feat(tier2): finish the finish line — Block H follow-ups, Block G geolocation, cross-cutting

Three threads of work land here together to close out Tier 2.

### Block H follow-ups — day-of check-in
- Scanner is now an "open on your phone" magic-link flow. Hosts on
  desktop mint a scoped JWT via POST /events/{id}/scanner-ticket and
  render its URL into a QR; phone scans it and lands on /scanner with
  the ticket as bearer. The ticket carries Audience=scanner so it can
  never substitute for a session token.
- Plus-one confirmation at the door: scan → POST /check-in/preview to
  fetch guest + expected party size → confirm buttons ("Just them",
  "Party of N", custom) → POST /check-in. No more silent arrival_count=1.
- Offline scan queue: failed POSTs go into an IndexedDB store and drain
  on the 'online' event with poison-message protection.
- Day-of arrivals headline widget on the event overview, gated to the
  host's local calendar date so it doesn't dominate the page weeks out.
- Tab nav restyled with inline heroicons + scrollable segmented control;
  Check-in moves to the rightmost slot.
- PWA: manifest + service worker scoped to /scanner, generated 192/512
  icons (Go scripted renderer in scripts/gen-scanner-icons.go).
- Confirmation email QR was rendering broken because html/template
  rewrites data: URLs to #ZgotmplZ; mark the value as template.URL.
- Email "open your invitation" link 404'd because we had no token to
  put after /rsvp/. Threaded AccessLink through the RSVPConfirmed NATS
  event from the API at submit time.

### Block G remainder — geolocation + threshold preview
- Pluggable GeoResolver in the fraud engine (NullResolver, IPApiResolver
  for the free ip-api.com fallback, MaxMindResolver behind GG_GEOIP_DB_PATH).
  Wrapped in a Redis cache (30d TTL). Geo flows through both gRPC and
  NATS scoring paths.
- geo_jump scoring feature: >500km in <1h flags ("accessed from Lagos
  and Paris within 12 minutes"); >500km in <6h is a softer signal. The
  existing single-signal cap keeps a lone geo_jump in MEDIUM.
- FraudScored event carries geo_country/city/lat/lon; ApplyScore uses
  COALESCE so a later re-score without geo doesn't wipe earlier data.
- Threshold-slider live preview: GET /events/{id}/security/thresholds/preview
  returns band counts the host's existing access events would have
  fallen into under the proposed thresholds. Debounced (250ms) widget
  under the Advanced sliders so the host gets concrete feedback instead
  of guessing.

### Cross-cutting — audit, tier-gating, feature flags
- audit_log table + internal/audit.Recorder (async fire-and-forget on
  detached context so an audit blip never fails the real action). Wired
  into branding update, thresholds update, allowlist add/remove,
  collaborator invite/role-change/remove, message create/send-now/cancel.
- Tier-gating: extended billing.Limits with MaxCollaborators,
  CustomBranding, Scanner, Broadcasts. Free = none; Pro = 5 + all;
  Business = unlimited. Gates the scanner-ticket, message create,
  branding put, and collaborator invite endpoints with 402 +
  structured upgrade payload. Auto-reminders, fraud detection, and
  analytics deliberately stay on every tier — those are safety + visibility
  features, not upsell levers.
- Feature flags: feature_flags table + internal/flags.Store with 30s
  in-memory refresh, stable sha256(key + user_id) percent bucketing,
  unknown-key-defaults-on. Six Tier 2 flags pre-seeded. Three handlers
  (branding, broadcasts, scanner) check the kill switch ahead of the
  tier gate so ops can pull a feature back without a redeploy.

