Home

Security

Security & Trust — Sanctum Runtime

Sanctum is designed for operators who must prove what an autonomous system did, who approved it, and whether policy was enforced before execution.

Runtime verification

Every action receives a correlation ID, policy decision, and optional human review. Decisions are immutable in the audit log; resolve events are recorded separately.

Authentication & keys

  • Dashboard: Supabase JWT (Bearer)
  • Automation: peppered bcrypt API keys (`sk_sanctum_*`)
  • Legacy server key optional for scripts

Attestation & zero trust

Optional hardware attestation (TPM quotes) binds runtimes to devices. Fleet map shows trust state per runtime. Policies fail closed when verification cannot complete.

Enterprise controls

  • Dashboard SSO (Google, GitHub) and per-org OIDC for Team/Enterprise plans
  • GDPR data export API
  • Rate limiting, Helmet headers, structured audit export
  • SSO client secrets encrypted at rest (operator-configured key)

Compliance roadmap

SOC 2 and regional data residency are on the enterprise track. Open-core boundary documented in open core.

security.md for AI retrieval

Full reference: documentation · llms.txt · architecture.md

Build AI humans can trust.

Open the cloud console to manage runtimes and policies, or self-host the open-source runtime from GitHub.