feat(auth): add PAT auth, scoped REST/MCP access, CSRF, and tenant isolation

Introduce Personal Access Tokens (nzp_*) as a stateless auth path alongside
JWT, gated per-endpoint by a scope middleware (nezha:{resource}:{verb}) with
fail-closed empty-scope defaults and a server-id whitelist. Self-management
endpoints (profile, api-tokens, oauth2 bind, refresh-token) explicitly reject
PATs to block privilege-escalation chains. A revoke registry tears down active
long-lived connections (terminal, fm, ws, transfer, mcp) the moment a PAT is
deleted, with a tombstone closing the revoke->register race.

Add an MCP endpoint that proxies tool calls (exec, fs read/write/delete,
transfer) to agents over gRPC, guarded by origin/DNS-rebinding checks, a
per-token rate limiter, audit logging, and a kill switch. Serialize all
sends through the IOStream wrapper to honour grpc-go's concurrency contract.

Add CSRF double-submit protection on unsafe cookie-authenticated methods,
exempting authenticated PAT requests by context identity (not a forgeable
Authorization header). Apply visibility/whitelist filtering consistently
across list, get-by-id, and mutate paths to enforce tenant isolation.

Migrate legacy mcp:* scopes: rewrite read/exec to nezha:* equivalents and
drop dangerous write/delete/wildcard grants.

Co-authored-by: cloudcode <cloudcode@users.noreply.github.com>

cmd/dashboard/controller/mcp_transfer_audit_throttle.go

@@ -0,0 +1,72 @@
+package controller
+
+import (
+	"sync"
+	"time"
+)
+
+// transferAnonAuditThrottle caps the number of audit rows written per
+// source IP within a sliding window for transfer requests that failed
+// before a valid `entry` could be loaded (bogus/expired/replayed token).
+//
+// Without this cap an unauthenticated attacker can POST millions of
+// /mcp/upload/<random> requests; every miss invokes
+// writeTransferFailureAudit which inserts into mcp_audit_log. The
+// throttle keeps a small per-IP token bucket in memory and drops audit
+// rows past the budget — successful and authenticated failures (entry
+// != nil) bypass this gate entirely so SIEM signal is unaffected.
+type transferAnonAuditThrottle struct {
+	mu     sync.Mutex
+	window time.Duration
+	limit  int
+	hits   map[string]*anonHitBucket
+	clock  func() time.Time
+}
+
+type anonHitBucket struct {
+	firstAt time.Time
+	count   int
+}
+
+func newTransferAnonAuditThrottle(window time.Duration, perWindow int) *transferAnonAuditThrottle {
+	return &transferAnonAuditThrottle{
+		window: window,
+		limit:  perWindow,
+		hits:   make(map[string]*anonHitBucket),
+		clock:  time.Now,
+	}
+}
+
+// shouldRecord reports whether the anonymous failure for this IP should
+// land in the audit table. Empty ip is treated as "always record" since
+// suppressing it would silently lose signal in test/headless contexts.
+func (t *transferAnonAuditThrottle) shouldRecord(ip string) bool {
+	if ip == "" {
+		return true
+	}
+	t.mu.Lock()
+	defer t.mu.Unlock()
+	now := t.clock()
+	t.pruneLocked(now)
+
+	b, ok := t.hits[ip]
+	if !ok || now.Sub(b.firstAt) >= t.window {
+		t.hits[ip] = &anonHitBucket{firstAt: now, count: 1}
+		return true
+	}
+	if b.count >= t.limit {
+		return false
+	}
+	b.count++
+	return true
+}
+
+func (t *transferAnonAuditThrottle) pruneLocked(now time.Time) {
+	for ip, b := range t.hits {
+		if now.Sub(b.firstAt) >= t.window {
+			delete(t.hits, ip)
+		}
+	}
+}
+
+var transferAnonAuditThrottleShared = newTransferAnonAuditThrottle(time.Minute, 5)