Voice & Text Chat Moderation System

## ROLE
You are a trust and safety engineer who has built moderation systems for gaming communities with 50M+ users. You understand that effective moderation requires a blend of automated detection, community-driven reporting, and human review.

## OBJECTIVE
Design a moderation system for [GAME TITLE] that handles text chat, voice chat, and player behavior across [PLATFORM COUNT] platforms with [LANGUAGE COUNT] supported languages.

## TASK

### Text Chat Moderation
- Profanity filter: configurable severity levels (off, mild, strict), support for leetspeak and Unicode abuse
- Hate speech detection: ML-based classifier for slurs, targeted harassment, discriminatory language
- Spam prevention: rate limiting, duplicate message detection, advertisement blocking
- Context awareness: competitive trash talk vs genuine harassment threshold
- Player-configurable filters: let players choose their own comfort level
- Automated actions: message removal, shadow mute, temporary chat ban based on severity

### Voice Chat Moderation
- Real-time speech-to-text: transcription for keyword detection and logging
- Audio fingerprinting: detect repeated toxic audio clips, soundboard abuse
- Volume abuse detection: mic spamming, intentional distortion, screeching
- Recording and review: temporary audio capture for report investigation
- Privacy compliance: data retention policies, consent notifications, GDPR/COPPA requirements
- Mute tools: individual mute, team mute, automatic muting of reported players

### Behavioral Moderation
- Griefing detection: team killing, intentional feeding, objective sabotage, AFK detection
- Gameplay disruption: blocking teammates, stealing resources, refusing to cooperate
- Unsportsmanlike conduct: rage quitting patterns, surrender abuse, GG spam
- Impersonation: fake developer tags, staff impersonation, phishing attempts
- Account sharing: detecting multiple users on one account for competitive integrity

### Reporting System
- In-game reporting flow: maximum 3 taps to report, categorized options, optional details
- Evidence attachment: auto-capture last 60 seconds of gameplay/chat on report
- Report validation: weight reports by reporter accuracy history
- Cross-reference: multiple reports against same player escalate priority
- Feedback: "Action was taken" notifications to encourage future reporting

### Penalty & Rehabilitation
- Progressive discipline: verbal warning, chat restrict, temp ban, permanent ban
- Behavior score: visible or hidden score that affects matchmaking and rewards
- Positive reinforcement: honor system, sportsmanship awards, good behavior rewards
- Reform pathway: how penalized players can demonstrate improved behavior
- Ban evasion detection: hardware ID, IP tracking, play pattern matching

## OUTPUT FORMAT
Moderation system design document with technical architecture, policy guidelines, escalation procedures, and analytics dashboard requirements.

## CONSTRAINTS
- False positive rate must be extremely low for automated penalties
- System must support multiple languages and cultural contexts
- Voice moderation must comply with recording consent laws per region
- Moderation must not introduce noticeable latency to gameplay
- Include special protections for minors (COPPA, age-gating)