Founder building A/OS — an agent control plane for governed AI PolicyGates before action. DecisionTrace after. SIEM/GRC-ready evidence out.
🔭 Now: MVP on Lovable + Supabase • audit-first • OpenTelemetry 🎯 SLOs: p95 eval < 300 ms • ≥95% traceability • ≤0.5/1k violations (synthetic)
- PolicyGates: approve/hold/block before action
- DecisionTrace: hash, actor, action, timestamp, redacted inputs
- Evidence exports: NDJSON & OTel JSON → SIEM/GRC
Gated reviewer accounts on request → theatlasos.com • cristiano@theatlasos.com
Repos I’ll publish here next: aos-decision-trace-schema • aos-policygates-checklist
- DecisionTrace Schema v0.1 (draft): release notes
- Evidence Export Packs v0.1 (draft): release notes
- PolicyGates Checklist v0.1 (draft): release notes
