Retire env-var SMTP basic-auth and send the contact form through Gmail over
OAuth2 (SMTP XOAUTH2), configured in Admin -> Settings -> Email via an in-app
"Connect Gmail" consent flow. Reuses the existing google SSO OAuth client; the
captured refresh token is stored AES-GCM-encrypted (write-only over the API,
never returned), mirroring the auth-provider and Discord-bot secret patterns.
- schema: new email_config singleton table (mirrors bot_config)
- model: emailConfig.{db,model} with encrypted refresh token + getSafe/getWithSecret
- mailer: nodemailer OAuth2 transport (client id/secret from the google provider
row), contact recipient = contact_email setting, mailto: fallback preserved,
plus sendTest()
- routes/controller: /admin/email config, connect start+callback (ssoState CSRF
+ PKCE), test, disconnect
- client: EmailDelivery section on the Settings page + api methods; Settings copy
now spells out that contact_email is the delivery recipient
- docs/env: drop SMTP_*/CONTACT_TO from env examples; update README/BACKEND_DESIGN
- tests: emailConfig.model + mailer suites (8 new; full suite 142 pass)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XKeCQEJZr1AFJN4Bgcmvh3
req.user comes from getById which strips password_hash, so has_password was
always false — the account page mis-rendered a real password account as the
SSO-only 'set a password' variant (and the change-password form omitted the
required current-password field). Read the raw row for that one flag.
Caught by a browser click-through of the /account portal. Adds a regression
test.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019rao86n5cXpwAyjdBFEshV
Light up the moderation dashboard's previously-empty widgets by persisting the
event streams the bot only reacted to in-memory before.
Schema (bot-owned)
- member_events: join/leave, with invite_code/inviter_* for best-effort invite
attribution on joins
- filter_hits: word / foreign-invite filter deletions (matched + action_taken)
- spam_hits: rate_limit / mass_mention / mass_emoji detections
Bot
- new models memberEvents/filterHits/spamHits
- guildMemberAdd records the join with invite attribution; new inviteTracker.js
keeps an invite-use cache (GuildInvites intent + inviteCreate/inviteDelete) and
diffs it on join to find which invite was used — best-effort, never blocks
auto-role
- new guildMemberRemove records leaves
- messageFilter records filter/spam hits alongside the existing warn/mute;
inviteFilter now returns the offending code; detectSpam identifies which spam
rule tripped (preserving the rate-limit-first side-effect order)
- mod_actions still logs the resulting warn/mute — the new tables are additive
Server
- summary extended with joins/leaves/invite_joins/filter_hits/spam_hits per window
- new feeds: /api/v1/admin/moderation/{members,filter-hits,spam-hits}
Client
- overview now shows 8 tiles (mod actions + joins/leaves/filter/spam, joins tile
notes "N via invite") plus an Events panel with Members/Filter/Spam tabs;
removed the coming-soon note
Verified: 119 server unit tests, client build, 14-check DB-backed smoke, and a
browser click-through of every tile and events tab (incl. invite attribution).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019rao86n5cXpwAyjdBFEshV
Surface the Discord bot's moderation data on the admin panel: a read-only
staff dashboard over the existing mod_actions log, per-user history, staff
notes, and a new moderator role. No bot changes.
Schema
- users.role ENUM gains 'moderator' (CREATE + idempotent ALTER for existing DBs)
- new server-owned mod_notes table (staff_only/admin_only visibility)
Server
- model/moderation: read mod_actions via the shared pool (documented read-only
cross of the bot/server ownership boundary), correlate accounts through
user_identities (provider='discord'), flag automated actions via
staff_user_id === bot_config.application_id; pure reshaping helpers isolated
in moderation.pure.js so they unit-test without opening a DB pool
- model/modNotes: list/add with role-gated admin_only visibility
- admin/moderation.controller + routes under /api/v1/admin/moderation/* gated by
requireRole('admin','moderator'); admin_only note writes require admin
- allow assigning 'moderator' in the user create/update validators
Client
- /admin/moderation overview (window tiles, type-filterable recent feed, user
lookup) and /user/:discordId history (tabs + notes with add-note)
- RoleGate; AdminLayout filters nav and confines moderators to their section
- moderator badge + action-type/auto badges
Deferred (see plan): 6b bot event capture (joins/leaves/filter/spam), 6c appeals
(needs public accounts), 6d /internal/mod-reverse bot reversal callback.
