handle envelop on grpc GetNetworkMap

On iOS the tunnel reconfiguration (setTunnelNetworkSettings) was driven
from three independent Go paths without serialization: route prefix
updates via the route notifier's own delivery loop, interface IP/IPv6
set synchronously from the engine start goroutine, and DNS config
applied from the DNS apply chain through a separate Swift object. The
three sources mutated the shared Swift settings-manager state
concurrently and triggered overlapping updateTunnel() calls, losing or
half-applying route and DNS settings.

Introduce client/internal/tunnelnotifier: a single notifier that
implements both listener.NetworkChangeListener and dns.IosDnsManager,
queues all four callbacks (OnNetworkChanged, SetInterfaceIP,
SetInterfaceIPv6, ApplyDns) in one FIFO and delivers them one-by-one
from a single goroutine, so calls into Swift never overlap and arrive in
order. RunOniOS wraps the two Swift objects into the notifier and closes
it after the run loop exits.

The route notifier keeps its prefix dedup but delegates delivery to the
shared notifier instead of maintaining its own queue and loop; the dedup
check and the enqueue stay under one mutex so queue order matches
state-update order. Setting the interface IP becomes asynchronous, which
is safe: on iOS wgInterface.Create() only uses the TunFd, and the FIFO
preserves the IP -> routes/DNS relative order.

The package is build-tag free so the unit tests run on any platform
under -race. Android is unaffected: it receives all settings atomically
in one configureInterface call and serializes TUN rebuilds on a single
handler thread.

## Describe your changes

## Issue ticket number and link

## Stack

<!-- branch-stack -->

### Checklist
- [x] Is it a bug fix
- [ ] Is a typo/documentation fix
- [ ] Is a feature enhancement
- [x] It is a refactor
- [ ] Created tests that fail without the change (if possible)
- [ ] This change does **not** modify the public API, gRPC protocols,
functionality behavior, CLI / service flags, or introduce a new feature
— **OR** I have discussed it with the NetBird team beforehand (link the
issue / Slack thread in the description). See
[CONTRIBUTING.md](https://github.com/netbirdio/netbird/blob/main/CONTRIBUTING.md#discuss-changes-with-the-netbird-team-first).

> By submitting this pull request, you confirm that you have read and
agree to the terms of the [Contributor License
Agreement](https://github.com/netbirdio/netbird/blob/main/CONTRIBUTOR_LICENSE_AGREEMENT.md).

## Documentation
Select exactly one:

- [ ] I added/updated documentation for this change
- [x] Documentation is **not needed** for this change (explain why)

### Docs PR URL (required if "docs added" is checked)
Paste the PR link from https://github.com/netbirdio/docs here:

https://github.com/netbirdio/docs/pull/__

<!-- codesmith:footer -->
---
<a
href="https://app.blacksmith.sh/netbirdio/codesmith/netbird/pr/6870"><picture><source
media="(prefers-color-scheme: dark)"
srcset="https://pr-comments-assets.blacksmith.sh/codesmith/view-with-codesmith-dark-v2.svg"><source
media="(prefers-color-scheme: light)"
srcset="https://pr-comments-assets.blacksmith.sh/codesmith/view-with-codesmith-light-v2.svg"><img
alt="View with [code]smith"
src="https://pr-comments-assets.blacksmith.sh/codesmith/view-with-codesmith-dark-v2.svg"></picture></a>
<a
href="https://backend.blacksmith.sh/track/enable-autofix?expires=1787404610&installation_model_id=427504&pr_number=6870&repository=netbirdio%2Fnetbird&return_to=https%3A%2F%2Fgithub.com%2Fnetbirdio%2Fnetbird%2Fpull%2F6870&signature=7db66c4ec4262cc21137b3d3e1fd5e905950eab82ef46236a13ac3d512ff2a16"><picture><source
media="(prefers-color-scheme: dark)"
srcset="https://pr-comments-assets.blacksmith.sh/codesmith/autofix-with-codesmith-dark.svg"><source
media="(prefers-color-scheme: light)"
srcset="https://pr-comments-assets.blacksmith.sh/codesmith/autofix-with-codesmith-light.svg"><img
alt="Autofix with [code]smith"
src="https://pr-comments-assets.blacksmith.sh/codesmith/autofix-with-codesmith-dark.svg"></picture></a>
<sup>Need help on this PR? Tag <code>@codesmith-bot</code> with what you
need. Autofix is disabled.</sup>

<!-- codesmith:autofix:disabled -->
<!-- /codesmith:footer -->

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit

* **New Features**
* Improved iOS tunnel networking by coordinating network, interface,
route, and DNS updates through a unified notifier.
* Updated iOS behavior so route/prefix changes are applied immediately
instead of via queued/background delivery.

* **Bug Fixes**
* Improved reliability by ensuring network and DNS-related callbacks are
invoked in order without overlap.
* Ensured pending updates are drained and handled correctly during
shutdown.

* **Tests**
* Added coverage for FIFO ordering, non-overlapping callback execution,
interleaved DNS/route updates, and graceful shutdown behavior.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->

.goreleaser_ui.yaml

@@ -24,6 +24,8 @@ builds:
     ldflags:
       - -s -w -X github.com/netbirdio/netbird/version.version={{.Version}} -X main.commit={{.Commit}} -X main.date={{.CommitDate}} -X main.builtBy=goreleaser
     mod_timestamp: "{{ .CommitTimestamp }}"
+    tags:
+      - production
 
   - id: netbird-ui-windows-amd64
     dir: client/ui
@@ -39,6 +41,8 @@ builds:
       - -s -w -X github.com/netbirdio/netbird/version.version={{.Version}} -X main.commit={{.Commit}} -X main.date={{.CommitDate}} -X main.builtBy=goreleaser
       - -H windowsgui
     mod_timestamp: "{{ .CommitTimestamp }}"
+    tags:
+      - production
 
   - id: netbird-ui-windows-arm64
     dir: client/ui
@@ -55,6 +59,8 @@ builds:
       - -s -w -X github.com/netbirdio/netbird/version.version={{.Version}} -X main.commit={{.Commit}} -X main.date={{.CommitDate}} -X main.builtBy=goreleaser
       - -H windowsgui
     mod_timestamp: "{{ .CommitTimestamp }}"
+    tags:
+      - production
 
 archives:
   - id: linux-arch

.goreleaser_ui_darwin.yaml

@@ -29,6 +29,8 @@ builds:
     ldflags:
       - -s -w -X github.com/netbirdio/netbird/version.version={{.Version}} -X main.commit={{.Commit}} -X main.date={{.CommitDate}} -X main.builtBy=goreleaser
     mod_timestamp: "{{ .CommitTimestamp }}"
+    tags:
+      - production
 
 universal_binaries:
   - id: netbird-ui-darwin

client/android/profile_manager.go

@@ -189,6 +189,19 @@ func (pm *ProfileManager) LogoutProfile(id string) error {
 	return nil
 }
 
+// RenameProfile changes a profile's display name. The profile ID, and therefore
+// its on-disk filename, is left untouched: only the "name" field of the config
+// is rewritten. This works for the default profile too, whose config lives in
+// netbird.cfg rather than under profiles/.
+func (pm *ProfileManager) RenameProfile(id string, newName string) error {
+	if err := pm.serviceMgr.RenameProfile(profilemanager.ID(id), androidUsername, newName); err != nil {
+		return fmt.Errorf("failed to rename profile: %w", err)
+	}
+
+	log.Infof("renamed profile %s to: %s", id, newName)
+	return nil
+}
+
 // RemoveProfile deletes a profile
 func (pm *ProfileManager) RemoveProfile(id string) error {
 	// Use ServiceManager (removes profile from profiles/ directory)

client/internal/connect.go

@@ -34,6 +34,7 @@ import (
 	"github.com/netbirdio/netbird/client/internal/profilemanager"
 	"github.com/netbirdio/netbird/client/internal/statemanager"
 	"github.com/netbirdio/netbird/client/internal/stdnet"
+	"github.com/netbirdio/netbird/client/internal/tunnelnotifier"
 	"github.com/netbirdio/netbird/client/internal/updater"
 	"github.com/netbirdio/netbird/client/internal/updater/installer"
 	nbnet "github.com/netbirdio/netbird/client/net"
@@ -136,10 +137,13 @@ func (c *ConnectClient) RunOniOS(
 	// Set GC percent to 5% to reduce memory usage as iOS only allows 50MB of memory for the extension.
 	debug.SetGCPercent(5)
 
+	notifier := tunnelnotifier.New(networkChangeListener, dnsManager)
+	defer notifier.Close()
+
 	mobileDependency := MobileDependency{
 		FileDescriptor:        fileDescriptor,
-		NetworkChangeListener: networkChangeListener,
-		DnsManager:            dnsManager,
+		NetworkChangeListener: notifier,
+		DnsManager:            notifier,
 		StateFilePath:         stateFilePath,
 		TempDir:               cacheDir,
 	}

client/internal/mobile_dependency.go

@@ -11,12 +11,14 @@ import (
 
 // MobileDependency collect all dependencies for mobile platform
 type MobileDependency struct {
-	// Android only
-	TunAdapter            device.TunAdapter
-	IFaceDiscover         stdnet.ExternalIFaceDiscover
+	// Android and iOS
 	NetworkChangeListener listener.NetworkChangeListener
-	HostDNSAddresses      []netip.AddrPort
-	DnsReadyListener      dns.ReadyListener
+
+	// Android only
+	TunAdapter       device.TunAdapter
+	IFaceDiscover    stdnet.ExternalIFaceDiscover
+	HostDNSAddresses []netip.AddrPort
+	DnsReadyListener dns.ReadyListener
 
 	//	iOS only
 	DnsManager     dns.IosDnsManager

client/internal/routemanager/dnsinterceptor/handler.go

@@ -95,7 +95,7 @@ func (d *DnsInterceptor) RemoveRoute() error {
 
 			// AllowedIPs should use real IPs
 			if d.currentPeerKey != "" {
-				if _, err := d.allowedIPsRefcounter.Decrement(prefix); err != nil {
+				if _, err := d.allowedIPsRefcounter.Decrement(prefix, d.currentPeerKey); err != nil {
 					merr = multierror.Append(merr, fmt.Errorf("remove allowed IP %s: %v", prefix, err))
 				}
 			}
@@ -172,7 +172,7 @@ func (d *DnsInterceptor) removeAllowedIP(realPrefix netip.Prefix) error {
 	}
 
 	// AllowedIPs use real IPs
-	if _, err := d.allowedIPsRefcounter.Decrement(realPrefix); err != nil {
+	if _, err := d.allowedIPsRefcounter.Decrement(realPrefix, d.currentPeerKey); err != nil {
 		return fmt.Errorf("remove allowed IP %s: %v", realPrefix, err)
 	}
 
@@ -205,7 +205,7 @@ func (d *DnsInterceptor) RemoveAllowedIPs() error {
 	for _, prefixes := range d.interceptedDomains {
 		for _, prefix := range prefixes {
 			// AllowedIPs use real IPs
-			if _, err := d.allowedIPsRefcounter.Decrement(prefix); err != nil {
+			if _, err := d.allowedIPsRefcounter.Decrement(prefix, d.currentPeerKey); err != nil {
 				merr = multierror.Append(merr, fmt.Errorf("remove allowed IP %s: %v", prefix, err))
 			}
 		}

client/internal/routemanager/dynamic/route.go

@@ -135,7 +135,7 @@ func (r *Route) RemoveAllowedIPs() error {
 	var merr *multierror.Error
 	for _, domainPrefixes := range r.dynamicDomains {
 		for _, prefix := range domainPrefixes {
-			if _, err := r.allowedIPsRefcounter.Decrement(prefix); err != nil {
+			if _, err := r.allowedIPsRefcounter.Decrement(prefix, r.currentPeerKey); err != nil {
 				merr = multierror.Append(merr, fmt.Errorf("remove allowed IP %s: %w", prefix, err))
 			}
 		}
@@ -320,7 +320,7 @@ func (r *Route) removeRoutes(prefixes []netip.Prefix) ([]netip.Prefix, error) {
 			merr = multierror.Append(merr, fmt.Errorf("remove dynamic route for IP %s: %w", prefix, err))
 		}
 		if r.currentPeerKey != "" {
-			if _, err := r.allowedIPsRefcounter.Decrement(prefix); err != nil {
+			if _, err := r.allowedIPsRefcounter.Decrement(prefix, r.currentPeerKey); err != nil {
 				merr = multierror.Append(merr, fmt.Errorf("remove allowed IP %s: %w", prefix, err))
 			}
 		}

client/internal/routemanager/manager.go

@@ -216,7 +216,7 @@ func (m *DefaultManager) setupRefCounters(useNoop bool) {
 		)
 	}
 
-	m.allowedIPsRefCounter = refcounter.New(
+	m.allowedIPsRefCounter = refcounter.NewAllowedIPs(
 		func(prefix netip.Prefix, peerKey string) (string, error) {
 			// save peerKey to use it in the remove function
 			return peerKey, m.wgInterface.AddAllowedIP(peerKey, prefix)

client/internal/routemanager/notifier/notifier_ios.go

@@ -3,7 +3,6 @@
 package notifier
 
 import (
-	"container/list"
 	"net/netip"
 	"slices"
 	"sort"
@@ -16,20 +15,12 @@ import (
 
 type Notifier struct {
 	mu              sync.Mutex
-	cond            *sync.Cond
 	currentPrefixes []string
 	listener        listener.NetworkChangeListener
-	queue           *list.List
-	closed          bool
 }
 
 func NewNotifier() *Notifier {
-	n := &Notifier{
-		queue: list.New(),
-	}
-	n.cond = sync.NewCond(&n.mu)
-	go n.deliverLoop()
-	return n
+	return &Notifier{}
 }
 
 func (n *Notifier) SetListener(listener listener.NetworkChangeListener) {
@@ -59,44 +50,19 @@ func (n *Notifier) OnNewPrefixes(prefixes []netip.Prefix) {
 	sort.Strings(newNets)
 
 	n.mu.Lock()
+	defer n.mu.Unlock()
 	if slices.Equal(n.currentPrefixes, newNets) {
-		n.mu.Unlock()
 		return
 	}
 	n.currentPrefixes = newNets
-	routes := strings.Join(n.currentPrefixes, ",")
-	n.queue.PushBack(routes)
-	n.cond.Signal()
-	n.mu.Unlock()
+	if n.listener != nil {
+		n.listener.OnNetworkChanged(strings.Join(n.currentPrefixes, ","))
+	}
 }
 
 func (n *Notifier) Close() {
-	n.mu.Lock()
-	n.closed = true
-	n.cond.Signal()
-	n.mu.Unlock()
 }
 
 func (n *Notifier) GetInitialRouteRanges() []string {
 	return nil
 }
-
-func (n *Notifier) deliverLoop() {
-	for {
-		n.mu.Lock()
-		for n.queue.Len() == 0 && !n.closed {
-			n.cond.Wait()
-		}
-		if n.closed && n.queue.Len() == 0 {
-			n.mu.Unlock()
-			return
-		}
-		routes := n.queue.Remove(n.queue.Front()).(string)
-		l := n.listener
-		n.mu.Unlock()
-
-		if l != nil {
-			l.OnNetworkChanged(routes)
-		}
-	}
-}

client/internal/routemanager/reconcile_test.go

@@ -54,7 +54,7 @@ func (m *reconcileWGMock) GetNet() *netstack.Net                      { return n
 func TestReconcilePeerAllowedIPs(t *testing.T) {
 	wg := &reconcileWGMock{}
 	m := &DefaultManager{wgInterface: wg}
-	m.allowedIPsRefCounter = refcounter.New[netip.Prefix, string, string](
+	m.allowedIPsRefCounter = refcounter.NewAllowedIPs(
 		func(_ netip.Prefix, peerKey string) (string, error) { return peerKey, nil },
 		func(netip.Prefix, string) error { return nil },
 	)

