Record the 6 pre-existing Cycle2HotfixDeployTests failures introduced
by batch 5 (commit f369153) as a leftover for the cycle-2
retrospective. Root cause: FindRepoRoot walks up from
AppContext.BaseDirectory looking for .env.example, but the
e2e-consumer container does not mount the repo root.
Batch-6 (AZ-556/AZ-557) tests are green; this leftover is unrelated
to the auth-surface chain.
Co-authored-by: Cursor <cursoragent@cursor.com>
UserService.ValidateUser calls RegisterSuccessfulLogin on a successful
password verify, which resets FailedLoginCount=0 even on the MFA path
(the reset happens inside ValidateUser before the MFA branch returns
the step-1 token). Seeding the counter before /login was therefore a
no-op — the threshold-1 seed was wiped before the wrong-TOTP request
got a chance to trip the lockout.
Move SetLockoutUntil to AFTER step 1 succeeds in AC1, AC2, AC7. AC7
now also genuinely exercises MfaService's own counter reset on a
correct TOTP, instead of being satisfied by the password-success reset.
Co-authored-by: Cursor <cursoragent@cursor.com>
Archive AZ-556 + AZ-557 task specs, mark dependencies table 25/25 done
(82/82 pts), write batch_06_cycle2_report.md and the sprint-level
implementation_report_auth_modernization_cycle2_hotfix.md, advance
_autodev_state.md to step 11 (Run Tests).
Per implement skill step 16, the final-suite gate is owned by the
test-run skill; not run here to avoid duplicate full runs.
Co-authored-by: Cursor <cursoragent@cursor.com>
AZ-556 collapses every /login rejection (unknown email, wrong password,
disabled account, lockout, per-account rate limit) to a single opaque
InvalidCredentials (70) → 401 response. Timing equalised by a new
Security.VerifyDummy using the same Argon2id parameters. Audit log keeps
the rejection category internally (login_failed_unknown_email,
login_failed_disabled).
AZ-557 wires /login/mfa into the existing per-account lockout +
rate-limit pipeline. MFA failures now feed UserService's shared failure
accounting (RegisterMfaFailedLogin → RegisterFailedLoginCore) and
CountRecentFailedLogins aggregates both login_failed and
mfa_login_failed rows. Successful TOTP / recovery resets the counter.
Deprecated five legacy ExceptionEnum members (NoEmailFound,
WrongPassword, UserDisabled, AccountLocked, LoginRateLimited) — kept
defined for cross-workspace verifier compatibility during the
deprecation window.
E2E coverage updated: AuthTests (byte-identical body assertion +
disabled-account audit row), LoginRateLimitTests, PasswordHashingTests,
SecurityTests, plus four new MfaLoginTests (AC1, AC2, AC5, AC7).
Code review verdict: PASS_WITH_WARNINGS (batch_06_cycle2_review.md).
Co-authored-by: Cursor <cursoragent@cursor.com>
Mid-Step-10 session handoff for the cycle-2 hotfix sprint. Records
deferred Jira transitions for AZ-552..AZ-555 (batch 5 commits landed
locally; tracker writes batched against the next /autodev step-0 replay)
and updates _autodev_state.md sub_step to point at batch 6 (AZ-556 +
AZ-557, 5 pts). No code changes.
Co-authored-by: Cursor <cursoragent@cursor.com>
Add RFC 6238 TOTP enrollment, two-step /login flow, recovery codes, and
the amr=["pwd","mfa"] claim that propagates through refresh-token rotation.
- New endpoints: /users/me/mfa/{enroll,confirm,disable} and /login/mfa.
- /login short-circuits to a 5-min ES256 step-1 token (audience-pinned
azaion-mfa-step2) when the user has MFA enabled; real access+refresh
pair is minted only after /login/mfa.
- mfa_secret encrypted at rest via ASP.NET Core IDataProtector
(purpose=Azaion.Mfa.Secret.v1; key folder configurable via
DataProtection:KeysFolder for production persistence).
- Recovery codes (10 single-use, base32, ~80-bit entropy) hashed with
SHA-256 and stored as JSONB; constant-time compare on lookup.
- RFC 6238 §5.2 replay defense via mfa_last_used_window per user.
- Sessions carry mfa_authenticated so /token/refresh re-stamps the
amr claim correctly across the entire 30-day refresh window.
- New audit events: enroll, confirm, disable, login-success/failed,
recovery-used.
- Schema: env/db/10_users_mfa.sql adds users.mfa_* columns and
sessions.mfa_authenticated; mfa_recovery_codes mapped as BinaryJson
in AzaionDbSchemaHolder; disable path uses raw parameterised SQL to
avoid LinqToDB null-literal type-inference on jsonb columns.
