ncreighton Overview
The AI Niche Community Discord/Slack Moderation and Growth Specialist is a production-ready automation system designed to manage, moderate, and grow engaged niche communities across Discord and Slack. This skill combines real-time moderation (spam detection, toxicity filtering, automated warnings), member intelligence (high-value member identification, retention scoring), and growth analytics (viral moment detection, sentiment-driven content recommendations, weekly health reports).
Why This Matters
Community managers spend 30-40% of their time on repetitive moderation tasks and manual member engagement tracking. This skill eliminates that overhead while providing data-driven insights to grow community quality and retention. It integrates with:
- Discord API — native bot integration for real-time message monitoring
- Slack API — workspace moderation and member analytics
- Google Perspective API — toxicity and spam classification
- OpenAI GPT-4 — contextual FAQ responses and sentiment analysis
- PostgreSQL/Firebase — member scoring and historical analytics
Quick Start
Try these prompts immediately to see the skill in action:
Example 1: Set Up Automated Moderation Rules
"Set up moderation for my Discord server. Flag messages with toxicity score >0.7,
auto-delete spam patterns (repeated emojis, links >5 per message), and send warnings
to members with >3 violations in 24 hours. Whitelist #announcements and #dev channels
from spam filters."
Example 2: Identify High-Value Members
"Analyze my Slack workspace and generate a member score for everyone based on:
message quality (not just frequency), helpful responses given to others,
participation in threads, and emoji reactions received. Rank top 20 members
and suggest who to feature in our weekly newsletter."
Example 3: Generate Weekly Community Health Report
"Create a weekly community health report for my Discord server showing:
total messages, new members joined, member retention rate, toxicity incidents,
top discussion topics by sentiment, and engagement trends. Include recommendations
for content topics based on conversation sentiment."
Example 4: Detect Viral Moments
"Monitor my Discord for viral moments: identify threads with >50 reactions,
messages that sparked high engagement, and sentiment shifts. When detected,
pin the message, notify community managers, and suggest follow-up content topics."
Example 5: Auto-Respond to FAQs
"Train the FAQ responder on our community guidelines, pricing, onboarding process,
and common technical questions. When members ask similar questions, auto-respond
with contextual answers and escalate complex issues to moderators."
Capabilities
1. Real-Time Moderation & Spam Detection
- Toxicity Filtering: Uses Google Perspective API to score messages for toxicity, profanity, and personal attacks (0-1 scale)
- Spam Detection: Identifies repeated links, excessive emojis, all-caps messages, and coordinated spam patterns
- Automated Warnings: Issues progressive warnings (1st: DM warning, 2nd: mute 1 hour, 3rd: escalate to mods)
- Custom Rules: Define channel-specific rules (e.g., stricter moderation in #general, relaxed in #off-topic)
- Whitelist/Blacklist: Exclude trusted members or channels from certain filters
Usage Example:
Configure moderation thresholds:
- Toxicity threshold: 0.7 (flag if score >0.7)
- Spam link limit: 3 links per message
- Caps lock threshold: 70% of message
- Repeated emoji limit: 5 consecutive
- Action: Auto-delete + warning on first offense
2. Member Intelligence & Retention Scoring
- Member Value Score: Composite metric based on:
- Message quality (sentiment, helpfulness, citations)
- Community contribution (answers given, threads started)
- Engagement consistency (days active, response time)
- Network influence (followers, reactions received)
- Retention Risk Scoring: Identifies members likely to churn (declining activity, negative sentiment)
- High-Value Member Alerts: Notifies admins when VIP members are at risk
- Cohort Analysis: Groups members by activity level and engagement type
Usage Example:
Member John_Dev scores:
- Message Quality: 8.5/10 (helpful, technical)
- Contribution: 9.2/10 (answered 47 questions, started 12 threads)
- Consistency: 7.8/10 (active 5/7 days)
- Influence: 8.9/10 (avg 12 reactions per message)
- Overall Score: 8.6/10 (VIP tier)
- Retention Risk: Low (stable engagement)
- Recommendation: Feature in monthly spotlight, invite to leadership council
3. Automated FAQ Responses
- Intent Recognition: Understands questions even with different wording
- Contextual Answers: Pulls from knowledge base and recent discussions
- Escalation Logic: Routes complex questions to human moderators
- Learning: Improves responses based on member feedback and mod corrections
- Multi-Language Support: Responds in member's detected language
Usage Example:
Member asks: "How do I get started with your product?"
