← ClaudeAtlas

auth-access-controllisted

Expert authentication and authorization engineering: identity lifecycle, session/token handling, password/OTP flows, RBAC/ABAC permission models, default-deny enforcement at every sensitive boundary, and safe audit logging. Use when implementing login/logout, role-permission matrices, policy gates, scoped multi-tenant access, or reviewing authn/authz for APIs and UI routes.
rogue-dev-studio/ai-agents-rogue-programmer · ★ 0 · API & Backend · score 63
Install: claude install-skill rogue-dev-studio/ai-agents-rogue-programmer
# Auth & Access Control (Canonical) **Expertise: expert.** Aliases: `auth`, `rbac`, `authorization`, `access-control`. ## When to use - Login, logout, refresh, reset password, OTP/session - Role, permission, policy, matrix akses - Scope data per tenant/wilayah/organisasi - Review endpoint/UI yang harus default deny ## When not to use - Hanya styling halaman login → `frontend-engineering` - Secret storage infra murni tanpa flow auth → `security` rule + DevOps ## Procedure 1. **Threat boundary** — Identifikasi aksi sensitif; default **deny**. 2. **Authn** — Kredensial aman (hash, TTL session/token, rotate/revoke). 3. **Authz** — Cek permission per aksi + scope data; jangan percaya “hidden UI”. 4. **Matrix** — Sumber kebenaran permission tersentral; hindari hardcode tersebar. 5. **Transport** — HTTPS; cookie flags / bearer hygiene sesuai stack; CSRF bila cookie session. 6. **Audit** — Log event keamanan tanpa menumpuk PII/secret. 7. **Verify** — Uji 401/403, escalation negatif, scope leak; libatkan Security pada trigger rule. Detail: `reference.md`. ## DoD - [ ] Default deny pada endpoint/aksi sensitif - [ ] Session/token lifecycle aman (expiry/revoke) - [ ] Permission + scope teruji (positif & negatif) - [ ] Tidak ada secret di log/repo - [ ] Review trigger `security` dipenuhi bila relevan