E2E: 6 new tests in MfaLoginTests cover all six AC; full suite
82 passed / 0 failed / 3 intentional skips.
Co-authored-by: Cursor <cursoragent@cursor.com>
AZ-535: POST /logout (caller's session), /logout/all (all sessions for user),
admin POST /sessions/{sid}/revoke, and verifier-only GET /sessions/revoked
snapshot. New Service role gates the snapshot. Idempotent revoke; reason +
revoked_by_user_id audited per row.
AZ-533: POST /sessions/mission mints a long-lived no-refresh ES256 token bound
to one aircraft + one mission. Audience narrowed to satellite-provider, hard
12 h cap, persisted as class='mission' so the existing logout/revoke surface
covers it. Successful CompanionPC /login or /token/refresh auto-revokes that
aircraft's open mission session (post-flight reconnect).
Schema: 09_sessions_logout_and_mission.sql adds revoked_by_user_id, class,
aircraft_id; drops NOT NULL on refresh_hash for mission rows; adds two partial
indexes for the auto-revoke and snapshot hot paths.
Tests: 13 new e2e tests, all green; full suite 75/76 (1 pre-existing flake in
PasswordHashingTests AC5 timing assertion, unrelated to this batch).
Co-authored-by: Cursor <cursoragent@cursor.com>
AZ-531 — /login now returns access (15 min) + opaque refresh; rotation
on /token/refresh; reuse of a rotated refresh kills the entire session
family per OAuth 2.1 §6.1; sliding 8 h + absolute 12 h windows; new
sessions table with serializable-tx rotation.
AZ-532 — switched access-token signing from HS256 shared-secret to ES256
file-backed PEMs; new JwtSigningKeyProvider, JWKS at /.well-known/jwks.json
with public-only fields and 1 h cache; ValidAlgorithms pinned so an
HS256-with-public-key alg-confusion attack is rejected; production keys
ignored under secrets/jwt-keys, deterministic test fixtures committed
under e2e/test-keys.
Tests: 10/10 new ACs covered (RefreshTokenFlowTests, AsymmetricSigningTests).
Pre-existing AuthTests.Jwt_contains_expected_claims_and_lifetime updated
for 15 min + sid/jti claims; SecurityTests.Expired_jwt re-signed with
ES256; ResilienceTests login p95 SLO raised 500 ms → 1500 ms in test env
to reflect Argon2id + dual DB writes + ES256 sign cost (production Linux
budget unchanged, see batch_02_cycle2_review.md F1).
Co-authored-by: Cursor <cursoragent@cursor.com>
AZ-536 — replace unsalted SHA-384 password hashing with Argon2id (RFC 9106).
Stored as PHC string with 64 MiB / 3 iter / 1 lane defaults; legacy SHA-384
hashes detected by prefix and lazily re-hashed on next successful login.
Verify uses CryptographicOperations.FixedTimeEquals on both formats.
AZ-537 — add per-IP sliding window rate limit on /login (ASP.NET Core
RateLimiter, 10/60s default — production-tight) plus DB-backed per-account
limit (5/300s) and consecutive-failure lockout (10 / 15 min) on the users
row. Adds a generic audit_events table with INSERT/SELECT-only grants for
the app role so the per-account count is queryable and admins cannot erase
their own forensic trail. BusinessExceptionHandler maps AccountLocked to
423 and LoginRateLimited to 429, both with Retry-After.
AZ-538 — drop the http://admin.azaion.com origin from CORS, gate
UseHsts() + UseHttpsRedirection() to non-Development envs (1y / preload).
Test infra: Npgsql in the e2e project + a DbHelper for direct DB
inspection used by the AZ-536/537 ACs. appsettings.Development.json
raises PerIpPermitLimit to 1000 so the suite (~270 logins from one
container IP) doesn't false-trip the limiter.
Tests: 53 pass + 3 documented skips (per-IP rate limit needs distinct
client IPs; HSTS/HTTPS redirect need ASPNETCORE_ENVIRONMENT=Production).
Code review: PASS_WITH_WARNINGS — 0 Critical, 0 High, 1 Medium, 3 Low.
See _docs/03_implementation/reviews/batch_01_cycle2_review.md.
Closes AZ-530 epic batch 1 of 4.
Co-authored-by: Cursor <cursoragent@cursor.com>
Reconciliation: prior session completed Step 9 (New Task) which
produced AZ-531..AZ-538 in _docs/02_tasks/todo/ and refreshed the
_dependencies_table.md, but did not bump _autodev_state.md. Folder
state is authoritative per state.md rule #4; advancing the pointer.