client/internal/routemanager/refcounter/allowedips.go

@@ -0,0 +1,206 @@
+package refcounter
+
+import (
+	"errors"
+	"fmt"
+	"net/netip"
+	"sort"
+	"sync"
+
+	"github.com/hashicorp/go-multierror"
+
+	nberrors "github.com/netbirdio/netbird/client/errors"
+)
+
+// allowedIPsEntry holds the per-peer reference counts for a single prefix and which peer is
+// currently installed in WireGuard. WireGuard allows a prefix on exactly one peer, so at most
+// one peer is active at a time even when several peers reference the prefix.
+type allowedIPsEntry struct {
+	// peers maps a peerKey to the number of references holding the prefix for that peer.
+	peers map[string]int
+	// active is the peerKey currently installed in WireGuard for this prefix ("" if none).
+	active string
+	// total is the sum of all per-peer reference counts (kept in sync with peers).
+	total int
+}
+
+// AllowedIPsRefCounter is a peer-aware reference counter for WireGuard AllowedIPs.
+//
+// The generic Counter keys only by prefix and remembers a single Out value set by the first
+// caller, which it never changes. That is wrong for AllowedIPs: two independent watchers (or
+// multiple resolved domains) can reference the same prefix through different peers, and when the
+// peer currently installed in WireGuard releases its last reference the prefix must be handed over
+// to a surviving peer instead of being left pointing at the released one.
+//
+// It calls add/remove (which program WireGuard) only on the transitions that matter:
+//   - add on the first reference for a prefix, or when swapping the active peer;
+//   - remove on the last reference for a prefix, or on the old peer during a swap.
+type AllowedIPsRefCounter struct {
+	mu      sync.Mutex
+	entries map[netip.Prefix]*allowedIPsEntry
+	add     AddFunc[netip.Prefix, string, string]
+	remove  RemoveFunc[netip.Prefix, string]
+}
+
+// NewAllowedIPs creates a new peer-aware AllowedIPs reference counter.
+// add programs a prefix on a peer in WireGuard and returns the peerKey to store as the active peer.
+// remove unprograms the prefix from the given peer.
+func NewAllowedIPs(add AddFunc[netip.Prefix, string, string], remove RemoveFunc[netip.Prefix, string]) *AllowedIPsRefCounter {
+	return &AllowedIPsRefCounter{
+		entries: map[netip.Prefix]*allowedIPsEntry{},
+		add:     add,
+		remove:  remove,
+	}
+}
+
+// Increment adds a reference to prefix for peerKey. WireGuard is programmed only for the first
+// reference to a prefix; while a different peer is already installed the prefix is left with it
+// (first peer wins, HA at the WireGuard layer is not possible) and only the reference count is kept.
+func (rm *AllowedIPsRefCounter) Increment(prefix netip.Prefix, peerKey string) (Ref[string], error) {
+	rm.mu.Lock()
+	defer rm.mu.Unlock()
+
+	e, ok := rm.entries[prefix]
+	if !ok {
+		e = &allowedIPsEntry{peers: map[string]int{}}
+		rm.entries[prefix] = e
+	}
+
+	logCallerF("Increasing allowed IP ref count for prefix %v peer %s [peer %d -> %d, total %d -> %d, active %q]",
+		prefix, peerKey, e.peers[peerKey], e.peers[peerKey]+1, e.total, e.total+1, e.active)
+
+	// Program WireGuard only when nothing is installed yet for this prefix.
+	if e.active == "" {
+		out, err := rm.add(prefix, peerKey)
+		if errors.Is(err, ErrIgnore) {
+			if e.total == 0 {
+				delete(rm.entries, prefix)
+			}
+			return Ref[string]{Count: e.total, Out: e.active}, nil
+		}
+		if err != nil {
+			if e.total == 0 {
+				delete(rm.entries, prefix)
+			}
+			return Ref[string]{}, fmt.Errorf("failed to add allowed IP %v for peer %s: %w", prefix, peerKey, err)
+		}
+		e.active = out
+	}
+
+	e.peers[peerKey]++
+	e.total++
+
+	return Ref[string]{Count: e.total, Out: e.active}, nil
+}
+
+// Decrement removes a reference to prefix for peerKey. When the peer currently installed in
+// WireGuard releases its last reference, the prefix is swapped to a surviving peer if one exists,
+// otherwise it is removed from WireGuard.
+func (rm *AllowedIPsRefCounter) Decrement(prefix netip.Prefix, peerKey string) (Ref[string], error) {
+	rm.mu.Lock()
+	defer rm.mu.Unlock()
+
+	e, ok := rm.entries[prefix]
+	if !ok {
+		logCallerF("No allowed IP reference found for prefix %v", prefix)
+		return Ref[string]{}, nil
+	}
+
+	if e.peers[peerKey] > 0 {
+		logCallerF("Decreasing allowed IP ref count for prefix %v peer %s [peer %d -> %d, total %d -> %d, active %q]",
+			prefix, peerKey, e.peers[peerKey], e.peers[peerKey]-1, e.total, e.total-1, e.active)
+		e.peers[peerKey]--
+		e.total--
+		if e.peers[peerKey] == 0 {
+			delete(e.peers, peerKey)
+		}
+	} else {
+		logCallerF("No allowed IP reference found for prefix %v peer %s", prefix, peerKey)
+	}
+
+	// If the peer currently installed in WireGuard still holds references, nothing to reprogram.
+	// Keying the check on the active peer (not the one just released) makes this self-healing:
+	// a prior swap whose remove/add failed leaves e.active pointing at a peer with no references,
+	// and this retries the hand-off on the next Decrement instead of getting stuck.
+	if e.active != "" && e.peers[e.active] > 0 {
+		return Ref[string]{Count: e.total, Out: e.active}, nil
+	}
+
+	// Detach the stale/gone active peer from WireGuard before reprogramming.
+	if e.active != "" {
+		if err := rm.remove(prefix, e.active); err != nil {
+			return Ref[string]{Count: e.total, Out: e.active}, fmt.Errorf("remove allowed IP %v for peer %s: %w", prefix, e.active, err)
+		}
+		e.active = ""
+	}
+
+	// Hand the prefix over to a surviving peer, or drop the entry when none remain.
+	if survivor, ok := pickSurvivor(e.peers); ok {
+		out, err := rm.add(prefix, survivor)
+		if err != nil {
+			return Ref[string]{Count: e.total, Out: ""}, fmt.Errorf("swap allowed IP %v to peer %s: %w", prefix, survivor, err)
+		}
+		e.active = out
+		return Ref[string]{Count: e.total, Out: e.active}, nil
+	}
+
+	delete(rm.entries, prefix)
+	return Ref[string]{Count: 0, Out: ""}, nil
+}
+
+// Flush removes all prefixes from WireGuard and clears the counter.
+func (rm *AllowedIPsRefCounter) Flush() error {
+	rm.mu.Lock()
+	defer rm.mu.Unlock()
+
+	var merr *multierror.Error
+	for prefix, e := range rm.entries {
+		if e.active == "" {
+			continue
+		}
+		logCallerF("Flushing allowed IP for prefix %v peer %s", prefix, e.active)
+		if err := rm.remove(prefix, e.active); err != nil {
+			merr = multierror.Append(merr, fmt.Errorf("remove allowed IP %v for peer %s: %w", prefix, e.active, err))
+		}
+	}
+
+	clear(rm.entries)
+
+	return nberrors.FormatErrorOrNil(merr)
+}
+
+// ReapplyMatching calls apply for every prefix whose currently installed (active) peer satisfies
+// pred, holding the lock for the whole pass. It is used to re-push allowed IPs onto a peer whose
+// WireGuard entry was rebuilt (e.g. a lazy connection cycling idle->wake) without a matching
+// refcounter change, which would otherwise leave the prefix installed in the counter but missing
+// on the device. Only the active peer is considered — a prefix that lost its installed peer to a
+// failed swap is skipped here and reconciled by the next Increment/Decrement.
+func (rm *AllowedIPsRefCounter) ReapplyMatching(pred func(out string) bool, apply func(key netip.Prefix) error) error {
+	rm.mu.Lock()
+	defer rm.mu.Unlock()
+
+	var merr *multierror.Error
+	for prefix, e := range rm.entries {
+		if e.active != "" && pred(e.active) {
+			if err := apply(prefix); err != nil {
+				merr = multierror.Append(merr, err)
+			}
+		}
+	}
+	return nberrors.FormatErrorOrNil(merr)
+}
+
+// pickSurvivor deterministically selects a peer still referencing the prefix. WireGuard cannot do
+// multipath for a single prefix, so any surviving peer is a valid winner; the choice is made stable
+// (lowest peerKey) for predictable behavior and testability.
+func pickSurvivor(peers map[string]int) (string, bool) {
+	if len(peers) == 0 {
+		return "", false
+	}
+	keys := make([]string, 0, len(peers))
+	for k := range peers {
+		keys = append(keys, k)
+	}
+	sort.Strings(keys)
+	return keys[0], true
+}

client/internal/routemanager/refcounter/allowedips_test.go

@@ -0,0 +1,241 @@
+package refcounter
+
+import (
+	"errors"
+	"net/netip"
+	"testing"
+)
+
+// fakeWG models WireGuard's cryptokey routing: a prefix can be installed on exactly one peer.
+// failAdd/failRemove make the next add/remove fail once, to exercise the self-healing error paths.
+type fakeWG struct {
+	installed  map[netip.Prefix]string
+	adds       int
+	removes    int
+	failAdd    bool
+	failRemove bool
+}
+
+func newFakeWG() *fakeWG {
+	return &fakeWG{installed: map[netip.Prefix]string{}}
+}
+
+func (f *fakeWG) counter() *AllowedIPsRefCounter {
+	return NewAllowedIPs(
+		func(prefix netip.Prefix, peerKey string) (string, error) {
+			if f.failAdd {
+				f.failAdd = false
+				return "", errors.New("add failed")
+			}
+			f.adds++
+			f.installed[prefix] = peerKey
+			return peerKey, nil
+		},
+		func(prefix netip.Prefix, peerKey string) error {
+			if f.failRemove {
+				f.failRemove = false
+				return errors.New("remove failed")
+			}
+			f.removes++
+			// only clear if this peer is the one installed, mirroring wg semantics
+			if f.installed[prefix] == peerKey {
+				delete(f.installed, prefix)
+			}
+			return nil
+		},
+	)
+}
+
+func mustPrefix(t *testing.T, s string) netip.Prefix {
+	t.Helper()
+	p, err := netip.ParsePrefix(s)
+	if err != nil {
+		t.Fatalf("parse prefix %q: %v", s, err)
+	}
+	return p
+}
+
+func mustIncrement(t *testing.T, c *AllowedIPsRefCounter, p netip.Prefix, peer string) Ref[string] {
+	t.Helper()
+	ref, err := c.Increment(p, peer)
+	if err != nil {
+		t.Fatalf("Increment(%v, %s): %v", p, peer, err)
+	}
+	return ref
+}
+
+func mustDecrement(t *testing.T, c *AllowedIPsRefCounter, p netip.Prefix, peer string) Ref[string] {
+	t.Helper()
+	ref, err := c.Decrement(p, peer)
+	if err != nil {
+		t.Fatalf("Decrement(%v, %s): %v", p, peer, err)
+	}
+	return ref
+}
+
+// TestAllowedIPs_SwapOnActivePeerRemoval reproduces the reported bug: two networks with the same
+// prefix routed by different peers. Removing the network whose peer is installed must hand the
+// prefix over to the surviving peer instead of leaving it on the removed one.
+func TestAllowedIPs_SwapOnActivePeerRemoval(t *testing.T) {
+	f := newFakeWG()
+	c := f.counter()
+	p := mustPrefix(t, "10.44.8.0/24")
+
+	mustIncrement(t, c, p, "peerA")
+	mustIncrement(t, c, p, "peerB")
+	// First peer wins while both are present.
+	if got := f.installed[p]; got != "peerA" {
+		t.Fatalf("expected peerA installed, got %q", got)
+	}
+
+	// Remove the active peer's network -> must swap to peerB.
+	mustDecrement(t, c, p, "peerA")
+	if got := f.installed[p]; got != "peerB" {
+		t.Fatalf("BUG: prefix stuck on removed peer, want peerB got %q", got)
+	}
+
+	// Remove the last one -> prefix gone.
+	mustDecrement(t, c, p, "peerB")
+	if _, ok := f.installed[p]; ok {
+		t.Fatalf("expected prefix removed, still installed on %q", f.installed[p])
+	}
+}
+
+// TestAllowedIPs_RemoveNonActivePeer removing a non-installed peer must not touch WireGuard.
+func TestAllowedIPs_RemoveNonActivePeer(t *testing.T) {
+	f := newFakeWG()
+	c := f.counter()
+	p := mustPrefix(t, "10.44.8.0/24")
+
+	mustIncrement(t, c, p, "peerA")
+	mustIncrement(t, c, p, "peerB")
+	removesBefore := f.removes
+
+	mustDecrement(t, c, p, "peerB")
+	if f.installed[p] != "peerA" {
+		t.Fatalf("active peer must stay peerA, got %q", f.installed[p])
+	}
+	if f.removes != removesBefore {
+		t.Fatalf("removing a non-active peer must not call wg remove")
+	}
+}
+
+// TestAllowedIPs_SamePeerMultipleRefs two references via the same peer must keep the prefix until
+// the last reference is released (the reason the per-peer count must be an int, not a set).
+func TestAllowedIPs_SamePeerMultipleRefs(t *testing.T) {
+	f := newFakeWG()
+	c := f.counter()
+	p := mustPrefix(t, "10.44.8.0/24")
+
+	mustIncrement(t, c, p, "peerA")
+	mustIncrement(t, c, p, "peerA")
+	if f.adds != 1 {
+		t.Fatalf("expected a single wg add for the same peer, got %d", f.adds)
+	}
+
+	mustDecrement(t, c, p, "peerA")
+	if f.installed[p] != "peerA" {
+		t.Fatalf("prefix must stay while a reference remains, got %q", f.installed[p])
+	}
+	if f.removes != 0 {
+		t.Fatalf("no wg remove expected while a reference remains, got %d", f.removes)
+	}
+
+	mustDecrement(t, c, p, "peerA")
+	if _, ok := f.installed[p]; ok {
+		t.Fatalf("prefix must be removed after last reference")
+	}
+}
+
+// TestAllowedIPs_RefCountAndActive checks the Ref returned to callers (used for the HA-disabled log).
+func TestAllowedIPs_RefCountAndActive(t *testing.T) {
+	f := newFakeWG()
+	c := f.counter()
+	p := mustPrefix(t, "10.44.8.0/24")
+
+	ref := mustIncrement(t, c, p, "peerA")
+	if ref.Count != 1 || ref.Out != "peerA" {
+		t.Fatalf("want {1, peerA}, got {%d, %q}", ref.Count, ref.Out)
+	}
+	ref = mustIncrement(t, c, p, "peerB")
+	if ref.Count != 2 || ref.Out != "peerA" {
+		t.Fatalf("want {2, peerA}, got {%d, %q}", ref.Count, ref.Out)
+	}
+}
+
+// TestAllowedIPs_Flush removes everything installed and clears the counter.
+func TestAllowedIPs_Flush(t *testing.T) {
+	f := newFakeWG()
+	c := f.counter()
+	p1 := mustPrefix(t, "10.44.8.0/24")
+	p2 := mustPrefix(t, "10.44.9.0/24")
+
+	mustIncrement(t, c, p1, "peerA")
+	mustIncrement(t, c, p2, "peerB")
+
+	if err := c.Flush(); err != nil {
+		t.Fatal(err)
+	}
+	if len(f.installed) != 0 {
+		t.Fatalf("expected all prefixes removed, got %v", f.installed)
+	}
+	// After flush, a fresh increment must add again.
+	mustIncrement(t, c, p1, "peerC")
+	if f.installed[p1] != "peerC" {
+		t.Fatalf("counter not reset after flush")
+	}
+}
+
+// TestAllowedIPs_SelfHealAfterSwapAddError ensures a failed add during a swap does not permanently
+// strand the prefix: the next Decrement (or Increment) must retry and install a surviving peer.
+func TestAllowedIPs_SelfHealAfterSwapAddError(t *testing.T) {
+	f := newFakeWG()
+	c := f.counter()
+	p := mustPrefix(t, "10.44.8.0/24")
+
+	mustIncrement(t, c, p, "peerA")
+	mustIncrement(t, c, p, "peerB")
+	mustIncrement(t, c, p, "peerC")
+
+	// Removing the active peerA triggers a swap to a survivor; make the add fail once.
+	f.failAdd = true
+	if _, err := c.Decrement(p, "peerA"); err == nil {
+		t.Fatalf("expected error from failed swap add")
+	}
+	if _, ok := f.installed[p]; ok {
+		t.Fatalf("nothing should be installed after a failed swap add, got %q", f.installed[p])
+	}
+
+	// A later Decrement of a non-active survivor must retry the hand-off (self-heal), not stay stuck.
+	ref := mustDecrement(t, c, p, "peerC")
+	if got := f.installed[p]; got == "" {
+		t.Fatalf("self-heal failed: prefix left unrouted after add recovered")
+	}
+	if ref.Out == "" {
+		t.Fatalf("expected an active peer after self-heal, got empty")
+	}
+}
+
+// TestAllowedIPs_SelfHealAfterRemoveError ensures a failed remove during a swap is retried instead
+// of leaving e.active stuck on a peer that no longer holds references.
+func TestAllowedIPs_SelfHealAfterRemoveError(t *testing.T) {
+	f := newFakeWG()
+	c := f.counter()
+	p := mustPrefix(t, "10.44.8.0/24")
+
+	mustIncrement(t, c, p, "peerA")
+	mustIncrement(t, c, p, "peerB")
+
+	// Releasing active peerA must detach it (remove) then add peerB; fail the remove once.
+	f.failRemove = true
+	if _, err := c.Decrement(p, "peerA"); err == nil {
+		t.Fatalf("expected error from failed remove")
+	}
+
+	// Next Decrement of the non-active survivor retries: removes stale peerA, installs peerB.
+	mustDecrement(t, c, p, "peerB")
+	// peerB had only one ref, so after retry the prefix is fully released.
+	if _, ok := f.installed[p]; ok {
+		t.Fatalf("expected prefix released after self-heal, still on %q", f.installed[p])
+	}
+}