Bot responds: "Welcome! Here's our onboarding guide: [link].
Key steps: (1) Create account, (2) Connect your first data source,
(3) Run your first analysis. Stuck? Reply with 'help' or ask in #support."
4. Viral Moment Detection & Content Recommendations
- Engagement Surge Detection: Identifies threads with unusual activity spikes
- Sentiment Trend Analysis: Detects when conversations shift positive/negative
- Topic Extraction: Pulls trending discussion topics from viral threads
- Content Recommendations: Suggests follow-up topics based on sentiment and engagement
- Viral Amplification: Auto-pins messages, notifies mods, suggests discussion starters
Usage Example:
Viral Moment Detected:
- Thread: "New feature announcement" (47 reactions, 23 replies in 2 hours)
- Sentiment: 92% positive
- Topics: Feature request, use cases, integration
- Action: Pinned message, notified #announcements
- Recommendation: Host a live demo session, create tutorial video
5. Weekly Community Health Reports
- Engagement Metrics: DAU, MAU, message volume, reply rates
- Member Lifecycle: New members, churned members, reactivated members
- Content Performance: Top threads, most-discussed topics, sentiment trends
- Moderation Summary: Violations, warnings issued, banned members
- Health Score: Overall community vitality (0-100 scale)
- Actionable Recommendations: Specific steps to improve retention and engagement
Usage Example:
WEEKLY COMMUNITY HEALTH REPORT (Jan 15-21)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
📊 Engagement
- Messages: 2,847 (+12% vs last week)
- Active Members: 287 (↑8)
- Avg Reply Time: 2.3 hours
- Thread Participation: 64% (↑3%)
👥 Member Lifecycle
- New Members: 34
- Churned (7d inactive): 8
- Reactivated: 12
- Health Score: 78/100 (↑2 points)
🔥 Top Topics (by engagement)
1. Feature requests (234 messages, 89% positive)
2. Use cases (156 messages, 94% positive)
3. Integrations (98 messages, 76% positive)
⚠️ Moderation
- Toxicity Flags: 3 (all handled)
- Spam Removed: 12 messages
- Warnings Issued: 2
- Bans: 0
💡 Recommendations
1. Host live demo on top feature request
2. Create integration guide for #3 trending topic
3. Spotlight 5 high-value members (retention)
4. Schedule AMAs with product team
Configuration
Environment Variables (Required)
# Discord Integration
DISCORD_BOT_TOKEN=your_discord_bot_token_here
DISCORD_GUILD_ID=your_server_id_here
# Slack Integration
SLACK_BOT_TOKEN=xoxb-your-slack-bot-token
SLACK_SIGNING_SECRET=your_slack_signing_secret
# AI & Moderation APIs
OPENAI_API_KEY=sk-your-openai-key
PERSPECTIVE_API_KEY=your-google-perspective-api-key
# Database (for member scoring & history)
DATABASE_URL=postgresql://user:password@localhost/community_db
# OR Firebase
FIREBASE_PROJECT_ID=your-firebase-project
FIREBASE_PRIVATE_KEY=your-firebase-key
# Optional: Analytics
MIXPANEL_TOKEN=your-mixpanel-token
Setup Instructions
Create Discord Bot (if using Discord):
- Go to Discord Developer Portal
- Create new application → Bot → Copy token
- Enable intents: Message Content, Guild Members
- Invite to server with permissions: Manage Messages, Moderate Members, Send Messages
Create Slack App (if using Slack):
- Go to Slack API Dashboard
- Create new app → Install to workspace
- Enable events: message.channels, member_joined_channel
- Set permissions: chat:write, users:read, conversations:read
Configure Moderation Rules:
moderation: toxicity_threshold: 0.7 spam_filters: - repeated_links: 3 - excessive_caps: 70% - emoji_spam: 5 auto_actions: first_violation: "dm_warning" second_violation: "mute_1h" third_violation: "escalate_to_mods" whitelisted_channels: ["#announcements", "#dev"]Train FAQ System:
- Upload knowledge base (markdown or PDF)
- Provide 10-20 example Q&A pairs
- Define escalation keywords for human review
Example Outputs
Moderation Alert
🚨 MODERATION ALERT
Member: @SpamBot_123
Message: "CHECK THIS OUT!!! FREE MONEY!!! CLICK LINK 👉 bit.ly/xxx [emoji spam]"
Toxicity Score: 0.82 (FLAGGED)
Spam Score: 0.91 (FLAGGED)
Action: Message deleted, member warned (1/3)
Reason: Toxicity (0.82 > 0.7 threshold) + Spam (links + emoji spam)
Member Spotlight Report
⭐ TOP 5 MEMBERS THIS WEEK
1. @Sarah_Dev (Score: 9.1/10)
- Answered 23 questions
- 156 reactions received
- Helped 18 unique members
2. @Alex_Designer (Score: 8.7/10)
- Started 8 high-quality threads
- 98% positive sentiment
- Mentored 3 new members
[... 3 more members ...]