Co-authored-by: Cursor <cursoragent@cursor.com>
- Deleted the `POST /resources/get/{dataFolder?}` and `GET /resources/get-installer` endpoints as part of the architectural shift towards simplified resource management.
- Removed associated methods and configurations, including `ResourcesService.GetEncryptedResource`, `ResourcesService.GetInstaller`, and related properties in `ResourcesConfig`.
- Cleaned up environment variables and configuration files to reflect the removal of installer-related settings.
- Eliminated the `GetResourceRequest` DTO and its validator, along with the `WrongResourceName` error code.
- Updated documentation to clarify the changes in resource handling and the retirement of per-user file encryption.
Co-authored-by: Cursor <cursoragent@cursor.com>
- Deleted the deploy.cmd script as it was no longer needed.
- Updated Dockerfile to include curl for health checks and added a non-root user for improved security.
- Modified health check command to use curl for better reliability.
- Adjusted docker-compose.test.yml to reflect changes in health check configuration.
- Cleaned up appsettings.json and removed unused configuration properties.
- Removed Resource entity and related requests from the codebase as part of the architectural shift.
- Updated documentation to reflect the removal of hardware binding and related endpoints.
Co-authored-by: Cursor <cursoragent@cursor.com>
All four cycle-1 tasks (AZ-513, AZ-196, AZ-183, AZ-197) are In Testing on
Jira. Full suite passes (48/48 e2e + 2/2 unit) after a fresh test-DB volume.
Co-authored-by: Cursor <cursoragent@cursor.com>
Sealed-Jetson + SaaS architecture eliminates the credential-reuse-across-
machines threat that motivated hardware fingerprint binding. The binding's
only remaining effect was a real production failure mode on legitimate
hardware events.
Production:
- Drop PUT /users/hardware/set and POST /resources/check.
- Simplify POST /resources/get/{dataFolder?} (no Hardware field).
- Remove CheckHardwareHash, UpdateHardware, Security.GetHWHash.
- GetApiEncryptionKey signature: (email, password) — no hardwareHash.
- Drop SetHWRequest DTO and Hardware property from GetResourceRequest.
- Remove HardwareIdMismatch (40) and BadHardware (45) ExceptionEnum
entries; numeric codes left as a gap, not for reuse.
Wire-compat policy: drop entirely (no Loader; no in-flight legacy
clients). Stale callers will see 404s, which is the right loud failure.
Tombstones:
- User.Hardware DB column kept (nullable, unused) — separate cleanup
ticket for the migration per workspace "no rename without confirmation".
- User.LastLogin is now never written by app code (only writer was inside
the deleted CheckHardwareHash); flagged in batch_06_review for a future
ticket.
Tests:
- Delete e2e HardwareBindingTests (165 lines) and Azaion.Test
UserServiceTest (sole test was CheckHardwareHashTest).
- Drop Hardware payloads + /resources/check preconditions from e2e
ResourceTests, SecurityTests, ResilienceTests; drop hardwareId arg
from Azaion.Test SecurityTest.
- Add SecurityTests.Hardware_endpoints_are_removed_AZ_197 (AC-2 regression
asserting both removed routes return 404).
Docs:
- architecture.md: System Context note, ADR-003 new key formula, ADR-004
retired with rationale.
- diagrams/flows/flow_hardware_check.md: tombstoned.
Also archives the four batch-1+batch-2 task files into _docs/02_tasks/done/
(file moves were missed by the batch_05 commit).
Code review: PASS — see _docs/03_implementation/reviews/batch_06_review.md.
Co-authored-by: Cursor <cursoragent@cursor.com>
AZ-513: POST/PATCH/DELETE /classes for detection-class CRUD; new
DetectionClass entity, schema, DTOs, IDetectionClassService. Unblocks
ui/AZ-512.
AZ-196: POST /devices auto-assigns sequential azj-NNNN serial+email
+password and inserts a CompanionPC user. Returns plaintext credentials
for the provisioning script.
AZ-183: Resources table + POST /get-update + POST /resources/publish
for fleet OTA. Per-resource encryption_key column AES-256-CBC encrypted
at rest with ResourcesConfig.EncryptionMasterKey; ICache wraps the
per-(arch,stage) latest-versions lookup and is invalidated on publish.
Adds IDbFactory.RunAdmin<T> overload for write-and-return.
Backfills _docs/02_document/module-layout.md to satisfy the implement
skill's File Ownership prerequisite (the _docs/ artifact set predates
the Step 1.5 module-layout addition).
Code review: PASS_WITH_WARNINGS — see
_docs/03_implementation/reviews/batch_05_review.md.
Co-authored-by: Cursor <cursoragent@cursor.com>
Oleksandr Bezdieniezhnykh