client/internal/routemanager/refcounter/types.go

@@ -5,5 +5,7 @@ import "net/netip"
 // RouteRefCounter is a Counter for Route, it doesn't take any input on Increment and doesn't use any output on Decrement
 type RouteRefCounter = Counter[netip.Prefix, struct{}, struct{}]
 
-// AllowedIPsRefCounter is a Counter for AllowedIPs, it takes a peer key on Increment and passes it back to Decrement
-type AllowedIPsRefCounter = Counter[netip.Prefix, string, string]
+// AllowedIPsRefCounter tracks WireGuard AllowedIPs per prefix. Unlike the generic Counter it is peer-aware:
+// a prefix can be claimed by several peers at once and WireGuard allows a given prefix on exactly one peer,
+// so the counter records the per-peer reference count and swaps the installed peer when the active one is released.
+// See allowedips.go.

client/internal/routemanager/static/route.go

@@ -15,6 +15,11 @@ type Route struct {
 	route                *route.Route
 	routeRefCounter      *refcounter.RouteRefCounter
 	allowedIPsRefcounter *refcounter.AllowedIPsRefCounter
+	// currentPeerKey is the routing peer this watcher currently has the prefix installed on
+	// (the HA winner elected by the watcher). It can differ from route.Peer and change on
+	// failover, so it is recorded on AddAllowedIPs and used on RemoveAllowedIPs to decrement
+	// the exact peer that was incremented.
+	currentPeerKey string
 }
 
 func NewRoute(params common.HandlerParams) *Route {
@@ -52,12 +57,15 @@ func (r *Route) AddAllowedIPs(peerKey string) error {
 			ref.Out,
 		)
 	}
+	r.currentPeerKey = peerKey
 	return nil
 }
 
 func (r *Route) RemoveAllowedIPs() error {
-	if _, err := r.allowedIPsRefcounter.Decrement(r.route.Network); err != nil {
-		return err
+	var err error
+	if _, decErr := r.allowedIPsRefcounter.Decrement(r.route.Network, r.currentPeerKey); decErr != nil {
+		err = fmt.Errorf("remove allowed IP %s: %w", r.route.Network, decErr)
 	}
-	return nil
+	r.currentPeerKey = ""
+	return err
 }

client/internal/routemanager/sysctl/sysctl_linux.go

@@ -20,6 +20,8 @@ const (
 	rpFilterPath          = "net.ipv4.conf.all.rp_filter"
 	rpFilterInterfacePath = "net.ipv4.conf.%s.rp_filter"
 	srcValidMarkPath      = "net.ipv4.conf.all.src_valid_mark"
+	percentEscape         = "%25"
+	dotEscape             = "%2E"
 )
 
 type iface interface {
@@ -56,7 +58,11 @@ func Setup(wgIface iface) (map[string]int, error) {
 			continue
 		}
 
-		i := fmt.Sprintf(rpFilterInterfacePath, intf.Name)
+		// Escape '%' and '.' so they survive the dot-to-slash conversion in Set()
+		safeName := strings.ReplaceAll(intf.Name, "%", percentEscape)
+		safeName = strings.ReplaceAll(safeName, ".", dotEscape)
+
+		i := fmt.Sprintf(rpFilterInterfacePath, safeName)
 		oldVal, err := Set(i, 2, true)
 		if err != nil {
 			result = multierror.Append(result, err)
@@ -70,7 +76,11 @@ func Setup(wgIface iface) (map[string]int, error) {
 
 // Set sets a sysctl configuration, if onlyIfOne is true it will only set the new value if it's set to 1
 func Set(key string, desiredValue int, onlyIfOne bool) (int, error) {
-	path := fmt.Sprintf("/proc/sys/%s", strings.ReplaceAll(key, ".", "/"))
+	path := strings.ReplaceAll(key, ".", "/")
+	// Unescape interface dots and percent signs
+	path = strings.ReplaceAll(path, dotEscape, ".")
+	path = strings.ReplaceAll(path, percentEscape, "%")
+	path = fmt.Sprintf("/proc/sys/%s", path)
 	currentValue, err := os.ReadFile(path)
 	if err != nil {
 		return -1, fmt.Errorf("read sysctl %s: %w", key, err)

client/internal/tunnelnotifier/notifier.go

@@ -0,0 +1,124 @@
+package tunnelnotifier
+
+import (
+	"container/list"
+	"sync"
+
+	"github.com/netbirdio/netbird/client/internal/dns"
+	"github.com/netbirdio/netbird/client/internal/listener"
+)
+
+type eventKind int
+
+const (
+	eventRoutes eventKind = iota
+	eventIfaceIP
+	eventIfaceIPv6
+	eventDNS
+)
+
+var (
+	_ listener.NetworkChangeListener = (*Notifier)(nil)
+	_ dns.IosDnsManager              = (*Notifier)(nil)
+)
+
+type event struct {
+	kind    eventKind
+	payload string
+}
+
+type Notifier struct {
+	mu     sync.Mutex
+	cond   *sync.Cond
+	queue  *list.List
+	closed bool
+	done   chan struct{}
+
+	listener   listener.NetworkChangeListener
+	dnsManager dns.IosDnsManager
+}
+
+func New(l listener.NetworkChangeListener, dm dns.IosDnsManager) *Notifier {
+	n := &Notifier{
+		queue:      list.New(),
+		done:       make(chan struct{}),
+		listener:   l,
+		dnsManager: dm,
+	}
+	n.cond = sync.NewCond(&n.mu)
+	go n.deliverLoop()
+	return n
+}
+
+func (n *Notifier) OnNetworkChanged(routes string) {
+	n.enqueue(event{kind: eventRoutes, payload: routes})
+}
+
+func (n *Notifier) SetInterfaceIP(ip string) {
+	n.enqueue(event{kind: eventIfaceIP, payload: ip})
+}
+
+func (n *Notifier) SetInterfaceIPv6(ip string) {
+	n.enqueue(event{kind: eventIfaceIPv6, payload: ip})
+}
+
+func (n *Notifier) ApplyDns(config string) {
+	n.enqueue(event{kind: eventDNS, payload: config})
+}
+
+// Close stops accepting new events and blocks until the delivery loop has
+// drained all queued events and exited.
+func (n *Notifier) Close() {
+	n.mu.Lock()
+	n.closed = true
+	n.cond.Signal()
+	n.mu.Unlock()
+	<-n.done
+}
+
+func (n *Notifier) enqueue(ev event) {
+	n.mu.Lock()
+	defer n.mu.Unlock()
+	if n.closed {
+		return
+	}
+	n.queue.PushBack(ev)
+	n.cond.Signal()
+}
+
+func (n *Notifier) deliverLoop() {
+	defer close(n.done)
+	for {
+		n.mu.Lock()
+		for n.queue.Len() == 0 && !n.closed {
+			n.cond.Wait()
+		}
+		if n.closed && n.queue.Len() == 0 {
+			n.mu.Unlock()
+			return
+		}
+		ev := n.queue.Remove(n.queue.Front()).(event)
+		l := n.listener
+		dm := n.dnsManager
+		n.mu.Unlock()
+
+		switch ev.kind {
+		case eventRoutes:
+			if l != nil {
+				l.OnNetworkChanged(ev.payload)
+			}
+		case eventIfaceIP:
+			if l != nil {
+				l.SetInterfaceIP(ev.payload)
+			}
+		case eventIfaceIPv6:
+			if l != nil {
+				l.SetInterfaceIPv6(ev.payload)
+			}
+		case eventDNS:
+			if dm != nil {
+				dm.ApplyDns(ev.payload)
+			}
+		}
+	}
+}

client/internal/tunnelnotifier/notifier_test.go

@@ -0,0 +1,192 @@
+package tunnelnotifier
+
+import (
+	"fmt"
+	"sync"
+	"sync/atomic"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+type call struct {
+	kind    string
+	payload string
+}
+
+type recorder struct {
+	mu       sync.Mutex
+	calls    []call
+	inFlight atomic.Int32
+	overlap  atomic.Bool
+	delay    time.Duration
+}
+
+func (r *recorder) record(kind, payload string) {
+	if r.inFlight.Add(1) != 1 {
+		r.overlap.Store(true)
+	}
+	if r.delay > 0 {
+		time.Sleep(r.delay)
+	}
+	r.mu.Lock()
+	r.calls = append(r.calls, call{kind: kind, payload: payload})
+	r.mu.Unlock()
+	r.inFlight.Add(-1)
+}
+
+func (r *recorder) count() int {
+	r.mu.Lock()
+	defer r.mu.Unlock()
+	return len(r.calls)
+}
+
+func (r *recorder) snapshot() []call {
+	r.mu.Lock()
+	defer r.mu.Unlock()
+	out := make([]call, len(r.calls))
+	copy(out, r.calls)
+	return out
+}
+
+type fakeListener struct {
+	rec *recorder
+}
+
+func (f *fakeListener) OnNetworkChanged(routes string) {
+	f.rec.record("routes", routes)
+}
+
+func (f *fakeListener) SetInterfaceIP(ip string) {
+	f.rec.record("ip", ip)
+}
+
+func (f *fakeListener) SetInterfaceIPv6(ip string) {
+	f.rec.record("ipv6", ip)
+}
+
+type fakeDNSManager struct {
+	rec *recorder
+}
+
+func (f *fakeDNSManager) ApplyDns(config string) {
+	f.rec.record("dns", config)
+}
+
+func TestFIFOOrder(t *testing.T) {
+	rec := &recorder{}
+	n := New(&fakeListener{rec: rec}, &fakeDNSManager{rec: rec})
+	defer n.Close()
+
+	n.SetInterfaceIP("10.0.0.1")
+	n.SetInterfaceIPv6("fd00::1")
+	n.ApplyDns(`{"domains":[]}`)
+	n.OnNetworkChanged("10.0.0.0/8,192.168.0.0/16")
+	n.ApplyDns(`{"domains":["example.com"]}`)
+
+	require.Eventually(t, func() bool { return rec.count() == 5 }, time.Second, time.Millisecond)
+
+	expected := []call{
+		{kind: "ip", payload: "10.0.0.1"},
+		{kind: "ipv6", payload: "fd00::1"},
+		{kind: "dns", payload: `{"domains":[]}`},
+		{kind: "routes", payload: "10.0.0.0/8,192.168.0.0/16"},
+		{kind: "dns", payload: `{"domains":["example.com"]}`},
+	}
+	assert.Equal(t, expected, rec.snapshot())
+}
+
+func TestNoOverlappingCalls(t *testing.T) {
+	rec := &recorder{delay: 100 * time.Microsecond}
+	n := New(&fakeListener{rec: rec}, &fakeDNSManager{rec: rec})
+	defer n.Close()
+
+	const producers = 8
+	const perProducer = 25
+
+	var wg sync.WaitGroup
+	for i := 0; i < producers; i++ {
+		wg.Add(1)
+		go func(id int) {
+			defer wg.Done()
+			for j := 0; j < perProducer; j++ {
+				payload := fmt.Sprintf("%d-%d", id, j)
+				switch j % 4 {
+				case 0:
+					n.OnNetworkChanged(payload)
+				case 1:
+					n.SetInterfaceIP(payload)
+				case 2:
+					n.SetInterfaceIPv6(payload)
+				case 3:
+					n.ApplyDns(payload)
+				}
+			}
+		}(i)
+	}
+	wg.Wait()
+
+	require.Eventually(t, func() bool { return rec.count() == producers*perProducer }, 5*time.Second, time.Millisecond)
+	assert.False(t, rec.overlap.Load())
+}
+
+func TestDNSAndRoutesInterleaved(t *testing.T) {
+	rec := &recorder{delay: 100 * time.Microsecond}
+	n := New(&fakeListener{rec: rec}, &fakeDNSManager{rec: rec})
+	defer n.Close()
+
+	const events = 50
+
+	var wg sync.WaitGroup
+	wg.Add(2)
+	go func() {
+		defer wg.Done()
+		for i := 0; i < events; i++ {
+			n.ApplyDns(fmt.Sprintf("dns-%d", i))
+		}
+	}()
+	go func() {
+		defer wg.Done()
+		for i := 0; i < events; i++ {
+			n.OnNetworkChanged(fmt.Sprintf("routes-%d", i))
+		}
+	}()
+	wg.Wait()
+
+	require.Eventually(t, func() bool { return rec.count() == 2*events }, 5*time.Second, time.Millisecond)
+	assert.False(t, rec.overlap.Load())
+
+	var dnsSeen, routesSeen int
+	for _, c := range rec.snapshot() {
+		switch c.kind {
+		case "dns":
+			assert.Equal(t, fmt.Sprintf("dns-%d", dnsSeen), c.payload)
+			dnsSeen++
+		case "routes":
+			assert.Equal(t, fmt.Sprintf("routes-%d", routesSeen), c.payload)
+			routesSeen++
+		}
+	}
+	assert.Equal(t, events, dnsSeen)
+	assert.Equal(t, events, routesSeen)
+}
+
+func TestCloseDrainsQueue(t *testing.T) {
+	rec := &recorder{delay: time.Millisecond}
+	n := New(&fakeListener{rec: rec}, &fakeDNSManager{rec: rec})
+
+	const events = 20
+	for i := 0; i < events; i++ {
+		n.OnNetworkChanged(fmt.Sprintf("routes-%d", i))
+	}
+	n.Close()
+
+	require.Equal(t, events, rec.count(), "Close must not return before all queued events are delivered")
+
+	n.OnNetworkChanged("after-close")
+	n.ApplyDns("after-close")
+	time.Sleep(50 * time.Millisecond)
+	assert.Equal(t, events, rec.count())
+}

client/ios/NetBirdSDK/version.go

@@ -0,0 +1,12 @@
+//go:build ios
+
+package NetBirdSDK
+
+import "github.com/netbirdio/netbird/version"
+
+// GoClientVersion returns the NetBird Go client version that was baked into
+// the framework at compile time via
+// -ldflags "-X github.com/netbirdio/netbird/version.version=<version>".
+func GoClientVersion() string {
+	return version.NetbirdVersion()
+}

client/ui/frontend/src/modules/auto-update/UpdateVersionCard.tsx

@@ -2,6 +2,7 @@ import { type ReactNode } from "react";
 import { useTranslation } from "react-i18next";
 import { Browser } from "@wailsio/runtime";
 import { DownloadIcon, NotepadText } from "lucide-react";
+import { Update as UpdateSvc } from "@bindings/services";
 import { Button } from "@/components/buttons/Button";
 import { useClientVersion } from "@/contexts/ClientVersionContext";
 import { cn } from "@/lib/cn";
@@ -14,6 +15,12 @@ function openUrl(url: string) {
     });
 }
 
