Skip to Main Content

MaddBoard Overseer is an AI decision system for high-risk infrastructure. It replaces opaque automation with evidence-backed, agent-debated recommendations that require human approval. Value: AI-Native: Orchestrates complex, evidence-gated workflows. Access: Brings expert SRE reasoning to small teams. Control: Amplifies human judgment via safety gates. Built & Demonstrated: Pipeline: 15 probes across 3 tiers; standardized output. Debate: 7 agents (6 Hats method) running parallel reasoning. Compiler: 4-pass validation with strict schemas. Mesh: 3-session coordinated architecture. Live: Real-world disk assessment with 100% evidence citation.

Team

liviuolos

Placement

Participant

More from Built with Opus 4.6: a Claude Code hackathon

CrossBeam tackles California's ADU permit crisis. 429,000+ permits issued since 2018, 90%+ sent back for corrections — each cycle costing weeks and thousands…

Mike project preview

Elisa is a block-based visual IDE where kids design software by snapping blocks together and AI agents build it. Users compose specs using seven primitives…

Jon McBee project preview

Built by a practicing cardiologist - because patients forget up to 80% of what their doctor says. PostVisit.ai bridges the gap between the doctor's office…

Michal Nedoszytko project preview

Conductr is a browser-based MIDI instrument where you play and Claude Opus 4.6 directs the band. Four tracks of drums, bass, melody, and harmony generate in…

Nanas Sound project preview

Zeppelin is a production-grade, S3-native vector search engine built entirely by Claude Opus 4.6 in long-running agentic sessions. 21,000 lines of Rust, 451…

Zeppelin project preview

FenceFlow eliminates paper-based workflows from fencing tournaments using Opus 4.6. Built from hosting 30+ events, it solves a real problem: staff squint at…

Nour Desouki project preview