Remove words from a specific filtering list. /sw logs | Logs | sparkword.logs | Display moderation logs captured by staff or console. /sw audit | Staff-Name | sparkword.audit | Show command execution logs for a specific staff member or all. /sw pu