+function openInstallerDownload() {
+    UpdateSvc.DownloadURL()
+        .then(openUrl)
+        .catch(() => openUrl(GITHUB_RELEASES));
+}
+
 export function UpdateVersionCard() {
     const { t } = useTranslation();
     const { updateVersion, enforced, triggerUpdate } = useClientVersion();
@@ -37,11 +44,7 @@ export function UpdateVersionCard() {
                         {t("update.card.installNow")}
                     </Button>
                 ) : (
-                    <Button
-                        variant={"primary"}
-                        size={"xs"}
-                        onClick={() => openUrl(GITHUB_RELEASES)}
-                    >
+                    <Button variant={"primary"} size={"xs"} onClick={openInstallerDownload}>
                         <DownloadIcon size={14} />
                         {t("update.card.getInstaller")}
                     </Button>

client/ui/main.go

@@ -281,6 +281,9 @@ func newApplication(onSecondInstance func()) *application.App {
 		Linux: application.LinuxOptions{
 			ProgramName: "netbird",
 		},
+		Windows: application.WindowsOptions{
+			WndProcInterceptor: endSessionInterceptor(),
+		},
 		SingleInstance: &application.SingleInstanceOptions{
 			UniqueID: "io.netbird.ui",
 			OnSecondInstanceLaunch: func(_ application.SecondInstanceData) {
@@ -367,6 +370,9 @@ func newMainWindow(app *application.App, prefStore *preferences.Store) *applicat
 
 	// Hide instead of quit on close; "really quit" is reached via tray -> Quit.
 	window.RegisterHook(events.Common.WindowClosing, func(e *application.WindowEvent) {
+		if services.ShuttingDown() {
+			return
+		}
 		e.Cancel()
 		window.Hide()
 	})

client/ui/services/shutdown.go

@@ -0,0 +1,24 @@
+package services
+
+import "sync/atomic"
+
+var (
+	sessionEnding atomic.Bool
+	quitting      atomic.Bool
+)
+
+func BeginSessionEnd() {
+	sessionEnding.Store(true)
+}
+
+func AbortSessionEnd() {
+	sessionEnding.Store(false)
+}
+
+func BeginShutdown() {
+	quitting.Store(true)
+}
+
+func ShuttingDown() bool {
+	return sessionEnding.Load() || quitting.Load()
+}

client/ui/services/update.go

@@ -10,6 +10,7 @@ import (
 
 	"github.com/netbirdio/netbird/client/proto"
 	"github.com/netbirdio/netbird/client/ui/updater"
+	"github.com/netbirdio/netbird/version"
 )
 
 // UpdateResult mirrors TriggerUpdateResponse.
@@ -33,6 +34,12 @@ func (s *Update) GetState() updater.State {
 	return s.holder.Get()
 }
 
+// DownloadURL returns the platform-appropriate installer download link for
+// manual (non-enforced) updates.
+func (s *Update) DownloadURL() string {
+	return version.DownloadUrl()
+}
+
 // Quit exits the app. Scheduled off the calling goroutine so the JS caller's
 // response returns before the runtime tears down.
 func (s *Update) Quit() {

client/ui/services/windowmanager.go

@@ -154,6 +154,9 @@ func NewWindowManager(app *application.App, mainWindow *application.WebviewWindo
 	})
 	// Hide (not destroy) on close to keep React state; reset to General for a flash-free reopen.
 	s.settings.RegisterHook(events.Common.WindowClosing, func(e *application.WindowEvent) {
+		if ShuttingDown() {
+			return
+		}
 		e.Cancel()
 		s.app.Event.Emit(EventSettingsOpen, "general")
 		s.settings.Hide()
@@ -393,6 +396,9 @@ func (s *WindowManager) CloseWelcome() {
 // OpenError shows the custom error dialog; title/message are pre-localised and ride in the
 // start URL. A second error replaces the open one via SetURL. Singleton, destroyed on close.
 func (s *WindowManager) OpenError(title, message string) {
+	if ShuttingDown() {
+		return
+	}
 	s.mu.Lock()
 	defer s.mu.Unlock()
 	startURL := errorDialogURL(title, message)

client/ui/shutdown_other.go

@@ -0,0 +1,7 @@
+//go:build !windows && !android && !ios && !freebsd && !js
+
+package main
+
+func endSessionInterceptor() func(hwnd uintptr, msg uint32, wParam, lParam uintptr) (uintptr, bool) {
+	return nil
+}

client/ui/shutdown_windows.go

@@ -0,0 +1,36 @@
+//go:build windows
+
+package main
+
+import (
+	"os"
+
+	log "github.com/sirupsen/logrus"
+
+	"github.com/netbirdio/netbird/client/ui/services"
+)
+
+const (
+	wmQueryEndSession = 0x0011
+	wmEndSession      = 0x0016
+)
+
+func endSessionInterceptor() func(hwnd uintptr, msg uint32, wParam, lParam uintptr) (uintptr, bool) {
+	return func(_ uintptr, msg uint32, wParam, _ uintptr) (uintptr, bool) {
+		switch msg {
+		case wmQueryEndSession:
+			services.BeginSessionEnd()
+			return 1, true
+		case wmEndSession:
+			if wParam == 0 {
+				services.AbortSessionEnd()
+				return 0, true
+			}
+			log.Info("windows session is ending; exiting immediately")
+			os.Exit(0)
+			return 0, true
+		default:
+			return 0, false
+		}
+	}
+}

client/ui/tray.go

@@ -32,9 +32,8 @@ const (
 
 	quitDownTimeout = 5 * time.Second
 
-	urlGitHubRepo     = "https://github.com/netbirdio/netbird"
-	urlGitHubReleases = "https://github.com/netbirdio/netbird/releases/latest"
-	urlDocs           = "https://docs.netbird.io"
+	urlGitHubRepo = "https://github.com/netbirdio/netbird"
+	urlDocs       = "https://docs.netbird.io"
 )
 
 // TrayServices bundles the services the tray menu needs, grouped so NewTray
@@ -453,6 +452,7 @@ func (t *Tray) buildMenu() *application.Menu {
 }
 
 func (t *Tray) handleQuit() {
+	services.BeginShutdown()
 	t.profileMu.Lock()
 	if t.switchCancel != nil {
 		t.switchCancel()

client/ui/tray_notify.go

@@ -25,6 +25,9 @@ type sendFn func(notifications.NotificationOptions) error
 // event-dispatch goroutine that panic is fatal process-wide; recover() turns
 // it into a logged no-op.
 func safeSendNotification(send sendFn, what string, opts notifications.NotificationOptions) (err error) {
+	if services.ShuttingDown() {
+		return nil
+	}
 	defer func() {
 		if r := recover(); r != nil {
 			log.Errorf("notify %s: recovered from panic (notification bus unavailable): %v", what, r)

client/ui/tray_update.go

@@ -13,6 +13,7 @@ import (
 
 	"github.com/netbirdio/netbird/client/ui/services"
 	"github.com/netbirdio/netbird/client/ui/updater"
+	"github.com/netbirdio/netbird/version"
 )
 
 // trayUpdater owns the tray UI that reacts to auto-update. Composed inside Tray.
@@ -76,15 +77,15 @@ func (u *trayUpdater) applyLanguage() {
 	u.refreshMenuItem(state)
 }
 
-// handleClick opens the GitHub releases page when not Enforced, otherwise shows
-// the progress page and asks the daemon to start the installer.
+// handleClick opens the installer download link when not Enforced, otherwise
+// shows the progress page and asks the daemon to start the installer.
 func (u *trayUpdater) handleClick() {
 	u.mu.Lock()
 	state := u.state
 	u.mu.Unlock()
 
 	if !state.Enforced {
-		_ = u.app.Browser.OpenURL(urlGitHubReleases)
+		_ = u.app.Browser.OpenURL(version.DownloadUrl())
 		return
 	}
 

docs/agent-networks/01-end-to-end-flows.md

@@ -109,7 +109,7 @@ sequenceDiagram
         Chk->>Inj: continue
         Inj->>Inj: inject NetBird identity headers per provider config
         Inj->>Grd: continue
-        Grd->>Grd: enforce model allowlist
+        Grd->>Grd: enforce per-provider allowlist (fail-closed backstop)
         Grd->>Up: forward (over WireGuard)
         Up-->>Resp: response (JSON or SSE stream)
         Resp->>Resp: parse usage tokens, completion
@@ -135,6 +135,21 @@ sequenceDiagram
   (`redact_pii = settings.RedactPii`). Phones, emails, credit cards,
   PII names — see `redact.go` for the full set. See
   [`modules/31-proxy-middleware-builtin.md`](modules/31-proxy-middleware-builtin.md).
+- The model allowlist is enforced in TWO places. `CheckLLMPolicyLimits`
+  is authoritative: it resolves the policy that governs this
+  (provider, caller-groups) and denies (`deny_code = llm_policy.model_blocked`)
+  when no applicable policy permits the model — so an allowlist scoped to
+  one group/provider never leaks to another, and an un-guardrailed policy
+  is genuinely unrestricted. `llm_guardrail` is a per-provider fail-closed
+  backstop: it only carries an allowlist for a provider every authorising
+  policy restricts, and blocks unknown/undetermined models even when
+  management is unreachable. Because that backstop allowlist is the UNION
+  of every restricting policy's models, per-group narrowing lives only in
+  the authoritative check: during a `CheckLLMPolicyLimits` outage
+  `llm_limit_check` fails open, so a caller can reach any model in the
+  provider's union — a group scoped to model A could reach model B if
+  another group restricts the same provider to B. This is the documented
+  fail-open trade-off; a future flag may switch it to fail-closed.
 - SSE streaming requires special handling on the response side; the
   parser must handle partial chunks without buffering the whole
   stream. See [`modules/32-proxy-llm-parsers.md`](modules/32-proxy-llm-parsers.md).

docs/agent-networks/modules/21-management-agentnetwork.md

@@ -122,7 +122,7 @@ At request time the path is independent: the proxy calls `SelectPolicyForRequest
   | on_request | 1 | `llm_router` | `{"providers":[{id, models[], upstream_*, auth_header_*, allowed_group_ids[]}]}` | **true** |
   | on_request | 2 | `llm_limit_check` | `{}` | – |
   | on_request | 3 | `llm_identity_inject` | `{"providers":[{provider_id, header_pair?, json_metadata?, extra_headers?}]}` | **true** |
-  | on_request | 4 | `llm_guardrail` | `{"model_allowlist"?, "prompt_capture":{enabled,redact_pii}}` | – |
+  | on_request | 4 | `llm_guardrail` | `{"provider_allowlists"?: {providerID: []model}, "prompt_capture":{enabled,redact_pii}}` | – |
   | on_response | 5 | `llm_limit_record` | `{}` (runs LAST at runtime) | – |
   | on_response | 6 | `cost_meter` | `{}` | – |
   | on_response | 7 | `llm_response_parser` | `{"capture_completion": <bool>, "redact_pii"?: true}` | – |

docs/agent-networks/modules/31-proxy-middleware-builtin.md

@@ -244,7 +244,7 @@ no mocks. Tests: `TestChain_AllowPath_StampsAttributionAndRecordsCounter`
 | `llm_router` | `{providers: [{id, models, upstream_scheme, upstream_host, upstream_path?, auth_header_name, auth_header_value, allowed_group_ids}]}` |
 | `llm_limit_check` | `{}` — pulls `MgmtClient` from `FactoryContext` |
 | `llm_identity_inject` | `{providers: [{provider_id, header_pair?|json_metadata?, extra_headers?}]}` |
-| `llm_guardrail` | `{model_allowlist: []string, prompt_capture: {enabled, redact_pii}}` |
+| `llm_guardrail` | `{provider_allowlists: {providerID: []string}, prompt_capture: {enabled, redact_pii}}` — allowlist keyed by resolved provider id; a provider absent from the map is unrestricted (fail-closed backstop; authoritative per-policy/group check is management's `CheckLLMPolicyLimits`) |
 | `llm_response_parser` | `{redact_pii?, capture_completion?: *bool}` |
 | `cost_meter` | `{pricing_path?}` (basename inside data-dir; defaults `pricing.yaml`) |
 | `llm_limit_record` | `{}` — same pattern as `llm_limit_check` |

