Skip to main content

Security & RBAC

DaoFlow uses a layered security model with roles, scopes, and API tokens. This ensures AI agents and humans operate with appropriate permissions.

Design Principles

  • Agent-first safety — agents default to read-only until explicitly granted write scopes
  • Granular scopes — 26 colon-delimited scopes covering every operation
  • Audit everything — every mutation creates an immutable audit record
  • Least privilege — tokens get only the scopes they need
  • Transparent denials — permission errors tell you exactly which scope is required

Security Layers

Principal (user/agent/service account)
→ Role (owner/admin/operator/developer/viewer/agent)
→ Capabilities (scopes granted by role)
→ API Token (optional scope restriction)
→ Effective Permissions (intersection of role + token scopes)

Quick Reference

TopicDescription
Roles6 built-in roles with different capability levels
Scopes26 colon-delimited permission scopes
API TokensCreating and managing scoped tokens
Agent PrincipalsDedicated identities for AI systems
Audit TrailImmutable log of all write operations