Verified: 116 server unit tests, client build, DB-backed model smoke, full
HTTP/RBAC e2e, and a browser click-through of the dashboard.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019rao86n5cXpwAyjdBFEshV
SSO login minted a full session immediately, ignoring the account's
totp_enabled flag — so a 2FA admin with a linked Google/Discord/OIDC
identity could sign in without their authenticator code, silently
downgrading the account to single-factor (the strength of the IdP login).
The local password flow already gates on needsTotp(); SSO did not.
Wire SSO through the same staged-TOTP gate:
- ssoState: createTotpPending/verifyTotpPending + a short-lived httpOnly
sso_totp cookie. The pending token carries stage:'totp' (session
validation rejects it) + kind:'sso_totp' (scoped to the SSO endpoint)
plus the resolved context (userId, provider, authMethod, returnTo).
- sso.controller: finishLogin now stages the challenge and redirects to
/admin/login?sso_totp=1 instead of creating a session when the account
has TOTP on. New finishSsoTotp verifies the code (backoff + bot-scoring
on failure, mirroring loginTotp) and only then mints the session.
- sso.routes: POST /auth/sso/totp behind the same backoff/slow/limiter
stack and code validation as the local TOTP endpoint.
- client: AdminLogin detects ?sso_totp=1 and completes over fetch via
api.ssoLoginTotp; the challenge never touches the URL or JS.
Keeps the second factor httpOnly throughout, consistent with the SSO tx
cookie. 12 new tests; full suite 106/106.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019rao86n5cXpwAyjdBFEshV
sessionFromDecoded validated sessions with a blocklist — it rejected a
token only when `decoded.stage` was present (the TOTP challenge). Because
every JWT is signed with the same JWT_SECRET and distinguished only by
claims, the SSO transaction cookie (sso_tx, which carries kind:'sso_tx'
and id:'sso' but no stage) passed validation and was accepted as a bogus
{ userId:'sso' } session.
requireAuth's DB re-load blocked protected admin routes, but non-DB
identity checks were fooled — notably siteMode's maintenance-preview
bypass, which trusts any truthy getUserFromRequest. An attacker could
start an SSO flow to obtain an sso_tx cookie and replay it as the auth
cookie / Bearer token to bypass the maintenance gate. The broader risk
was latent: any future code path trusting attachSession/getUserFromRequest
without a DB round-trip inherited an auth bypass.
Make session validation positively typed: real sessions are now stamped
with typ:'session' (createSession + mintMobileTokens), and
sessionFromDecoded accepts a token only when that marker is present. As
belt-and-suspenders it also rejects any token carrying a non-session
marker (stage || kind). Flow/challenge tokens are never stamped, so they
can no longer be mistaken for sessions.
Note: existing web cookie sessions predating this change lack the typ
claim and will be rejected once — users re-login. Mobile clients recover
automatically on next refresh.
Adds regression tests: the sso_tx flow token and a bare identity token
are both rejected by validateSession / decodeIdentity / getUserFromRequest.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Web sessions were stateless JWTs with no server-side store: the revocation
hooks in session.service were stubs that only logged. As a result web logout
was client-side only (a copied cookie stayed valid until natural JWT expiry)
and a password change never invalidated existing sessions. The mobile bearer
flow already had revocable, DB-stored tokens; this brings the web/cookie flow
to parity.