e2e/agentnetwork/guardrail_block_test.go

@@ -0,0 +1,209 @@
+//go:build e2e
+
+package agentnetwork
+
+import (
+	"context"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	"github.com/netbirdio/netbird/e2e/harness"
+	"github.com/netbirdio/netbird/shared/management/http/api"
+)
+
+// pathRoutedGuardrailCase is one provider's self-contained scenario: its own
+// provider, its own guardrail whose allowlist holds ONLY that provider's
+// allowed model, and its own policy. Each case runs in isolation (its own
+// proxy + client), so the guardrail the proxy enforces contains exactly this
+// provider's model — never a mixed cross-provider list.
+type pathRoutedGuardrailCase struct {
+	name       string
+	catalogID  string // agent-network catalog provider id
+	wire       string // harness.WireVertex | harness.WireBedrock
+	allowEntry string // the single model id put on the guardrail allowlist
+	allowModel string // model id sent that MUST be served (200)
+	blockModel string // model id sent that MUST be denied (403 model_blocked)
+}
+
+// TestGuardrailBlocksUnselectedModel_PathRouted is the end-to-end regression
+// guard for the customer report that a model-allowlist guardrail attached to a
+// policy has no effect for PATH-ROUTED providers — where the model travels in
+// the URL, not the JSON body: Google Vertex (…/models/{model}:rawPredict) and
+// AWS Bedrock (/model/{id}/invoke).
+//
+// Each provider is tested in isolation with a guardrail allowlisting a single
+// model of its own: the allowed model (in the URL path) is served (200) and an
+// unselected model (in the URL path) is denied 403 by the guardrail
+// (llm_policy.model_blocked) before the upstream. The Vertex case mirrors the
+// customer verbatim — allow Sonnet, and the unselected model is the exact
+// claude-opus-4-6 they reported reaching the model unblocked. The Bedrock case
+// sends a region-prefixed, versioned inference-profile id so URL-path model
+// normalization is exercised too.
+//
+// The provider is catch-all (no models), so the router forwards any model and a
+// 403 can only come from the guardrail, never model_not_routable. Only the
+// upstream LLM is mocked (the vLLM nginx answers any path with 200); management
+// synth/reconcile, the proxy middleware chain (URL-path model extraction,
+// router, guardrail) and the tunnel are all real, and the guardrail denies
+// before the upstream is dialed so the mock cannot influence the block. A
+// static bearer api key is used so the router injects a static Authorization
+// header instead of minting a GCP token — the only reason path-routed providers
+// normally need live credentials — so the test runs with none and is always on.
+func TestGuardrailBlocksUnselectedModel_PathRouted(t *testing.T) {
+	cases := []pathRoutedGuardrailCase{
+		{
+			name:       "vertex",
+			catalogID:  "vertex_ai_api",
+			wire:       harness.WireVertex,
+			allowEntry: "claude-sonnet-4-5",
+			allowModel: "claude-sonnet-4-5",
+			blockModel: "claude-opus-4-6", // the customer-reported model
+		},
+		{
+			name:       "bedrock",
+			catalogID:  "bedrock_api",
+			wire:       harness.WireBedrock,
+			allowEntry: "anthropic.claude-sonnet-4-5", // normalized catalog id
+			allowModel: "us.anthropic.claude-sonnet-4-5-v1:0",
+			blockModel: "us.anthropic.claude-opus-4-8-v1:0",
+		},
+	}
+
+	for _, tc := range cases {
+		tc := tc
+		t.Run(tc.name, func(t *testing.T) {
+			runPathRoutedGuardrailCase(t, tc)
+		})
+	}
+}
+
+func runPathRoutedGuardrailCase(t *testing.T, tc pathRoutedGuardrailCase) {
+	t.Helper()
+
+	const (
+		vertexProject = "e2e-project"
+		vertexRegion  = "global"
+	)
+
+	ctx, cancel := context.WithTimeout(context.Background(), 15*time.Minute)
+	defer cancel()
+
+	vllm, err := harness.StartVLLM(ctx, srv)
+	require.NoError(t, err, "start mock upstream")
+	t.Cleanup(func() { _ = vllm.Terminate(context.Background()) })
+
+	grp, err := srv.API().Groups.Create(ctx, api.PostApiGroupsJSONRequestBody{Name: "e2e-guardrail-" + tc.name})
+	require.NoError(t, err, "create group")
+	t.Cleanup(func() { _ = srv.API().Groups.Delete(context.Background(), grp.Id) })
+
+	ephemeral := false
+	sk, err := srv.API().SetupKeys.Create(ctx, api.PostApiSetupKeysJSONRequestBody{
+		Name:       "e2e-guardrail-" + tc.name + "-client",
+		Type:       "reusable",
+		ExpiresIn:  86400,
+		UsageLimit: 0,
+		AutoGroups: []string{grp.Id},
+		Ephemeral:  &ephemeral,
+	})
+	require.NoError(t, err, "mint setup key")
+	require.NotEmpty(t, sk.Key, "setup key plaintext")
+
+	// Catch-all provider (no models) so the router forwards any model; a static
+	// bearer key means the router injects a static auth header instead of minting
+	// a GCP token. Bootstraps the cluster if it isn't already.
+	staticKey := "static-e2e-token"
+	prov, err := srv.CreateProvider(ctx, api.AgentNetworkProviderRequest{
+		Name:             tc.name,
+		ProviderId:       tc.catalogID,
+		UpstreamUrl:      vllm.URL,
+		ApiKey:           &staticKey,
+		Enabled:          ptr(true),
+		BootstrapCluster: ptr(harness.AgentNetworkCluster),
+	})
+	require.NoError(t, err, "create %s provider", tc.name)
+	t.Cleanup(func() { _ = srv.DeleteProvider(context.Background(), prov.Id) })
+
+	// Guardrail allowlisting ONLY this provider's allowed model.
+	var gr api.AgentNetworkGuardrailRequest
+	gr.Name = "e2e-guardrail-" + tc.name
+	gr.Checks.ModelAllowlist.Enabled = true
+	gr.Checks.ModelAllowlist.Models = []string{tc.allowEntry}
+	guard, err := srv.CreateGuardrail(ctx, gr)
+	require.NoError(t, err, "create guardrail")
+	t.Cleanup(func() { _ = srv.DeleteGuardrail(context.Background(), guard.Id) })
+
+	enabled := true
+	pol, err := srv.CreatePolicy(ctx, api.AgentNetworkPolicyRequest{
+		Name:                   "e2e-guardrail-" + tc.name,
+		Enabled:                &enabled,
+		SourceGroups:           []string{grp.Id},
+		DestinationProviderIds: []string{prov.Id},
+		GuardrailIds:           &[]string{guard.Id},
+	})
+	require.NoError(t, err, "create policy")
+	t.Cleanup(func() { _ = srv.DeletePolicy(context.Background(), pol.Id) })
+
+	settings, err := srv.GetSettings(ctx)
+	require.NoError(t, err, "read settings")
+	require.NotEmpty(t, settings.Endpoint, "endpoint must be assigned")
+
+	proxyToken, err := srv.CreateProxyTokenCLI(ctx, "e2e-guardrail-"+tc.name+"-proxy")
+	require.NoError(t, err, "mint proxy token")
+	px, err := harness.StartProxy(ctx, srv, proxyToken)
+	require.NoError(t, err, "start proxy")
+	t.Cleanup(func() { _ = px.Terminate(context.Background()) })
+
+	cl, err := harness.StartClient(ctx, srv, sk.Key)
+	require.NoError(t, err, "start client")
+	t.Cleanup(func() { _ = cl.Terminate(context.Background()) })
+
+	require.NoError(t, cl.WaitConnected(ctx, 90*time.Second), "client must connect to management")
+	// Probe first: the GET resolves the endpoint and its first packet wakes the
+	// lazy proxy peer, so WaitProxyPeer then observes it connected.
+	proxyIP, err := cl.ResolveProxyIP(ctx, settings.Endpoint)
+	require.NoError(t, err, "resolve endpoint to proxy IP")
+	if err := cl.WaitProxyPeer(ctx, 180*time.Second); err != nil {
+		t.Fatalf("client did not see the proxy peer: %v\n=== proxy logs ===\n%s", err, px.Logs(context.Background()))
+	}
+
+	send := func(model string) (int, string) {
+		var code int
+		var body string
+		var cerr error
+		switch tc.wire {
+		case harness.WireVertex:
+			code, body, cerr = cl.Vertex(ctx, settings.Endpoint, proxyIP, vertexProject, vertexRegion, model, "Reply with exactly: pong", "")
+		case harness.WireBedrock:
+			code, body, cerr = cl.Bedrock(ctx, settings.Endpoint, proxyIP, model, "Reply with exactly: pong", "")
+		default:
+			t.Fatalf("unsupported wire %q", tc.wire)
+		}
+		require.NoError(t, cerr, "request must reach the proxy for %s", tc.name)
+		return code, body
+	}
+
+	// Allowed model (in the URL path) is served. Retry to absorb tunnel/DNS
+	// jitter on the first call over the freshly warmed tunnel.
+	var code int
+	var body string
+	deadline := time.Now().Add(90 * time.Second)
+	for time.Now().Before(deadline) {
+		code, body = send(tc.allowModel)
+		if code == 200 {
+			break
+		}
+		time.Sleep(5 * time.Second)
+	}
+	assert.Equal(t, 200, code,
+		"allowed %s model (URL path) must be served; body: %s\n=== proxy logs ===\n%s", tc.name, body, px.Logs(context.Background()))
+
+	// Unselected model (in the URL path) must be blocked by the guardrail.
+	code, body = send(tc.blockModel)
+	assert.Equal(t, 403, code,
+		"unselected %s model (URL path) must be denied, not served; body: %s\n=== proxy logs ===\n%s", tc.name, body, px.Logs(context.Background()))
+	assert.Contains(t, body, "llm_policy.model_blocked",
+		"%s denial must come from the guardrail allowlist, not routing; body: %s", tc.name, body)
+}

e2e/agentnetwork/guardrail_groupswitch_test.go

@@ -0,0 +1,205 @@
+//go:build e2e
+
+package agentnetwork
+
+import (
+	"context"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	"github.com/netbirdio/netbird/e2e/harness"
+	"github.com/netbirdio/netbird/shared/management/http/api"
+)
+
+// TestGuardrailGroupSwitchTakesEffectAfterTTL proves that moving a peer between
+// groups flips its model-allowlist decision once the proxy's tunnel-peer cache
+// expires. The peer's groups reach the guardrail via ValidateTunnelPeer, which
+// the proxy caches; the switch is invisible until that cache expires. The proxy
+// runs with a short NB_PROXY_TUNNEL_CACHE_TTL so the flip happens in seconds
+// instead of the 5-minute default.
+//
+// Setup: one catch-all provider declaring modelA + modelB; polA (grpA -> allow
+// modelA) and polB (grpB -> allow modelB). The client starts in grpA. modelA is
+// served and modelB denied; after switching the client grpA -> grpB, modelB is
+// served and modelA denied. The cross-group deny comes from management's
+// per-policy/group CheckLLMPolicyLimits (the proxy backstop carries the union).
+func TestGuardrailGroupSwitchTakesEffectAfterTTL(t *testing.T) {
+	ctx, cancel := context.WithTimeout(context.Background(), 15*time.Minute)
+	defer cancel()
+
+	const (
+		modelA = "e2e-model-a"
+		modelB = "e2e-model-b"
+		// Short tunnel-cache TTL so a group switch propagates in seconds.
+		// Exercises the NB_PROXY_TUNNEL_CACHE_TTL override.
+		cacheTTL = 3 * time.Second
+	)
+
+	vllm, err := harness.StartVLLM(ctx, srv)
+	require.NoError(t, err, "start mock upstream")
+	t.Cleanup(func() { _ = vllm.Terminate(context.Background()) })
+
+	grpA, err := srv.API().Groups.Create(ctx, api.PostApiGroupsJSONRequestBody{Name: "e2e-gswitch-a"})
+	require.NoError(t, err, "create group A")
+	t.Cleanup(func() { _ = srv.API().Groups.Delete(context.Background(), grpA.Id) })
+
+	grpB, err := srv.API().Groups.Create(ctx, api.PostApiGroupsJSONRequestBody{Name: "e2e-gswitch-b"})
+	require.NoError(t, err, "create group B")
+	t.Cleanup(func() { _ = srv.API().Groups.Delete(context.Background(), grpB.Id) })
+
+	ephemeral := false
+	sk, err := srv.API().SetupKeys.Create(ctx, api.PostApiSetupKeysJSONRequestBody{
+		Name:       "e2e-gswitch-client",
+		Type:       "reusable",
+		ExpiresIn:  86400,
+		UsageLimit: 0,
+		AutoGroups: []string{grpA.Id}, // client starts in group A
+		Ephemeral:  &ephemeral,
+	})
+	require.NoError(t, err, "mint setup key")
+	require.NotEmpty(t, sk.Key, "setup key plaintext")
+
+	staticKey := "static-e2e-token"
+	prov, err := srv.CreateProvider(ctx, api.AgentNetworkProviderRequest{
+		Name:        "gswitch",
+		ProviderId:  "openai_api",
+		UpstreamUrl: vllm.URL,
+		ApiKey:      &staticKey,
+		Enabled:     ptr(true),
+		Models: &[]api.AgentNetworkProviderModel{
+			{Id: modelA, InputPer1k: 0.001, OutputPer1k: 0.001},
+			{Id: modelB, InputPer1k: 0.001, OutputPer1k: 0.001},
+		},
+		BootstrapCluster: ptr(harness.AgentNetworkCluster),
+	})
+	require.NoError(t, err, "create provider")
+	t.Cleanup(func() { _ = srv.DeleteProvider(context.Background(), prov.Id) })
+
+	mkGuard := func(name, model string) api.AgentNetworkGuardrail {
+		var gr api.AgentNetworkGuardrailRequest
+		gr.Name = name
+		gr.Checks.ModelAllowlist.Enabled = true
+		gr.Checks.ModelAllowlist.Models = []string{model}
+		g, gerr := srv.CreateGuardrail(ctx, gr)
+		require.NoError(t, gerr, "create guardrail %s", name)
+		t.Cleanup(func() { _ = srv.DeleteGuardrail(context.Background(), g.Id) })
+		return g
+	}
+	gA := mkGuard("e2e-gswitch-a", modelA)
+	gB := mkGuard("e2e-gswitch-b", modelB)
+
+	enabled := true
+	polA, err := srv.CreatePolicy(ctx, api.AgentNetworkPolicyRequest{
+		Name:                   "e2e-gswitch-a",
+		Enabled:                &enabled,
+		SourceGroups:           []string{grpA.Id},
+		DestinationProviderIds: []string{prov.Id},
+		GuardrailIds:           &[]string{gA.Id},
+	})
+	require.NoError(t, err, "create policy A")
+	t.Cleanup(func() { _ = srv.DeletePolicy(context.Background(), polA.Id) })
+
+	polB, err := srv.CreatePolicy(ctx, api.AgentNetworkPolicyRequest{
+		Name:                   "e2e-gswitch-b",
+		Enabled:                &enabled,
+		SourceGroups:           []string{grpB.Id},
+		DestinationProviderIds: []string{prov.Id},
+		GuardrailIds:           &[]string{gB.Id},
+	})
+	require.NoError(t, err, "create policy B")
+	t.Cleanup(func() { _ = srv.DeletePolicy(context.Background(), polB.Id) })
+
+	settings, err := srv.GetSettings(ctx)
+	require.NoError(t, err, "read settings")
+	require.NotEmpty(t, settings.Endpoint, "endpoint must be assigned")
+
+	proxyToken, err := srv.CreateProxyTokenCLI(ctx, "e2e-gswitch-proxy")
+	require.NoError(t, err, "mint proxy token")
+	px, err := harness.StartProxy(ctx, srv, proxyToken, map[string]string{
+		"NB_PROXY_TUNNEL_CACHE_TTL": cacheTTL.String(),
+	})
+	require.NoError(t, err, "start proxy")
+	t.Cleanup(func() { _ = px.Terminate(context.Background()) })
+
+	cl, err := harness.StartClient(ctx, srv, sk.Key)
+	require.NoError(t, err, "start client")
+	t.Cleanup(func() { _ = cl.Terminate(context.Background()) })
+
+	require.NoError(t, cl.WaitConnected(ctx, 90*time.Second), "client must connect to management")
+	proxyIP, err := cl.ResolveProxyIP(ctx, settings.Endpoint)
+	require.NoError(t, err, "resolve endpoint to proxy IP")
+	if err := cl.WaitProxyPeer(ctx, 180*time.Second); err != nil {
+		t.Fatalf("client did not see the proxy peer: %v\n=== proxy logs ===\n%s", err, px.Logs(context.Background()))
+	}
+
+	send := func(model string) (int, string) {
+		code, body, cerr := cl.Chat(ctx, settings.Endpoint, proxyIP, harness.WireChat, model, "Reply with exactly: pong", "")
+		require.NoError(t, cerr, "request must reach the proxy")
+		return code, body
+	}
+	sendUntil := func(model string, want int, timeout time.Duration) (int, string) {
+		var code int
+		var body string
+		deadline := time.Now().Add(timeout)
+		for time.Now().Before(deadline) {
+			code, body = send(model)
+			if code == want {
+				return code, body
+			}
+			time.Sleep(2 * time.Second)
+		}
+		return code, body
+	}
+
+	// Phase 1 — client is in group A: modelA served, modelB denied.
+	code, body := sendUntil(modelA, 200, 90*time.Second)
+	assert.Equal(t, 200, code,
+		"group-A model must be served while the client is in group A; body: %s\n=== proxy logs ===\n%s", body, px.Logs(context.Background()))
+	code, body = send(modelB)
+	assert.Equal(t, 403, code,
+		"group-B model must be denied while the client is in group A; body: %s", body)
+	assert.Contains(t, body, "llm_policy.model_blocked")
+
+	// Switch the client peer from group A to group B.
+	peerID := clientPeerInGroup(t, ctx, grpA.Id)
+	_, err = srv.API().Groups.Update(ctx, grpB.Id, api.PutApiGroupsGroupIdJSONRequestBody{
+		Name:  grpB.Name,
+		Peers: &[]string{peerID},
+	})
+	require.NoError(t, err, "add peer to group B")
+	_, err = srv.API().Groups.Update(ctx, grpA.Id, api.PutApiGroupsGroupIdJSONRequestBody{
+		Name:  grpA.Name,
+		Peers: &[]string{},
+	})
+	require.NoError(t, err, "remove peer from group A")
+
+	// Phase 2 — after the short TTL expires the proxy re-validates the peer,
+	// sees group B, and the decision flips. Poll to absorb TTL + re-validation.
+	code, body = sendUntil(modelB, 200, 60*time.Second)
+	assert.Equal(t, 200, code,
+		"after the group switch + TTL, the group-B model must be served; body: %s\n=== proxy logs ===\n%s", body, px.Logs(context.Background()))
+	code, body = send(modelA)
+	assert.Equal(t, 403, code,
+		"after the switch, the old group-A model must be denied; body: %s", body)
+	assert.Contains(t, body, "llm_policy.model_blocked")
+}
+
+// clientPeerInGroup returns the id of the single peer that is a member of the
+// given group — the test client. The proxy peer is never added to test groups.
+func clientPeerInGroup(t *testing.T, ctx context.Context, groupID string) string {
+	t.Helper()
+	peers, err := srv.API().Peers.List(ctx)
+	require.NoError(t, err, "list peers")
+	for _, p := range peers {
+		for _, g := range p.Groups {
+			if g.Id == groupID {
+				return p.Id
+			}
+		}
+	}
+	t.Fatalf("no peer found in group %s", groupID)
+	return ""
+}

