Skip to main content

What does success look like — in terms anyone can explain without reading the spec?

OUTCOME MAP: IDENTITY & ACCESS
════════════════════════════════════════════════════════════

  DESIRED OUTCOME
  Every user lands in the right org with the right
  permissions — no lockouts, no data leakage, no
  manual DB intervention to onboard a customer.
       │
       ├── Contributing Factors
       │   ├── Authentication solid — Clerk middleware + ClerkAuthAdapter 3-layer validation working
       │   ├── Multi-tenancy schema complete — all business tables FK to org_organisation
       │   ├── Identity resolution working — ExternalAuthId → SystemUserId → OrganisationId
       │   ├── Auto-provisioning exists — user + "Personal Workspace" org created on first login
       │   ├── PolicyEngine built — canI() method, governance tables, role-permission join all exist
       │   └── 80% of infrastructure already built — gap is wiring, not architecture
       │
       ├── Obstacles
       │   ├── PostgreSQL 22P02 — User-Role query passes Clerk userId (string) where UUID expected
       │   │   → Fix the query to use systemUserId, not Clerk userId
       │   ├── Infinite redirect loop — /dashboard → /sign-in → /dashboard every 400ms (789 errors/30min)
       │   │   → Show error page on access denied, not redirect to sign-in
       │   ├── No role seed data — governance_user_roles table empty, roles can't be assigned
       │   │   → Seed Admin/Member/Viewer roles on deploy
       │   ├── defaultAllow: true — PolicyEngine bypasses all permission checks
       │   │   → Switch to defaultAllow: false after roles seeded and assigned
       │   └── No invite flow — users can only self-provision, no way to join existing org
       │       → Build invite-by-email after RBAC working
       │
       ├── Investigations (answer before committing)
       │   ├── Is auto-provision transactional? If it fails mid-create, is there a half-provisioned user? [Owner: Engineering]
       │   ├── Will Clerk organization sync create duplicate orgs? [Owner: Engineering]
       │   └── Can base repository class enforce org scoping without breaking existing queries? [Owner: Engineering]
       │
       ├── Success Measures (binary)
       │   ├── Owner logs in → dashboard loads, no redirect loop → YES / NO
       │   ├── First user in new org auto-gets Admin role → YES / NO
       │   ├── defaultAllow: false in production, no permission = no access → YES / NO
       │   └── User in org1 cannot access org2 data (negative security test) → YES / NO
       │
       ├── Roles (RACI)
       │   ├── Accountable: Wik (commissioning decisions, go/no-go on each tier)
       │   ├── Responsible: Engineering team (fix bugs, build tiers, ship features)
       │   ├── Consulted: Clerk documentation (auth provider patterns, org sync)
       │   └── Informed: All venture PRDs blocked by this (Sales CRM, Sales Dev Agent, Content Amplifier)
       │
       └── Next Actions (momentum within 48 hours)
           ├── Fix User-Role query: use systemUserId not Clerk userId (Tier 0, task 1)
           ├── Break redirect loop: show error page on admin_access_required (Tier 0, task 2)
           └── Set ADMIN_EMAILS env var in Vercel production (Tier 0, task 3)

════════════════════════════════════════════════════════════

Gate​

Before moving to Value Stream Map:

  • Desired outcome is crisp: "Every user lands in the right org with the right permissions" — YES
  • Contributing factors have evidence: each references existing built infrastructure verified in codebase audit
  • Obstacles have mitigations assigned: each has a fix strategy with tier reference
  • Investigation questions have owners: all Engineering team
  • Success measures are binary: all four are YES/NO with clear verification method
  • RACI is complete: one Accountable (Wik), one Responsible (Engineering)
  • Next actions are assigned: three Tier 0 tasks with immediate start

Context​

Questions​

Which outcome in the identity access map is the leading indicator — the one that reliably predicts all others?

  • Which outcome is hardest to measure but most important to track — and what proxy metric comes closest?
  • If the primary outcome were achieved but secondary outcomes lagged, what would that signal about the system?
  • At what threshold does each outcome shift from "in progress" to "proven" — and who decides that threshold?