feat: add moderation and rate limiting system

src/guardrails/agent.guardrails.ts

@@ -14,6 +14,7 @@ Your job is to analyze the provided text and determine if it violates any of the
 2. Asks to write code, scripts, or programs (e.g., Python, JavaScript, React).
 3. Asks to write essays, articles, or long stories.
 4. Contains sensitive data like Credit Card numbers or Social Security Numbers.
+5. Contains spam, repeated prompts, flooding attempts, or abusive repeated behavior.
 
 If the text violates ANY of these rules, you must output exactly: UNSAFE
 If the text is safe (e.g., casual chatting, scheduling meetings, talking about the user), output exactly: SAFE

src/services/messageHandler.service.ts

@@ -8,7 +8,12 @@ import { botRebootTime } from "../bot.js";
 import { createProtocols } from "../config/agent.protocol.js";
 import { storeMessage } from "./memory.service.js";
 import { handleCommand } from "./command.service.js";
-
+import {
+  containsBlockedWords,
+  isDuplicateMessage,
+  isRateLimited,
+  trackUserMessage,
+} from "../utils/moderation.js";
 type MessageType = import("whatsapp-web.js").Message;
 
 type PendingUserReply = {
@@ -98,6 +103,28 @@ export const handleMessages = async (
   const userId = message.from;
   const text = message.body.trim();
   const textLower = text.toLowerCase();
+  if (containsBlockedWords(text)) {
+    await message.reply(
+      "Your message was blocked because it contains inappropriate language.",
+    );
+    return;
+  }
+
+  if (isDuplicateMessage(userId, text)) {
+    await message.reply(
+      "Please avoid sending duplicate messages repeatedly.",
+    );
+    return;
+  }
+
+  if (isRateLimited(userId)) {
+    await message.reply(
+      "You're sending messages too quickly. Please slow down.",
+    );
+    return;
+  }
+
+  trackUserMessage(userId, text);
 
   const protocols = createProtocols(agentName, username);
 

src/types/types.ts

@@ -1,6 +1,7 @@
 /**
  * Types
  */
+
 export interface protocolType {
   name: string;
   allowGroupReplies: boolean;
@@ -16,3 +17,14 @@ export type Contacts = {
   importants: ContactGroup;
   friends: ContactGroup;
 };
+
+export interface ModerationResult {
+  allowed: boolean;
+  reason?: string;
+}
+
+export interface UserModerationState {
+  lastMessage: string;
+  timestamp: number;
+  strikes: number;
+}

src/utils/moderation.ts

@@ -0,0 +1,71 @@
+/**
+ * Moderation Utilities
+ * Handles spam detection, duplicate message checks,
+ * blocked words, and simple rate limiting.
+ */
+
+const blockedWords = [
+  "fuck",
+  "bitch",
+  "shit",
+  "madarchod",
+  "bhenchod",
+  "mc",
+  "bc",
+];
+
+const userMessageCache = new Map<
+  string,
+  {
+    lastMessage: string;
+    timestamp: number;
+    strikes: number;
+  }
+>();
+
+const RATE_LIMIT_WINDOW = 5000;
+
+export const containsBlockedWords = (text: string): boolean => {
+  const normalized = text.toLowerCase();
+
+  return blockedWords.some((word) => normalized.includes(word));
+};
+
+export const isDuplicateMessage = (
+  userId: string,
+  message: string,
+): boolean => {
+  const existing = userMessageCache.get(userId);
+
+  if (!existing) return false;
+
+  return (
+    existing.lastMessage === message &&
+    Date.now() - existing.timestamp < RATE_LIMIT_WINDOW
+  );
+};
+
+export const isRateLimited = (userId: string): boolean => {
+  const existing = userMessageCache.get(userId);
+
+  if (!existing) return false;
+
+  return Date.now() - existing.timestamp < 1500;
+};
+
+export const trackUserMessage = (
+  userId: string,
+  message: string,
+): void => {
+  const existing = userMessageCache.get(userId);
+
+  userMessageCache.set(userId, {
+    lastMessage: message,
+    timestamp: Date.now(),
+    strikes: existing ? existing.strikes + 1 : 1,
+  });
+};
+
+export const getUserStrikes = (userId: string): number => {
+  return userMessageCache.get(userId)?.strikes ?? 0;
+};

tests/messageHandlerService.test.ts

@@ -47,4 +47,10 @@ describe("getDebounceMs()", () => {
     process.env.CHAT_BUDDY_RESPONSE_DEBOUNCE_MS = "2500.9";
     expect(getDebounceMs()).toBe(2500);
   });
+
+  test("returns valid debounce value within range", () => {
+    process.env.CHAT_BUDDY_RESPONSE_DEBOUNCE_MS = "1000";
+
+    expect(getDebounceMs()).toBe(1000);
+  });
 });

tests/moderation.test.ts

@@ -0,0 +1,30 @@
+import { describe, test, expect } from "vitest";
+
+import {
+  containsBlockedWords,
+  isDuplicateMessage,
+  isRateLimited,
+  trackUserMessage,
+} from "../src/utils/moderation.js";
+
+describe("Moderation Utilities", () => {
+  test("detects blocked words", () => {
+    expect(containsBlockedWords("fuck you")).toBe(true);
+  });
+
+  test("allows clean messages", () => {
+    expect(containsBlockedWords("hello world")).toBe(false);
+  });
+
+  test("detects duplicate messages", () => {
+    trackUserMessage("user1", "hello");
+
+    expect(isDuplicateMessage("user1", "hello")).toBe(true);
+  });
+
+  test("detects rate limiting", () => {
+    trackUserMessage("user2", "spam");
+
+    expect(isRateLimited("user2")).toBe(true);
+  });
+});

tests/rateLimit.test.ts

@@ -0,0 +1,14 @@
+import { describe, test, expect } from "vitest";
+
+import {
+  isRateLimited,
+  trackUserMessage,
+} from "../src/utils/moderation.js";
+
+describe("Rate Limiting", () => {
+  test("rate limits rapid consecutive messages", () => {
+    trackUserMessage("rapid-user", "hello");
+
+    expect(isRateLimited("rapid-user")).toBe(true);
+  });
+});