e2e/agentnetwork/guardrail_multipolicy_test.go

@@ -0,0 +1,201 @@
+//go:build e2e
+
+package agentnetwork
+
+import (
+	"context"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	"github.com/netbirdio/netbird/e2e/harness"
+	"github.com/netbirdio/netbird/shared/management/http/api"
+)
+
+// TestGuardrailMultiPolicyModelAllowlist: modelSelected served (200), grpOther's
+// modelOther denied for the grpMain client (403 model_blocked, no cross-group
+// leak), and openModel on the un-guardrailed policy's provider served (200).
+func TestGuardrailMultiPolicyModelAllowlist(t *testing.T) {
+	ctx, cancel := context.WithTimeout(context.Background(), 15*time.Minute)
+	defer cancel()
+
+	const (
+		modelSelected = "e2e-selected"
+		modelOther    = "e2e-other"
+		openModel     = "e2e-open"
+	)
+
+	vllm, err := harness.StartVLLM(ctx, srv)
+	require.NoError(t, err, "start mock upstream")
+	t.Cleanup(func() { _ = vllm.Terminate(context.Background()) })
+
+	grpMain, err := srv.API().Groups.Create(ctx, api.PostApiGroupsJSONRequestBody{Name: "e2e-guardrail-mp-main"})
+	require.NoError(t, err, "create main group")
+	t.Cleanup(func() { _ = srv.API().Groups.Delete(context.Background(), grpMain.Id) })
+
+	grpOther, err := srv.API().Groups.Create(ctx, api.PostApiGroupsJSONRequestBody{Name: "e2e-guardrail-mp-other"})
+	require.NoError(t, err, "create other group")
+	t.Cleanup(func() { _ = srv.API().Groups.Delete(context.Background(), grpOther.Id) })
+
+	ephemeral := false
+	sk, err := srv.API().SetupKeys.Create(ctx, api.PostApiSetupKeysJSONRequestBody{
+		Name:       "e2e-guardrail-mp-client",
+		Type:       "reusable",
+		ExpiresIn:  86400,
+		UsageLimit: 0,
+		AutoGroups: []string{grpMain.Id}, // client joins grpMain only
+		Ephemeral:  &ephemeral,
+	})
+	require.NoError(t, err, "mint setup key")
+	require.NotEmpty(t, sk.Key, "setup key plaintext")
+
+	staticKey := "static-e2e-token"
+	models := func(ids ...string) *[]api.AgentNetworkProviderModel {
+		out := make([]api.AgentNetworkProviderModel, 0, len(ids))
+		for _, id := range ids {
+			out = append(out, api.AgentNetworkProviderModel{Id: id, InputPer1k: 0.001, OutputPer1k: 0.001})
+		}
+		return &out
+	}
+
+	// pRestricted declares the two guardrailed models so routing is deterministic
+	// (model -> provider). Created first, so it carries the bootstrap cluster.
+	pRestricted, err := srv.CreateProvider(ctx, api.AgentNetworkProviderRequest{
+		Name:             "restricted",
+		ProviderId:       "openai_api",
+		UpstreamUrl:      vllm.URL,
+		ApiKey:           &staticKey,
+		Enabled:          ptr(true),
+		Models:           models(modelSelected, modelOther),
+		BootstrapCluster: ptr(harness.AgentNetworkCluster),
+	})
+	require.NoError(t, err, "create restricted provider")
+	t.Cleanup(func() { _ = srv.DeleteProvider(context.Background(), pRestricted.Id) })
+
+	pOpen, err := srv.CreateProvider(ctx, api.AgentNetworkProviderRequest{
+		Name:        "open",
+		ProviderId:  "openai_api",
+		UpstreamUrl: vllm.URL,
+		ApiKey:      &staticKey,
+		Enabled:     ptr(true),
+		Models:      models(openModel),
+	})
+	require.NoError(t, err, "create open provider")
+	t.Cleanup(func() { _ = srv.DeleteProvider(context.Background(), pOpen.Id) })
+
+	mkGuardrail := func(name, model string) api.AgentNetworkGuardrail {
+		var gr api.AgentNetworkGuardrailRequest
+		gr.Name = name
+		gr.Checks.ModelAllowlist.Enabled = true
+		gr.Checks.ModelAllowlist.Models = []string{model}
+		g, gerr := srv.CreateGuardrail(ctx, gr)
+		require.NoError(t, gerr, "create guardrail %s", name)
+		t.Cleanup(func() { _ = srv.DeleteGuardrail(context.Background(), g.Id) })
+		return g
+	}
+	gMain := mkGuardrail("e2e-guardrail-mp-main", modelSelected)
+	gOther := mkGuardrail("e2e-guardrail-mp-other", modelOther)
+
+	enabled := true
+	// polMain: grpMain restricted to modelSelected on pRestricted.
+	polMain, err := srv.CreatePolicy(ctx, api.AgentNetworkPolicyRequest{
+		Name:                   "e2e-guardrail-mp-main",
+		Enabled:                &enabled,
+		SourceGroups:           []string{grpMain.Id},
+		DestinationProviderIds: []string{pRestricted.Id},
+		GuardrailIds:           &[]string{gMain.Id},
+	})
+	require.NoError(t, err, "create main policy")
+	t.Cleanup(func() { _ = srv.DeletePolicy(context.Background(), polMain.Id) })
+
+	// polOther: grpOther restricted to modelOther on the SAME provider. The
+	// client is not in grpOther, so modelOther must never be usable by it.
+	polOther, err := srv.CreatePolicy(ctx, api.AgentNetworkPolicyRequest{
+		Name:                   "e2e-guardrail-mp-other",
+		Enabled:                &enabled,
+		SourceGroups:           []string{grpOther.Id},
+		DestinationProviderIds: []string{pRestricted.Id},
+		GuardrailIds:           &[]string{gOther.Id},
+	})
+	require.NoError(t, err, "create other policy")
+	t.Cleanup(func() { _ = srv.DeletePolicy(context.Background(), polOther.Id) })
+
+	// polOpen: grpMain on pOpen with NO guardrail — unrestricted.
+	polOpen, err := srv.CreatePolicy(ctx, api.AgentNetworkPolicyRequest{
+		Name:                   "e2e-guardrail-mp-open",
+		Enabled:                &enabled,
+		SourceGroups:           []string{grpMain.Id},
+		DestinationProviderIds: []string{pOpen.Id},
+	})
+	require.NoError(t, err, "create open policy")
+	t.Cleanup(func() { _ = srv.DeletePolicy(context.Background(), polOpen.Id) })
+
+	settings, err := srv.GetSettings(ctx)
+	require.NoError(t, err, "read settings")
+	require.NotEmpty(t, settings.Endpoint, "endpoint must be assigned")
+
+	proxyToken, err := srv.CreateProxyTokenCLI(ctx, "e2e-guardrail-mp-proxy")
+	require.NoError(t, err, "mint proxy token")
+	px, err := harness.StartProxy(ctx, srv, proxyToken)
+	require.NoError(t, err, "start proxy")
+	t.Cleanup(func() { _ = px.Terminate(context.Background()) })
+
+	cl, err := harness.StartClient(ctx, srv, sk.Key)
+	require.NoError(t, err, "start client")
+	t.Cleanup(func() { _ = cl.Terminate(context.Background()) })
+
+	require.NoError(t, cl.WaitConnected(ctx, 90*time.Second), "client must connect to management")
+	proxyIP, err := cl.ResolveProxyIP(ctx, settings.Endpoint)
+	require.NoError(t, err, "resolve endpoint to proxy IP")
+	if err := cl.WaitProxyPeer(ctx, 180*time.Second); err != nil {
+		t.Fatalf("client did not see the proxy peer: %v\n=== proxy logs ===\n%s", err, px.Logs(context.Background()))
+	}
+
+	send := func(model string) (int, string) {
+		code, body, cerr := cl.Chat(ctx, settings.Endpoint, proxyIP, harness.WireChat, model, "Reply with exactly: pong", "")
+		require.NoError(t, cerr, "request must reach the proxy")
+		return code, body
+	}
+	// sendUntil200 absorbs first-call tunnel/DNS jitter on the freshly warmed tunnel.
+	sendUntil200 := func(model string) (int, string) {
+		var code int
+		var body string
+		deadline := time.Now().Add(90 * time.Second)
+		for time.Now().Before(deadline) {
+			code, body = send(model)
+			if code == 200 {
+				break
+			}
+			time.Sleep(5 * time.Second)
+		}
+		return code, body
+	}
+
+	t.Run("selected model allowed for its group", func(t *testing.T) {
+		code, body := sendUntil200(modelSelected)
+		assert.Equal(t, 200, code,
+			"grpMain's allowlisted model must be served; body: %s\n=== proxy logs ===\n%s", body, px.Logs(context.Background()))
+	})
+
+	t.Run("other group's model does not leak", func(t *testing.T) {
+		// modelOther is allowlisted only for grpOther. The grpMain client must be
+		// denied by management's per-policy/group check — not waved through by an
+		// account-wide union. This is the security-critical wrong-ALLOW guard.
+		code, body := send(modelOther)
+		assert.Equal(t, 403, code,
+			"another group's allowlisted model must be denied for this caller; body: %s\n=== proxy logs ===\n%s", body, px.Logs(context.Background()))
+		assert.Contains(t, body, "llm_policy.model_blocked",
+			"denial must be a model-allowlist decision; body: %s", body)
+	})
+
+	t.Run("unguarded policy leaves its provider unrestricted", func(t *testing.T) {
+		// polOpen carries no guardrail, so pOpen is unrestricted for grpMain. The
+		// old account-wide union would have blocked openModel (it is on no
+		// allowlist); it must now be served — the false-DENY guard.
+		code, body := sendUntil200(openModel)
+		assert.Equal(t, 200, code,
+			"an un-guardrailed policy's provider must not be blocked by another policy's allowlist; body: %s\n=== proxy logs ===\n%s", body, px.Logs(context.Background()))
+	})
+}

