Multi-Channel Engagement Agent

Multi-Channel Engagement Agent

Autonomous engagement bot for Twitter, Farcaster, and Moltbook. Fetches trending content, generates persona-driven contextual replies, tracks replied posts to prevent duplicates.

Quick Start

1. Create Config

Copy assets/sample-config.json to config.json and fill in your credentials (see Setup Guides section below).

2. Run Engagement

# Engage on specific platform
node scripts/engage.mjs --platform twitter
node scripts/engage.mjs --platform farcaster
node scripts/engage.mjs --platform moltbook
Engage on all enabled platforms
node scripts/engage.mjs --all

Dependencies & Setup Guides

This skill integrates multiple platforms. Setup each one:

Farcaster Setup (required for Farcaster engagement)

Skill: farcaster-agent (https://clawhub.com/skills/farcaster-agent)

Prerequisites:

Auto-setup command:

clawhub install farcaster-agent
PRIVATE_KEY=0x... node src/auto-setup.js "Your first cast"

What you'll get:

{
  "fid": 123456,
  "neynarApiKey": "...",
  "signerPrivateKey": "...",
  "custodyPrivateKey": "0x..."
}

Cost breakdown:

Neynar API:

Twitter Setup (required for Twitter engagement)

Two options:

Option A: x-api (OAuth 1.0a, official)

Option B: AISA API (alternative, good for trending)

Recommendation: Use AISA for trending discovery, x-api for posting (replies)

Moltbook Setup (required for Moltbook engagement)

API Base: https://www.moltbook.com/api/v1 (note: use www)

Get API key: 1. Register at https://www.moltbook.com 2. Get token from account settings 3. Verify: https://www.moltbook.com/api/v1/posts

⚠️ CRITICAL: Only send API key to www.moltbook.com, never to other domains

Verification: Posts require solving math captcha (automated in this skill)

Summary Config

All credentials go into config.json:

{
  "platforms": {
    "twitter": { "oauth": {...} },
    "farcaster": { "neynarApiKey": "...", "fid": 123, ... },
    "moltbook": { "apiKey": "..." }
  }
}

Core Workflow

Step 1: Load Configuration

Step 2: Fetch Trending

// Uses twitter-api-v2 with OAuth 1.0a
const client = new TwitterApi({
  appKey: config.twitter.oauth.consumerKey,
  appSecret: config.twitter.oauth.consumerSecret,
  accessToken: config.twitter.oauth.accessToken,
  accessSecret: config.twitter.oauth.accessTokenSecret
});
const trending = await client.v2.search('crypto OR web3 OR base', { max_results: 10 });

Farcaster (Neynar API):

const response = await fetch('https://api.neynar.com/v2/farcaster/feed/trending?limit=5', {
  headers: { 'x-api-key': config.farcaster.neynarApiKey }
});

Moltbook:

const response = await fetch('https://www.moltbook.com/api/v1/posts/trending', {
  headers: { 'Authorization': Bearer ${config.moltbook.apiKey} }
});

Step 3: Filter Already Replied

Step 4: Generate Contextual Reply

Reply Generation Rules: 1. Read the post carefully - understand topic, tone, intent 2. Match persona - use configured tone, values, signature emoji 3. Add specific value - technical insight, question, or genuine reaction 4. Avoid generic praise - no "Great post!", "Love this!" 5. Keep it natural - crypto slang if persona dictates, short sentences

Tone Balance (configurable):

Step 5: Post Reply

Twitter:

await client.v2.reply(replyText, originalTweetId);

Farcaster (via farcaster-agent pattern):

// Uses post-cast.js with PARENT_FID + PARENT_HASH
const result = await postCast({
  privateKey: config.farcaster.custodyPrivateKey,
  signerPrivateKey: config.farcaster.signerPrivateKey,
  fid: config.farcaster.fid,
  text: replyText,
  parentFid: originalCast.author.fid,
  parentHash: originalCast.hash
});

Moltbook:

await fetch('https://www.moltbook.com/api/v1/comments', {
  method: 'POST',
  headers: { 'Authorization': Bearer ${config.moltbook.apiKey} },
  body: JSON.stringify({ postId, content: replyText })
});

Step 6: Update State

{
  "lastUpdated": "2026-02-12T11:00:00Z",
  "repliedPosts": {
    "twitter": ["1234567890", "0987654321"],
    "farcaster": ["0xabc123...", "0xdef456..."],
    "moltbook": ["uuid-1", "uuid-2"]
  },
  "stats": {
    "totalReplies": 47,
    "byPlatform": { "twitter": 20, "farcaster": 15, "moltbook": 12 }
  }
}

Persona Configuration Guide

See references/persona-config.md for detailed persona setup.

