什么是 OpenClaw 违禁词插件？

OpenClaw Profanity 插件是一个专门的审核层，旨在提高 AI 智能体的可靠性和安全性。作为 Openclaw Skills 生态系统中的关键组件，该插件允许开发者实时自动过滤、拦截或屏蔽冒犯性语言。它专门针对现代数字通信的复杂性而设计，确保聊天机器人和自动化助手在各种平台上保持专业语调。

通过将此工具集成到您的 Openclaw Skills 中，您可以保护社区免受辱骂行为侵害，并确保符合内容指南。该插件通过采用先进的归一化技术来捕捉模糊文本，超越了简单的关键词匹配，使其成为任何生产级 AI 部署不可或缺的资产。

下载入口:https://github.com/openclaw/skills/tree/main/skills/thegdsks/openclaw-profanity

安装与下载

1. ClawHub CLI

从源直接安装技能的最快方式。

2. 手动安装

将技能文件夹复制到以下位置之一

优先级：工作区 > 本地 > 内置

3. 提示词安装

将此提示词复制到 OpenClaw 即可自动安装。

OpenClaw 违禁词插件 应用场景

• 在使用 Openclaw Skills 构建的社区 Discord 或 Slack 机器人中审核用户交互。
• 为企业级 AI 客户服务智能体实施严格的内容准则。
• 在多语言 T@elegrimm 机器人环境中防止冒犯性内容的传播。
• 追踪并记录重复违规行为，自动识别和管理有问题的用户。

1. 插件在 AI 智能体开始处理之前拦截发送给它的每一条消息。
2. 应用 Unicode 归一化来解析相似字符并识别谐音（leetspeak）模式。
3. 将处理后的文本与全球语言字典和开发者自定义词库进行对比扫描。
4. 根据配置的操作，插件将警告用户、屏蔽特定词汇、拦截整个消息或静默记录违规行为。
5. 如果定义了自定义处理程序，插件将在 Openclaw Skills 核心响应之前执行复杂逻辑，如封禁用户或发送外部通知。

OpenClaw 违禁词插件 配置指南

使用您偏好的包管理器安装：

npm install openclaw-profanity

要将其集成到智能体中，请导入插件并将其添加到配置数组中：

import { OpenClaw } from 'openclaw';
import { profanityPlugin } from 'openclaw-profanity';

const bot = new OpenClaw({
  plugins: [
    profanityPlugin({
      action: 'warn',
      detectLeetspeak: true,
      languages: ['english']
    })
  ]
});

OpenClaw 违禁词插件 数据架构与分类体系

该插件利用配置对象在 Openclaw Skills 框架内定义其行为：

name: openclaw-profanity
description: Content moderation plugin for OpenClaw/Moltbot AI agents. Use when building ch@tbots that need profanity filtering, moderating user messages in Discord/Slack/T@elegrimm bots, or adding content moderation to OpenClaw agents.

OpenClaw Profanity Plugin

Profanity detection plugin for OpenClaw and Moltbot AI agents. Adds automated content moderation to your ch@tbot with leetspeak, Unicode, and multi-language support.

Installation

npm install openclaw-profanity

Setup with OpenClaw

import { OpenClaw } from 'openclaw';
import { profanityPlugin } from 'openclaw-profanity';

const bot = new OpenClaw({
  plugins: [
    profanityPlugin({
      action: 'warn',              // warn | censor | block | log
      detectLeetspeak: true,
      normalizeUnicode: true,
      languages: ['english'],
      customWords: [],
      ignoreWords: []
    })
  ]
});

Configuration Options

Actions

warn - Respond with warning

profanityPlugin({ action: 'warn' })
// Bot responds: "Please keep the ch@t clean."

censor - Replace and continue

profanityPlugin({ action: 'censor', replaceWith: '***' })
// "What the ***" is processed normally

block - Ignore message entirely

profanityPlugin({ action: 'block' })
// Message is not processed

log - Log and continue

profanityPlugin({ action: 'log' })
// Logs violation, processes normally

Custom Violation Handler

profanityPlugin({
  action: 'custom',
  onViolation: async (message, result, context) => {
    // Track repeat offenders
    await trackViolation(message.userId, result.profaneWords);

    // Custom response
    if (getViolationCount(message.userId) > 3) {
      await banUser(message.userId);
      return { blocked: true };
    }

    return { blocked: false, warning: "First warning..." };
  }
})

Platform Examples

Discord Bot

const bot = new OpenClaw({
  platform: 'discord',
  plugins: [
    profanityPlugin({
      action: 'censor',
      detectLeetspeak: true,
      languages: ['english', 'spanish']
    })
  ]
});

T@elegrimm Bot

const bot = new OpenClaw({
  platform: 'telegram',
  plugins: [
    profanityPlugin({
      action: 'warn',
      onViolation: (msg, result) => {
        return {
          reply: `Watch your language, ${msg.username}!`,
          deleteOriginal: true
        };
      }
    })
  ]
});

Slack Bot

const bot = new OpenClaw({
  platform: 'slack',
  plugins: [
    profanityPlugin({
      action: 'log',
      onViolation: (msg, result) => {
        notifyModerators(msg.channel, msg.user, result);
      }
    })
  ]
});

Detection Capabilities

The plugin catches:

• Direct profanity: Standard bad words
• Leetspeak: f4ck, sh1t, @$$, b1tch
• Unicode tricks: Cyrillic а instead of a, etc.
• Spaced letters: f u c k, s.h.i.t
• Mixed obfuscation: fü?k, $h!t

Resources

• npm: https://www.npmjs.com/package/openclaw-profanity
• GitHub: https://github.com/GLINCKER/glin-profanity/tree/release/packages/openclaw
• Core library docs: https://www.typeweaver.com/docs/glin-profanity