e2e/agentnetwork/guardrail_pergroup_providers_test.go

@@ -0,0 +1,422 @@
+//go:build e2e
+
+package agentnetwork
+
+import (
+	"context"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	"github.com/netbirdio/netbird/e2e/harness"
+	"github.com/netbirdio/netbird/shared/management/http/api"
+)
+
+// pergroupCase describes one provider surface for the per-group allowlist matrix.
+// selectedReq/otherReq are the model identifiers as they travel in the request
+// (URL path for Bedrock/Vertex, body "model" for chat/messages). selectedAllow/
+// otherAllow are the (normalized) forms the guardrail allowlist holds — for
+// Bedrock these differ from the request form so path normalization is exercised.
+type pergroupCase struct {
+	name          string
+	catalogID     string
+	wire          string // "chat", "messages", "vertex", "bedrock"
+	models        *[]api.AgentNetworkProviderModel
+	selectedReq   string
+	selectedAllow string
+	otherReq      string
+	otherAllow    string
+
+	providerID string // filled during setup
+}
+
+// TestGuardrailPerGroupAllowlist_AllProviders proves the per-policy/group model
+// allowlist end to end across every always-on provider surface, including the
+// path-routed ones (Vertex, Bedrock) where the model travels in the URL.
+//
+// For each provider two policies target it: grpMain (the client) is allowed only
+// selectedReq; grpOther (which the client is NOT in) is allowed only otherReq.
+// The client must get selectedReq served (200) and otherReq denied (403,
+// llm_policy.model_blocked) — the cross-group no-leak property. The deny is the
+// authoritative per-policy/group decision from management (the proxy per-provider
+// backstop carries the union of both models), so this also confirms management
+// receives the correct normalized model for path-routed providers.
+func TestGuardrailPerGroupAllowlist_AllProviders(t *testing.T) {
+	ctx, cancel := context.WithTimeout(context.Background(), 20*time.Minute)
+	defer cancel()
+
+	const (
+		vertexProject = "e2e-project"
+		vertexRegion  = "global"
+	)
+
+	priced := func(ids ...string) *[]api.AgentNetworkProviderModel {
+		out := make([]api.AgentNetworkProviderModel, 0, len(ids))
+		for _, id := range ids {
+			out = append(out, api.AgentNetworkProviderModel{Id: id, InputPer1k: 0.001, OutputPer1k: 0.001})
+		}
+		return &out
+	}
+
+	cases := []*pergroupCase{
+		{
+			name: "openai", catalogID: "openai_api", wire: harness.WireChat,
+			models:      priced("oai-model-a", "oai-model-b"),
+			selectedReq: "oai-model-a", selectedAllow: "oai-model-a",
+			otherReq: "oai-model-b", otherAllow: "oai-model-b",
+		},
+		{
+			name: "anthropic", catalogID: "anthropic_api", wire: harness.WireMessages,
+			models:      priced("ant-model-a", "ant-model-b"),
+			selectedReq: "ant-model-a", selectedAllow: "ant-model-a",
+			otherReq: "ant-model-b", otherAllow: "ant-model-b",
+		},
+		{
+			// Vertex catalog ids travel bare in the rawPredict path.
+			name: "vertex", catalogID: "vertex_ai_api", wire: "vertex",
+			selectedReq: "claude-sonnet-4-5", selectedAllow: "claude-sonnet-4-5",
+			otherReq: "claude-opus-4-6", otherAllow: "claude-opus-4-6",
+		},
+		{
+			// Bedrock request ids are region-prefixed/versioned; the parser
+			// normalizes them to the catalog key the allowlist holds.
+			name: "bedrock", catalogID: "bedrock_api", wire: "bedrock",
+			selectedReq: "us.anthropic.claude-sonnet-4-5-v1:0", selectedAllow: "anthropic.claude-sonnet-4-5",
+			otherReq: "us.anthropic.claude-opus-4-8-v1:0", otherAllow: "anthropic.claude-opus-4-8",
+		},
+	}
+
+	vllm, err := harness.StartVLLM(ctx, srv)
+	require.NoError(t, err, "start mock upstream")
+	t.Cleanup(func() { _ = vllm.Terminate(context.Background()) })
+
+	grpMain, err := srv.API().Groups.Create(ctx, api.PostApiGroupsJSONRequestBody{Name: "e2e-pergroup-main"})
+	require.NoError(t, err, "create main group")
+	t.Cleanup(func() { _ = srv.API().Groups.Delete(context.Background(), grpMain.Id) })
+
+	grpOther, err := srv.API().Groups.Create(ctx, api.PostApiGroupsJSONRequestBody{Name: "e2e-pergroup-other"})
+	require.NoError(t, err, "create other group")
+	t.Cleanup(func() { _ = srv.API().Groups.Delete(context.Background(), grpOther.Id) })
+
+	ephemeral := false
+	sk, err := srv.API().SetupKeys.Create(ctx, api.PostApiSetupKeysJSONRequestBody{
+		Name:       "e2e-pergroup-client",
+		Type:       "reusable",
+		ExpiresIn:  86400,
+		UsageLimit: 0,
+		AutoGroups: []string{grpMain.Id},
+		Ephemeral:  &ephemeral,
+	})
+	require.NoError(t, err, "mint setup key")
+	require.NotEmpty(t, sk.Key, "setup key plaintext")
+
+	staticKey := "static-e2e-token"
+	enabled := true
+
+	for i, c := range cases {
+		req := api.AgentNetworkProviderRequest{
+			Name:        "e2e-pergroup-" + c.name,
+			ProviderId:  c.catalogID,
+			UpstreamUrl: vllm.URL,
+			ApiKey:      &staticKey,
+			Enabled:     ptr(true),
+			Models:      c.models,
+		}
+		if i == 0 {
+			req.BootstrapCluster = ptr(harness.AgentNetworkCluster)
+		}
+		prov, perr := srv.CreateProvider(ctx, req)
+		require.NoError(t, perr, "create provider %s", c.name)
+		c.providerID = prov.Id
+		t.Cleanup(func() { _ = srv.DeleteProvider(context.Background(), prov.Id) })
+
+		gSel := mkAllowGuardrail(t, ctx, "e2e-pergroup-"+c.name+"-sel", c.selectedAllow)
+		gOth := mkAllowGuardrail(t, ctx, "e2e-pergroup-"+c.name+"-oth", c.otherAllow)
+
+		polMain, merr := srv.CreatePolicy(ctx, api.AgentNetworkPolicyRequest{
+			Name:                   "e2e-pergroup-" + c.name + "-main",
+			Enabled:                &enabled,
+			SourceGroups:           []string{grpMain.Id},
+			DestinationProviderIds: []string{prov.Id},
+			GuardrailIds:           &[]string{gSel.Id},
+		})
+		require.NoError(t, merr, "create main policy %s", c.name)
+		t.Cleanup(func() { _ = srv.DeletePolicy(context.Background(), polMain.Id) })
+
+		polOther, oerr := srv.CreatePolicy(ctx, api.AgentNetworkPolicyRequest{
+			Name:                   "e2e-pergroup-" + c.name + "-other",
+			Enabled:                &enabled,
+			SourceGroups:           []string{grpOther.Id},
+			DestinationProviderIds: []string{prov.Id},
+			GuardrailIds:           &[]string{gOth.Id},
+		})
+		require.NoError(t, oerr, "create other policy %s", c.name)
+		t.Cleanup(func() { _ = srv.DeletePolicy(context.Background(), polOther.Id) })
+	}
+
+	settings, err := srv.GetSettings(ctx)
+	require.NoError(t, err, "read settings")
+	require.NotEmpty(t, settings.Endpoint, "endpoint must be assigned")
+
+	proxyToken, err := srv.CreateProxyTokenCLI(ctx, "e2e-pergroup-proxy")
+	require.NoError(t, err, "mint proxy token")
+	px, err := harness.StartProxy(ctx, srv, proxyToken)
+	require.NoError(t, err, "start proxy")
+	t.Cleanup(func() { _ = px.Terminate(context.Background()) })
+
+	cl, err := harness.StartClient(ctx, srv, sk.Key)
+	require.NoError(t, err, "start client")
+	t.Cleanup(func() { _ = cl.Terminate(context.Background()) })
+
+	require.NoError(t, cl.WaitConnected(ctx, 90*time.Second), "client must connect to management")
+	proxyIP, err := cl.ResolveProxyIP(ctx, settings.Endpoint)
+	require.NoError(t, err, "resolve endpoint to proxy IP")
+	if err := cl.WaitProxyPeer(ctx, 180*time.Second); err != nil {
+		t.Fatalf("client did not see the proxy peer: %v\n=== proxy logs ===\n%s", err, px.Logs(context.Background()))
+	}
+
+	send := func(c *pergroupCase, model string) (int, string) {
+		var code int
+		var body string
+		var cerr error
+		switch c.wire {
+		case "vertex":
+			code, body, cerr = cl.Vertex(ctx, settings.Endpoint, proxyIP, vertexProject, vertexRegion, model, "Reply with exactly: pong", "")
+		case "bedrock":
+			code, body, cerr = cl.Bedrock(ctx, settings.Endpoint, proxyIP, model, "Reply with exactly: pong", "")
+		default:
+			code, body, cerr = cl.Chat(ctx, settings.Endpoint, proxyIP, c.wire, model, "Reply with exactly: pong", "")
+		}
+		require.NoError(t, cerr, "request must reach the proxy for %s", c.name)
+		return code, body
+	}
+
+	for _, c := range cases {
+		t.Run(c.name, func(t *testing.T) {
+			// grpMain's own model is served. Retry to absorb tunnel/DNS jitter on
+			// the first call over the freshly warmed tunnel.
+			var code int
+			var body string
+			deadline := time.Now().Add(90 * time.Second)
+			for time.Now().Before(deadline) {
+				code, body = send(c, c.selectedReq)
+				if code == 200 {
+					break
+				}
+				time.Sleep(5 * time.Second)
+			}
+			assert.Equal(t, 200, code,
+				"%s: grpMain's allowlisted model must be served; body: %s\n=== proxy logs ===\n%s", c.name, body, px.Logs(context.Background()))
+
+			// grpOther's model must NOT leak to the grpMain client.
+			code, body = send(c, c.otherReq)
+			assert.Equal(t, 403, code,
+				"%s: another group's allowlisted model must be denied for this caller; body: %s\n=== proxy logs ===\n%s", c.name, body, px.Logs(context.Background()))
+			assert.Contains(t, body, "llm_policy.model_blocked",
+				"%s: denial must be a model-allowlist decision, not routing; body: %s", c.name, body)
+		})
+	}
+}
+
+// TestGuardrailMultiGroupUser proves the per-policy/group decision for a caller
+// that belongs to MULTIPLE groups at once. Two scenarios, one shared stack:
+//
+//   - union across the user's groups: the client is in gUX and gUY, each with
+//     its own policy+guardrail on provider P1 (gUX->union-a, gUY->union-b). The
+//     client may use BOTH models (the union of its groups' allowlists) while a
+//     third, un-allowlisted model is denied.
+//   - an un-guardrailed group lifts the restriction: the client is in gMP and
+//     gMQ on provider P2, where gMP restricts to mix-a but gMQ's policy carries
+//     NO guardrail. Because one applicable policy is unrestricted, the client may
+//     use a model on no allowlist (mix-z) as well as mix-a.
+func TestGuardrailMultiGroupUser(t *testing.T) {
+	ctx, cancel := context.WithTimeout(context.Background(), 15*time.Minute)
+	defer cancel()
+
+	const (
+		unionA = "mg-union-a"
+		unionB = "mg-union-b"
+		unionC = "mg-union-c" // allowlisted by neither group
+		mixA   = "mg-mix-a"
+		mixZ   = "mg-mix-z" // on no allowlist; reachable only via the un-guardrailed policy
+	)
+
+	priced := func(ids ...string) *[]api.AgentNetworkProviderModel {
+		out := make([]api.AgentNetworkProviderModel, 0, len(ids))
+		for _, id := range ids {
+			out = append(out, api.AgentNetworkProviderModel{Id: id, InputPer1k: 0.001, OutputPer1k: 0.001})
+		}
+		return &out
+	}
+
+	vllm, err := harness.StartVLLM(ctx, srv)
+	require.NoError(t, err, "start mock upstream")
+	t.Cleanup(func() { _ = vllm.Terminate(context.Background()) })
+
+	mkGroup := func(name string) *api.Group {
+		g, gerr := srv.API().Groups.Create(ctx, api.PostApiGroupsJSONRequestBody{Name: name})
+		require.NoError(t, gerr, "create group %s", name)
+		t.Cleanup(func() { _ = srv.API().Groups.Delete(context.Background(), g.Id) })
+		return g
+	}
+	gUX := mkGroup("e2e-mg-union-x")
+	gUY := mkGroup("e2e-mg-union-y")
+	gMP := mkGroup("e2e-mg-mix-p")
+	gMQ := mkGroup("e2e-mg-mix-q")
+
+	ephemeral := false
+	sk, err := srv.API().SetupKeys.Create(ctx, api.PostApiSetupKeysJSONRequestBody{
+		Name:       "e2e-mg-client",
+		Type:       "reusable",
+		ExpiresIn:  86400,
+		UsageLimit: 0,
+		AutoGroups: []string{gUX.Id, gUY.Id, gMP.Id, gMQ.Id}, // client in all four groups
+		Ephemeral:  &ephemeral,
+	})
+	require.NoError(t, err, "mint setup key")
+	require.NotEmpty(t, sk.Key, "setup key plaintext")
+
+	staticKey := "static-e2e-token"
+	enabled := true
+
+	// P1 — union scenario: two restricting policies, one per group.
+	p1, err := srv.CreateProvider(ctx, api.AgentNetworkProviderRequest{
+		Name:             "e2e-mg-union",
+		ProviderId:       "openai_api",
+		UpstreamUrl:      vllm.URL,
+		ApiKey:           &staticKey,
+		Enabled:          ptr(true),
+		Models:           priced(unionA, unionB, unionC),
+		BootstrapCluster: ptr(harness.AgentNetworkCluster),
+	})
+	require.NoError(t, err, "create union provider")
+	t.Cleanup(func() { _ = srv.DeleteProvider(context.Background(), p1.Id) })
+
+	polUX, err := srv.CreatePolicy(ctx, api.AgentNetworkPolicyRequest{
+		Name:                   "e2e-mg-union-x",
+		Enabled:                &enabled,
+		SourceGroups:           []string{gUX.Id},
+		DestinationProviderIds: []string{p1.Id},
+		GuardrailIds:           &[]string{mkAllowGuardrail(t, ctx, "e2e-mg-union-x", unionA).Id},
+	})
+	require.NoError(t, err, "create union policy X")
+	t.Cleanup(func() { _ = srv.DeletePolicy(context.Background(), polUX.Id) })
+
+	polUY, err := srv.CreatePolicy(ctx, api.AgentNetworkPolicyRequest{
+		Name:                   "e2e-mg-union-y",
+		Enabled:                &enabled,
+		SourceGroups:           []string{gUY.Id},
+		DestinationProviderIds: []string{p1.Id},
+		GuardrailIds:           &[]string{mkAllowGuardrail(t, ctx, "e2e-mg-union-y", unionB).Id},
+	})
+	require.NoError(t, err, "create union policy Y")
+	t.Cleanup(func() { _ = srv.DeletePolicy(context.Background(), polUY.Id) })
+
+	// P2 — mixed scenario: one restricting policy + one un-guardrailed policy.
+	p2, err := srv.CreateProvider(ctx, api.AgentNetworkProviderRequest{
+		Name:        "e2e-mg-mix",
+		ProviderId:  "openai_api",
+		UpstreamUrl: vllm.URL,
+		ApiKey:      &staticKey,
+		Enabled:     ptr(true),
+		Models:      priced(mixA, mixZ),
+	})
+	require.NoError(t, err, "create mix provider")
+	t.Cleanup(func() { _ = srv.DeleteProvider(context.Background(), p2.Id) })
+
+	polMP, err := srv.CreatePolicy(ctx, api.AgentNetworkPolicyRequest{
+		Name:                   "e2e-mg-mix-p",
+		Enabled:                &enabled,
+		SourceGroups:           []string{gMP.Id},
+		DestinationProviderIds: []string{p2.Id},
+		GuardrailIds:           &[]string{mkAllowGuardrail(t, ctx, "e2e-mg-mix-p", mixA).Id},
+	})
+	require.NoError(t, err, "create mix policy P")
+	t.Cleanup(func() { _ = srv.DeletePolicy(context.Background(), polMP.Id) })
+
+	polMQ, err := srv.CreatePolicy(ctx, api.AgentNetworkPolicyRequest{
+		Name:                   "e2e-mg-mix-q",
+		Enabled:                &enabled,
+		SourceGroups:           []string{gMQ.Id},
+		DestinationProviderIds: []string{p2.Id}, // NO guardrail -> unrestricted
+	})
+	require.NoError(t, err, "create mix policy Q")
+	t.Cleanup(func() { _ = srv.DeletePolicy(context.Background(), polMQ.Id) })
+
+	settings, err := srv.GetSettings(ctx)
+	require.NoError(t, err, "read settings")
+	require.NotEmpty(t, settings.Endpoint, "endpoint must be assigned")
+
+	proxyToken, err := srv.CreateProxyTokenCLI(ctx, "e2e-mg-proxy")
+	require.NoError(t, err, "mint proxy token")
+	px, err := harness.StartProxy(ctx, srv, proxyToken)
+	require.NoError(t, err, "start proxy")
+	t.Cleanup(func() { _ = px.Terminate(context.Background()) })
+
+	cl, err := harness.StartClient(ctx, srv, sk.Key)
+	require.NoError(t, err, "start client")
+	t.Cleanup(func() { _ = cl.Terminate(context.Background()) })
+
+	require.NoError(t, cl.WaitConnected(ctx, 90*time.Second), "client must connect to management")
+	proxyIP, err := cl.ResolveProxyIP(ctx, settings.Endpoint)
+	require.NoError(t, err, "resolve endpoint to proxy IP")
+	if err := cl.WaitProxyPeer(ctx, 180*time.Second); err != nil {
+		t.Fatalf("client did not see the proxy peer: %v\n=== proxy logs ===\n%s", err, px.Logs(context.Background()))
+	}
+
+	send := func(model string) (int, string) {
+		code, body, cerr := cl.Chat(ctx, settings.Endpoint, proxyIP, harness.WireChat, model, "Reply with exactly: pong", "")
+		require.NoError(t, cerr, "request must reach the proxy")
+		return code, body
+	}
+	sendUntil200 := func(model string) (int, string) {
+		var code int
+		var body string
+		deadline := time.Now().Add(90 * time.Second)
+		for time.Now().Before(deadline) {
+			code, body = send(model)
+			if code == 200 {
+				break
+			}
+			time.Sleep(5 * time.Second)
+		}
+		return code, body
+	}
+
+	t.Run("union across the user's groups", func(t *testing.T) {
+		code, body := sendUntil200(unionA)
+		assert.Equal(t, 200, code, "model allowed by group X must be served; body: %s\n=== proxy logs ===\n%s", body, px.Logs(context.Background()))
+
+		code, body = sendUntil200(unionB)
+		assert.Equal(t, 200, code, "model allowed by group Y must also be served (union across the user's groups); body: %s", body)
+
+		code, body = send(unionC)
+		assert.Equal(t, 403, code, "a model on neither group's allowlist must be denied; body: %s", body)
+		assert.Contains(t, body, "llm_policy.model_blocked")
+	})
+
+	t.Run("an un-guardrailed group lifts the restriction", func(t *testing.T) {
+		code, body := sendUntil200(mixA)
+		assert.Equal(t, 200, code, "the restricted group's model must be served; body: %s\n=== proxy logs ===\n%s", body, px.Logs(context.Background()))
+
+		code, body = sendUntil200(mixZ)
+		assert.Equal(t, 200, code,
+			"a non-allowlisted model must be served because the user is also in a group whose policy has no guardrail; body: %s\n=== proxy logs ===\n%s", body, px.Logs(context.Background()))
+	})
+}
+
+// mkAllowGuardrail creates a guardrail whose model allowlist is enabled and holds
+// exactly the given model, registering cleanup.
+func mkAllowGuardrail(t *testing.T, ctx context.Context, name, model string) api.AgentNetworkGuardrail {
+	t.Helper()
+	var gr api.AgentNetworkGuardrailRequest
+	gr.Name = name
+	gr.Checks.ModelAllowlist.Enabled = true
+	gr.Checks.ModelAllowlist.Models = []string{model}
+	g, err := srv.CreateGuardrail(ctx, gr)
+	require.NoError(t, err, "create guardrail %s", name)
+	t.Cleanup(func() { _ = srv.DeleteGuardrail(context.Background(), g.Id) })
+	return g
+}