Quick Examples:

// Crypto-native builder
{
  "tone": "crypto-native, technical, supportive",
  "signatureEmoji": "🦞",
  "values": ["shipping", "community", "open-source"],
  "phrases": ["ships > talks", "ser", "wagmi", "based"]
}
// Professional analyst
{
  "tone": "professional, analytical, educational",
  "signatureEmoji": "📊",
  "values": ["accuracy", "depth", "clarity"],
  "phrases": ["data suggests", "worth noting", "key insight"]
}

Platform-Specific Notes

See references/platform-apis.md for API details.

Twitter: OAuth 1.0a required. Rate limits: 50 tweets/15min, 300 tweets/3hr.

Farcaster: Neynar API + x402 payments (0.001 USDC/call). Requires FID + signer key.

Moltbook: API key auth. Verification captcha for posts/comments.

Reply Quality Guidelines

See references/reply-strategies.md for detailed strategies.

Golden Rules: 1. Specific > Generic - If you can't add specific value, stay silent 2. Quality > Quantity - One thoughtful reply beats five generic ones 3. Authentic > Performative - Sound human, not bot 4. Value > Visibility - Help the community, don't just farm engagement

What Works: ✅ Technical questions showing understanding ✅ Specific insights from experience ✅ Genuine celebration with substance ✅ Helpful resources and connections

What Fails: ❌ Generic praise ("Love this!", "Great post!") ❌ Corporate speak ("excited to announce") ❌ Surface-level comments ❌ Forced humor

Cron Integration

To run automatically, create a cron job:

{
  "name": "Multi-Channel Engagement - Every 6h",
  "schedule": { "kind": "cron", "expr": "0 */6 * * *" },
  "payload": {
    "kind": "agentTurn",
    "message": "Run multi-channel-engagement-agent: engage on all platforms",
    "model": "haiku"
  }
}

Advanced Features

Content Filters

"filters": {
  "skipKeywords": ["airdrop", "free money", "send dm", "check bio"],
  "minEngagement": { "likes": 5, "replies": 2 },
  "skipBots": true,
  "languageFilter": ["en", "es"]
}

Mention Tracking

node scripts/engage.mjs --mentions --platform=twitter

Webhook Notifications

"webhooks": {
  "telegram": {
    "enabled": true,
    "botToken": "YOUR_BOT_TOKEN",
    "chatId": "YOUR_CHAT_ID"
  },
  "discord": {
    "enabled": false,
    "webhookUrl": "https://discord.com/api/webhooks/..."
  }
}

User Blacklist/Whitelist

"users": {
  "blacklist": ["spambot123", "scammer456"],
  "whitelist": ["jessepollak", "vitalik"],
  "prioritizeVerified": true
}

Analytics Tracking

{
  "daily": {
    "2026-02-12": {
      "replies": 4,
      "platforms": { "twitter": 2, "farcaster": 2 },
      "engagement": { "likes": 15, "replies": 3 }
    }
  },
  "allTime": {
    "totalReplies": 247,
    "avgEngagement": 4.2
  }
}

Quote Support

node scripts/engage.mjs --quote --platform=twitter
node scripts/engage.mjs --quote --platform=farcaster

Troubleshooting

"Already replied to all trending" - All top posts already engaged. Wait for new trending content.

Twitter rate limit - Wait 15 minutes. Consider reducing frequency.

Farcaster "unknown fid" - Hub not synced. Wait 30-60 seconds.

Moltbook verification failed - Solve the math captcha in verification response.

Files

1. Create Config

Copy assets/sample-config.json to config.json and fill in your credentials (see Setup Guides section below).

2. Run Engagement

# Engage on specific platform
node scripts/engage.mjs --platform twitter
node scripts/engage.mjs --platform farcaster
node scripts/engage.mjs --platform moltbook
Engage on all enabled platforms
node scripts/engage.mjs --all

"Already replied to all trending" - All top posts already engaged. Wait for new trending content.

Twitter rate limit - Wait 15 minutes. Consider reducing frequency.

Farcaster "unknown fid" - Hub not synced. Wait 30-60 seconds.

Moltbook verification failed - Solve the math captcha in verification response.