Two-layer revocation, both enforced in requireAuth (which already loads the
fresh user row each request):
- Per-session denylist: new `revoked_sessions` table keyed on the JWT `jti`
(already minted per session). A single logout adds this session's jti;
rows self-expire at the token's own exp and are pruned on boot. New model
`revokedSessions` mirrors the `mobileSessions` db/model split.
- Per-user cutoff: new `users.tokens_valid_after` column. A password change
(and the new `invalidateSessions` helper) bumps it to NOW(); any token whose
iat is at or before the cutoff is rejected. The comparison is inclusive so a
token minted in the same wall-clock second as the change is still revoked.
Wiring:
- session.service: revokeSession / invalidateSession / invalidateAllUserSessions
now delegate to the stores; sessions carry `expiresAt` (JWT exp) so logout can
set a self-pruning denylist row.
- /logout gains best-effort attachSession so the controller can revoke this
session's jti and log auth.logout; stays a no-op for anonymous callers.
- users.model.update bumps the cutoff whenever the password hash is rotated.
- schema.sql: revoked_sessions table + tokens_valid_after column, added to the
CREATE and to the idempotent migration block (ensureSchema on boot).
Verified end-to-end against the local dev DB: a captured cookie is rejected
after logout, and an existing session is rejected after a password change while
re-login with the new password succeeds. Full server test suite green (96).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019rao86n5cXpwAyjdBFEshV
The GET /internal/bot-config route returns the DECRYPTED Discord bot
token and was mounted on the same Express app / port 3000 that Pangolin
proxies publicly. Its only guard was the BOT_INTERNAL_KEY shared secret,
and .env.example shipped a placeholder default — so a forwarded path or a
weak/unrotated key would expose the plaintext token to the internet.
Move server<->bot internal traffic onto its own listener and fail fast on
a weak key:
- Add server/src/internalApp.js: a standalone Express app mounting
requireInternalKey + /internal (and a no-secret /health), mirroring the
bot's unpublished port-4100 pattern.
- server.js starts a second listener on INTERNAL_PORT (default 3001),
closed on graceful shutdown.
- Remove the /internal mount from the public v1.router; the public app now
404s /api/v1/internal/bot-config even with a valid key.
- Fail fast: new utils/botInternalKey.js rejects an empty, placeholder, or
<16-char BOT_INTERNAL_KEY — fatal in production (exit 1), warning in dev.
- docker-compose: bot SITE_INTERNAL_URL -> app:3001/internal/bot-config;
document that INTERNAL_PORT stays unpublished.
- .env.example (root/server/bot): document INTERNAL_PORT, the fail-fast
behavior, and a defense-in-depth Pangolin deny rule for /api/v1/internal.
Tests: add requireInternalKey.test.js and botInternalKey.test.js
(node --test: 93 pass).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019rao86n5cXpwAyjdBFEshV
Refactor authentication into a provider-agnostic session layer and build
two new auth surfaces on top of it, without changing local password/TOTP
behavior. Every flow now issues sessions through
sessionService.createSession(user, authMethod).
Part 1 — Session abstraction (backward-compatible refactor):
- New server/src/auth/: token.js (JWT/cookie primitives), session.service.js
(create/validate/partial-TOTP/revoke), session.middleware.js
(attachSession/requireAuth/requireRole). utils/auth.js is now a thin
compat facade so existing imports are unchanged.
Part 2 — Mobile bearer auth (additive):
- /api/v1/auth/mobile/{login,refresh,logout}: short-lived access JWT +
long-lived refresh token, stored hashed and rotated on use, in a new
mobile_refresh_tokens table. Reuses web bot-scoring/backoff; single-request
TOTP. token.signToken gains a backward-compatible expiresIn option.
Part 3 — Pluggable SSO (Google, Discord, generic OIDC):
- OAuth2Provider base + built-in Google/Discord (fixed endpoints) + generic
OIDC, a registry with health/validation, PKCE+CSRF transaction state, and
discovery (GET /auth/providers), start/link/callback routes.