🎯 Retention Risk Alert
@OldMember_42 (formerly Score: 8.9/10)
- Activity ↓ 78% (was 45 msg/week, now 10)
- Last message: 6 days ago
- Sentiment: Neutral → Negative
- Recommendation: Reach out personally, invite to special event
FAQ Response Example
User: "What's the pricing for enterprise?"
Bot: "Great question! Our Enterprise plan includes:
• Unlimited users & API calls
• Dedicated support (1h response time)
• Custom integrations
• SLA guarantee
Pricing starts at $5K/month (custom quotes available).
Schedule a demo: [calendly link]
Or chat with our sales team in #sales-inquiries"
Tips & Best Practices
1. Moderation Calibration
- Start with toxicity threshold at 0.75 (conservative), lower to 0.65 after 2 weeks of tuning
- Monitor false positives weekly; adjust spam filters based on actual violations
- Use "mute" before "ban" to give members a second chance
- Review escalated messages manually to improve AI accuracy
2. Member Engagement Strategy
- Spotlight top 5 members weekly in #announcements (increases retention by ~15%)
- Proactively reach out to members with retention risk scores >0.6
- Create "member tiers" (Contributor, Expert, Leader) based on scores
- Invite high-value members to leadership council or beta testing
3. Content Strategy
- Review viral moment recommendations daily; prioritize trending topics
- Host AMAs on top 3 trending discussion topics monthly
- Create tutorials/guides for frequently asked questions
- Use sentiment trends to pivot community focus (if negative, increase recognition events)
4. FAQ Training
- Start with 20-30 Q&A pairs; add 5-10 new ones weekly based on real questions
- Test FAQ bot in #test-channel before production deployment
- Review bot responses daily for first 2 weeks; adjust as needed
- Create escalation tags for questions requiring human judgment
5. Analytics & Reporting
- Share weekly health reports with community leads (builds transparency)
- Track week-over-week trends; investigate sudden drops in engagement
- Use member cohort analysis to identify at-risk groups early
- Set health score targets (e.g., "maintain 75+ score") and track progress
Safety & Guardrails
What This Skill Will NOT Do
- No Surveillance: Does not track private DMs or read deleted messages
- No Discriminatory Moderation: Will not flag messages based on protected characteristics (race, religion, gender, disability)
- No False Bans: Requires human confirmation before permanent bans
- No Spam Escalation: Will not auto-ban without mod review; uses graduated warnings
- No Data Sharing: Member data stays in your database; never shared with 3rd parties
- No Impersonation: Bot always identifies itself; never pretends to be human
- No Automated Decisions on Account Deletion: Requires human mod approval
Boundaries & Limitations
- Moderation Bias: AI toxicity detection (Google Perspective API) can have cultural/linguistic bias. Always review flagged content manually.
- Context Blindness: Bot may miss sarcasm, inside jokes, or context-dependent toxicity. Pair with human judgment.
- Escalation Overload: In communities >10K members, may generate high false-positive rates; recommend manual tuning.
- FAQ Hallucination: GPT-4 may generate plausible-sounding but inaccurate answers. Always review FAQ responses before production.
- Member Scoring Opacity: Composite scores can be hard to interpret; provide detailed breakdowns to members if challenged.
- Rate Limits: Discord API allows ~50 requests/sec; Slack allows ~20 requests/sec. In high-traffic communities, may experience lag.
Privacy Compliance
- GDPR: Supports data export/deletion on request
- CCPA: Complies with California privacy rights
- Discord/Slack ToS: Fully compliant with platform policies
- Data Retention: Configurable (default: 90 days for moderation logs, 1 year for member scores)
Troubleshooting
Common Issues & Solutions
Q: Bot is not responding to commands in Discord
- A: Check that bot has "Send Messages" and "Read Message History" permissions in the channel
- Verify DISCORD_BOT