### Verified
- go test ./... + fraud-engine pytest (12/12 incl. 3 new geo_jump tests + 5
  new flags tests).
- docker compose build + up across api, fraud-engine, notifier, frontend.
- /health endpoints 200; migrations 0014 + 0015 applied; 6 flags
  seeded; audit_log table + partial indexes confirmed.
- Fraud-engine logs confirm geo resolver kind=CachedGeoResolver provider=auto.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

internal/audit/audit.go

@@ -0,0 +1,111 @@
+// Package audit records meaningful host-facing writes to an
+// append-only table so the timeline (and any future compliance
+// export) has a single source of truth.
+//
+// Design notes:
+//
+//   - All writes are best-effort + fire-and-forget. Audit logging must
+//     never block or fail the real action — a host saving branding
+//     should not 500 because the audit insert flaked. We log a warning
+//     and move on.
+//
+//   - Use the package-level `Record(...)` helper from handler code; it
+//     decorates the call with the request-id middleware adds to the
+//     context and dispatches to the wired Recorder. A nil Recorder
+//     (zero-value Server in tests) is a no-op.
+//
+//   - Action names follow `entity.verb` (e.g. `branding.update`,
+//     `collaborator.invite`). Verbs are past-tense-implied — the row's
+//     existence is the past tense.
+//
+//   - Metadata is freeform JSON; keep it small and reviewable. Don't
+//     put secrets in here — this table is queried by support.
+package audit
+
+import (
+	"context"
+	"encoding/json"
+	"log/slog"
+
+	"github.com/google/uuid"
+	"github.com/jackc/pgx/v5/pgxpool"
+)
+
+// Recorder writes audit rows asynchronously. Construct one per process
+// and pass it into the handlers that need it (server.go bundles it on
+// the few handlers that audit).
+type Recorder struct {
+	pool   *pgxpool.Pool
+	logger *slog.Logger
+}
+
+func New(pool *pgxpool.Pool, logger *slog.Logger) *Recorder {
+	if logger == nil {
+		logger = slog.Default()
+	}
+	return &Recorder{pool: pool, logger: logger}
+}
+
+// Params carries everything the audit_log INSERT needs. Most fields
+// are nullable so a record can describe an action that doesn't have a
+// specific target (e.g. "feature_flag.toggle") or doesn't sit under
+// an event (account-level writes).
+type Params struct {
+	UserID     *uuid.UUID
+	EventID    *uuid.UUID
+	Action     string         // e.g. "branding.update"
+	EntityType string         // e.g. "event", "guest", "message"
+	TargetID   *uuid.UUID     // the row that was acted on (nullable)
+	Metadata   map[string]any // free-form context — keep small
+	RequestID  string         // correlation id from middleware (if any)
+}
+
+// Record inserts one audit_log row. Returns immediately; the insert
+// runs on a detached goroutine with the package logger. Errors are
+// warned, never returned.
+func (r *Recorder) Record(ctx context.Context, p Params) {
+	if r == nil || r.pool == nil {
+		return
+	}
+	// Detach the context so cancelling the inbound HTTP request
+	// doesn't cancel the audit write mid-flight.
+	go r.write(context.WithoutCancel(ctx), p)
+}
+
+func (r *Recorder) write(ctx context.Context, p Params) {
+	var meta []byte
+	if p.Metadata != nil {
+		var err error
+		meta, err = json.Marshal(p.Metadata)
+		if err != nil {
+			r.logger.Warn("audit: marshal metadata", "err", err, "action", p.Action)
+			meta = []byte(`{}`)
+		}
+	} else {
+		meta = []byte(`{}`)
+	}
+
+	const q = `
+        INSERT INTO audit_log
+            (user_id, event_id, action, entity_type, target_id, metadata, request_id)
+        VALUES ($1, $2, $3, $4, $5, $6::jsonb, NULLIF($7, ''))
+    `
+	if _, err := r.pool.Exec(ctx, q,
+		p.UserID, p.EventID, p.Action, nilIfEmpty(p.EntityType),
+		p.TargetID, meta, p.RequestID,
+	); err != nil {
+		r.logger.Warn("audit: insert failed",
+			"err", err,
+			"action", p.Action,
+			"event_id", p.EventID,
+			"user_id", p.UserID,
+		)
+	}
+}
+
+func nilIfEmpty(s string) any {
+	if s == "" {
+		return nil
+	}
+	return s
+}