Play 61
Content Moderation v2
Advanced multi-modal content moderation with cultural context and human appeal workflows.
Next-generation content moderation combining Azure AI Content Safety with GPT-powered cultural context analysis across text, image, and video modalities. Features severity-based routing to human reviewers, custom category training for domain-specific policies, real-time dashboards with false-positive tracking, and automated appeal workflows — all backed by Cosmos DB for audit trails and Service Bus for reliable async processing.
Architecture Pattern
Multi-modal classification: severity routing, human-in-the-loop, policy enforcement
Azure Services
DevKit (.github Agentic OS)
- agent.md — root orchestrator with builder→reviewer→tuner handoffs
- 3 agents — Moderation Builder (gpt-4o), Reviewer (gpt-4o-mini), Tuner (gpt-4o-mini)
- 3 skills — deploy (244 lines), evaluate (127 lines), tune (186 lines)
- 4 prompts — /deploy, /test, /review, /evaluate with agent routing
- .vscode/mcp.json — FrootAI MCP with Content Safety + OpenAI inputs + envFile
TuneKit (AI Config)
- config/openai.json — gpt-4o for cultural context, mini for triage
- config/guardrails.json — strictest safety=0.0, multi-modal rules
- evaluation/eval.py — Precision >95%, False positive <5%
Tuning Parameters
Machine evidence
FrootAI evidence lifecycle
This is an internal evidence maturity label, not third-party certification, accreditation, legal compliance, or a production guarantee. Missing or expired evidence demotes automatically; catalog claims cannot promote a play.
This play currently has design evidence only. A runnable scenario, endpoint evaluation, and build receipts are the next contiguous gates.
Repo Intelligence
v1A no-clone, revision-pinned map for agents and humans. Observed evidence is separated from inferred flow so the output stays useful without pretending to be a full call graph.