- Link-only policy: SSO signs in only to an already-linked account; external
identities are never auto-provisioned. Client secrets encrypted at rest
(AES-256-GCM, utils/secretBox.js). Admin CRUD (/admin/auth/providers) and
account linking (/admin/account/identities). New auth_providers +
user_identities tables.
Frontend:
- Login page renders provider buttons from /auth/providers (inline SVG icons,
graceful with zero providers). New Authentication admin view
(Local/Google/Discord/Custom). Account page linked-accounts section.
Tests: 83 passing (session, mobile, providers, registry, secretBox, ssoState,
ssoCallback) — all DB-free via fetch mocks + model stubs. README + .env.example
updated.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Expose the botScore middleware's in-memory scoring/ban state to admins.
Previously state lived only in the store Map with no persistence or API — the
only visibility was tailing container logs.
- botScore: bounded ring buffer (300) recording scan/login-fail/honeypot and
ban events (most-recent-first); listState() snapshot of all scored IPs;
unban() to clear a single IP.
- New admin-only endpoints GET /admin/bot-activity and
POST /admin/bot-activity/unban (RBAC admin gate, IP validated). Unban is
activity-logged with the admin username.
- Bot Activity tab: currently-banned table with Unban, plus a recent-events
feed, following the existing admin table patterns.
- Tests for the buffer, listState, and unban (guard lets an unbanned IP back
through). README updated.
Read + emergency-unban only — no ban-add or weight-editing surface. Buffer is
in-memory, matching the store; not persisted.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Adds a layered set of protections around the admin login and the app edge.
Trust proxy (server/src/utils/trustProxy.js)
- Configurable via TRUST_PROXY; pin to the newt agent ("ptero") LAN IP so
X-Forwarded-For is trusted ONLY from that peer. A blanket "true" is
rejected (coerced to 1) to prevent XFF spoofing that would dodge every
IP-based control. DEBUG_TRUST_PROXY logs peer/XFF/req.ip to re-verify the
proxy IP without a redeploy. Documents the Omada static-reservation
assumption.
Login throttling (server/src/middleware/loginProtection.js, rateLimit.js)
- express-slow-down progressive delay + the existing hard rate cap + a
separate per-IP exponential backoff that persists across the rate window.
All failures return one generic message (no user/pass disclosure).
Honeypot (login form + auth.controller)
- Hidden, plausibly-named field ("company"); a filled value fails
generically and is scored as an unambiguous bot.
Optional per-user TOTP 2FA (speakeasy/qrcode)
- totp_secret/totp_enabled columns (+ idempotent migration). Self-service
Account page: enroll via QR, confirm a code to enable, code-gated disable.
- Login is two-step for enrolled users: after the password, a short-lived
signed challenge (stage:'totp', not a session) is required before the
real session is issued.
Bot / scanner scoring + IP ban (server/src/middleware/botScore.js)
- Weighted CMS-scanner paths (this app uses none). Junk paths 404 FIRST,
unconditionally — independent of score/ban state, so a scanner rotating
through fresh Cloudflare IPs gets no free pass. /wp-admin/install.php is
the top-weighted near-1-hit ban (worst offender in prod logs). Per-IP
score with quiet-period decay temp-bans an IP from ALL routes once past a
(deliberately low) threshold, to protect /admin from credential stuffing.
Failed logins and honeypot hits feed the same score.
- Periodic sweep evicts stale, unbanned, quiet entries so the in-memory
store can't grow unbounded; the interval is unref'd and cleared on
graceful shutdown.
Tests: node --test suite (40) covering trust-proxy parsing + live req.ip
(incl. pinned-IP), rate limiter + exponential backoff, honeypot rejection,
TOTP verify (enabled/disabled) + challenge-isn't-a-session, bot-score
threshold/decay/ban + junk-404-independence + install.php + store sweep.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