e2e/harness/proxy.go

@@ -38,7 +38,11 @@ type Proxy struct {
 // network, registered via the given account proxy token and serving the
 // AgentNetworkCluster over a self-signed wildcard cert. It does not wait for
 // peer connectivity — callers poll management for the proxy peer.
-func StartProxy(ctx context.Context, c *Combined, proxyToken string) (*Proxy, error) {
+// StartProxy launches the reverse-proxy container. Optional envOverrides are
+// merged into the container environment after the defaults, so callers can set
+// or override any NB_PROXY_* var (e.g. NB_PROXY_TUNNEL_CACHE_TTL for tests that
+// need a short authorization-cache window).
+func StartProxy(ctx context.Context, c *Combined, proxyToken string, envOverrides ...map[string]string) (*Proxy, error) {
 	root, err := repoRoot()
 	if err != nil {
 		return nil, err
@@ -93,6 +97,12 @@ func StartProxy(ctx context.Context, c *Combined, proxyToken string) (*Proxy, er
 		WaitingFor: wait.ForLog("Initial mapping sync complete").WithStartupTimeout(90 * time.Second),
 	}
 
+	for _, ov := range envOverrides {
+		for k, v := range ov {
+			req.Env[k] = v
+		}
+	}
+
 	ctr, err := testcontainers.GenericContainer(ctx, testcontainers.GenericContainerRequest{
 		ContainerRequest: req,
 		Started:          true,

infrastructure_files/getting-started.sh

@@ -228,8 +228,6 @@ read_enable_crowdsec() {
   echo "CrowdSec checks client IPs against a community threat intelligence database" > /dev/stderr
   echo "and blocks known malicious sources before they reach your services." > /dev/stderr
   echo "A local CrowdSec LAPI container will be added to your deployment." > /dev/stderr
-  echo "It also enables the AppSec (WAF) endpoint, so services can inspect HTTP" > /dev/stderr
-  echo "requests for exploits. Both stay off per service until you enable them." > /dev/stderr
   echo -n "Enable CrowdSec? [y/N]: " > /dev/stderr
   read -r CHOICE < /dev/tty
 
@@ -499,8 +497,7 @@ generate_configuration_files() {
         # TCP ServersTransport for PROXY protocol v2 to the proxy backend
         render_traefik_dynamic > traefik-dynamic.yaml
         if [[ "$ENABLE_CROWDSEC" == "true" ]]; then
-          mkdir -p crowdsec/acquis.d
-          render_crowdsec_appsec_acquis > crowdsec/acquis.d/appsec.yaml
+          mkdir -p crowdsec
         fi
       fi
       ;;
@@ -534,23 +531,6 @@ generate_configuration_files() {
   return 0
 }
 
-# The AppSec (WAF) listener only exists if an appsec acquisition datasource is
-# configured. One datasource is one listener carrying one merged rule set: the
-# protocol has no rule-set selector, so per-service rule variation would need
-# either a second datasource on another port or pre_eval hooks filtering on
-# req.Host.
-render_crowdsec_appsec_acquis() {
-  cat <<EOF
-source: appsec
-listen_addr: 0.0.0.0:7422
-appsec_configs:
-  - crowdsecurity/appsec-default
-labels:
-  type: appsec
-EOF
-  return 0
-}
-
 start_services_and_show_instructions() {
   # For built-in Traefik, start containers immediately
   # For NPM, start containers first (NPM needs services running to create proxy)
@@ -762,11 +742,7 @@ render_docker_compose_traefik_builtin() {
     restart: unless-stopped
     networks: [netbird]
     environment:
-      # appsec-generic-rules is required alongside appsec-virtual-patching:
-      # the appsec-default config references crowdsecurity/generic-* and
-      # crowdsecurity/experimental-*, which only that collection provides, and
-      # the engine exits at startup if they are missing.
-      COLLECTIONS: crowdsecurity/linux crowdsecurity/appsec-virtual-patching crowdsecurity/appsec-generic-rules
+      COLLECTIONS: crowdsecurity/linux
     volumes:
       - ./crowdsec:/etc/crowdsec
       - crowdsec_db:/var/lib/crowdsec/data
@@ -1031,11 +1007,6 @@ EOF
     cat <<EOF
 NB_PROXY_CROWDSEC_API_URL=http://crowdsec:8080
 NB_PROXY_CROWDSEC_API_KEY=$CROWDSEC_BOUNCER_KEY
-# AppSec (WAF) request inspection. Separate endpoint from the LAPI above and
-# validated with the same bouncer key. Setting it makes the proxy advertise the
-# AppSec capability, which is what lets a service select appsec_mode; nothing is
-# inspected until a service opts in.
-NB_PROXY_CROWDSEC_APPSEC_URL=http://crowdsec:7422/
 EOF
   fi
 

management/internals/modules/agentnetwork/manager.go

@@ -86,12 +86,17 @@ type Manager interface {
 
 // PolicySelectionInput is the per-request selection envelope. The
 // proxy populates it from CapturedData (account, user, groups) plus
-// the provider llm_router resolved.
+// the provider llm_router resolved and the model it extracted.
 type PolicySelectionInput struct {
 	AccountID  string
 	UserID     string
 	GroupIDs   []string
 	ProviderID string
+	// Model is the already-normalised upstream model id the proxy extracted
+	// (parser strips Bedrock region/version, Vertex @version), so a
+	// case-insensitive compare suffices. Empty = undetermined → not permitted
+	// (fail closed).
+	Model string
 }
 
 // PolicySelectionResult names the policy that "pays" for this request

management/internals/modules/agentnetwork/policyselect.go

@@ -5,6 +5,7 @@ import (
 	"fmt"
 	"math"
 	"sort"
+	"strings"
 	"time"
 
 	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork/types"
@@ -35,6 +36,10 @@ const (
 	denyCodeAccountTokenCapExceeded = "llm_account.token_cap_exceeded"
 	//nolint:gosec // account deny code label, not a credential
 	denyCodeAccountBudgetCapExceeded = "llm_account.budget_cap_exceeded"
+	// denyCodeModelBlocked is returned when policies govern the request's
+	// (provider, caller-groups) but none permits the model. Matches the proxy
+	// guardrail's code so both layers surface the same label.
+	denyCodeModelBlocked = "llm_policy.model_blocked"
 )
 
 // consumptionCache holds the consumption counters prefetched for one
@@ -159,6 +164,25 @@ func (m *managerImpl) SelectPolicyForRequest(ctx context.Context, in PolicySelec
 	}
 	candidates := filterApplicablePolicies(policies, in)
 
+	// Model-allowlist gate scoped to the matched policies: keep candidates whose
+	// guardrails permit the model (none enabled = unrestricted), deny when
+	// policies apply but none permits it. Skip the load when none has a guardrail.
+	if len(candidates) > 0 && anyPolicyHasGuardrails(candidates) {
+		guardrailsByID, gErr := m.loadGuardrailsByID(ctx, in.AccountID)
+		if gErr != nil {
+			return nil, gErr
+		}
+		permitted := filterModelPermittedPolicies(candidates, guardrailsByID, in.Model)
+		if len(permitted) == 0 {
+			return &PolicySelectionResult{
+				Allow:      false,
+				DenyCode:   denyCodeModelBlocked,
+				DenyReason: modelBlockedReason(in.Model),
+			}, nil
+		}
+		candidates = permitted
+	}
+
 	// Prefetch every consumption counter the ceiling + candidate policies will
 	// read, in a single store round-trip, then score against the cache.
 	cache, err := m.prefetchConsumption(ctx, in, rules, candidates, now)
@@ -250,6 +274,90 @@ func filterApplicablePolicies(policies []*types.Policy, in PolicySelectionInput)
 	return out
 }
 
+// anyPolicyHasGuardrails reports whether any policy references at least one
+// guardrail, so the selector can skip loading guardrails when none do.
+func anyPolicyHasGuardrails(policies []*types.Policy) bool {
+	for _, p := range policies {
+		if p != nil && len(p.GuardrailIDs) > 0 {
+			return true
+		}
+	}
+	return false
+}
+
+// loadGuardrailsByID loads the account's guardrails indexed by ID. Used by the
+// model-allowlist gate to resolve each candidate policy's attached guardrails.
+func (m *managerImpl) loadGuardrailsByID(ctx context.Context, accountID string) (map[string]*types.Guardrail, error) {
+	guardrails, err := m.store.GetAccountAgentNetworkGuardrails(ctx, store.LockingStrengthNone, accountID)
+	if err != nil {
+		return nil, fmt.Errorf("list account guardrails: %w", err)
+	}
+	byID := make(map[string]*types.Guardrail, len(guardrails))
+	for _, g := range guardrails {
+		if g != nil {
+			byID[g.ID] = g
+		}
+	}
+	return byID, nil
+}
+
+// filterModelPermittedPolicies returns the subset of policies whose guardrails
+// permit the model. Order is preserved so downstream scoring is unaffected.
+func filterModelPermittedPolicies(policies []*types.Policy, byID map[string]*types.Guardrail, model string) []*types.Policy {
+	out := make([]*types.Policy, 0, len(policies))
+	for _, p := range policies {
+		if policyPermitsModel(p, byID, model) {
+			out = append(out, p)
+		}
+	}
+	return out
+}
+
+// policyPermitsModel reports whether a policy permits the model. No
+// allowlist-enabled guardrail = unrestricted (permits any, incl. empty);
+// otherwise the model must be in the union of its allowlists, so an
+// empty/undetermined model fails closed.
+func policyPermitsModel(p *types.Policy, byID map[string]*types.Guardrail, model string) bool {
+	if p == nil {
+		return false
+	}
+	wanted := normaliseModelID(model)
+	restricted := false
+	for _, gID := range p.GuardrailIDs {
+		g, ok := byID[gID]
+		if !ok || g == nil || !g.Checks.ModelAllowlist.Enabled {
+			continue
+		}
+		restricted = true
+		if wanted == "" {
+			continue
+		}
+		for _, allowed := range g.Checks.ModelAllowlist.Models {
+			if normaliseModelID(allowed) == wanted {
+				return true
+			}
+		}
+	}
+	return !restricted
+}
+
+// normaliseModelID lowercases and trims a model identifier so the allowlist
+// compare is case-insensitive and trim-tolerant. Mirrors the proxy guardrail's
+// normaliseModel so both layers agree on what "same model" means.
+func normaliseModelID(model string) string {
+	return strings.ToLower(strings.TrimSpace(model))
+}
+
+// modelBlockedReason builds the human-readable deny reason for a model-allowlist
+// rejection. The model is quoted when known; an undetermined model is reported
+// as such so the access log distinguishes "wrong model" from "no model".
+func modelBlockedReason(model string) string {
+	if normaliseModelID(model) == "" {
+		return "request model could not be determined for the policy allowlist"
+	}
+	return fmt.Sprintf("model %q is not permitted by any applicable policy allowlist", model)
+}
+
 // candidate is the per-policy intermediate the selector ranks. A
 // policy that's been exhausted on any enabled cap never makes it
 // into this slice; the selector's deny envelope carries the latest

management/internals/modules/agentnetwork/policyselect_model_test.go

@@ -0,0 +1,329 @@
+package agentnetwork
+
+import (
+	"context"
+	"errors"
+	"testing"
+	"time"
+
+	"github.com/golang/mock/gomock"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork/types"
+	"github.com/netbirdio/netbird/management/server/store"
+)
+
+// guardedPolicy builds an enabled, uncapped policy that authorises sourceGroups
+// to reach providerID under the given guardrails. Uncapped keeps the selector's
+// headroom scoring trivial so these tests isolate the model-allowlist gate.
+func guardedPolicy(id, account string, sourceGroups []string, providerID string, guardrailIDs ...string) *types.Policy {
+	return &types.Policy{
+		ID:                     id,
+		AccountID:              account,
+		Enabled:                true,
+		SourceGroups:           sourceGroups,
+		DestinationProviderIDs: []string{providerID},
+		GuardrailIDs:           guardrailIDs,
+		CreatedAt:              time.Now().UTC(),
+	}
+}
+
+// allowlistGuardrail builds a guardrail whose model allowlist is enabled and
+// carries the given models.
+func allowlistGuardrail(id, account string, models ...string) *types.Guardrail {
+	return &types.Guardrail{
+		ID:        id,
+		AccountID: account,
+		Checks: types.GuardrailChecks{
+			ModelAllowlist: types.GuardrailModelAllowlist{Enabled: true, Models: models},
+		},
+	}
+}
+
+func expectPolicies(mockStore *store.MockStore, account string, policies ...*types.Policy) {
+	mockStore.EXPECT().
+		GetAccountAgentNetworkPolicies(gomock.Any(), gomock.Any(), account).
+		Return(policies, nil)
+}
+
+func expectGuardrails(mockStore *store.MockStore, account string, guardrails ...*types.Guardrail) {
+	mockStore.EXPECT().
+		GetAccountAgentNetworkGuardrails(gomock.Any(), gomock.Any(), account).
+		Return(guardrails, nil)
+}
+
+// TestSelectPolicy_ModelBlockedByAllowlist proves the authoritative allowlist
+// decision: a policy authorises the (provider, group) but restricts the model,
+// and the requested model isn't on the list, so the request is denied.
+func TestSelectPolicy_ModelBlockedByAllowlist(t *testing.T) {
+	ctrl := gomock.NewController(t)
+	mgr, mockStore := newSelectorMgr(t, ctrl)
+
+	policy := guardedPolicy("pol-A", "acc-1", []string{"grp-eng"}, "prov-1", "g-1")
+	expectPolicies(mockStore, "acc-1", policy)
+	expectGuardrails(mockStore, "acc-1", allowlistGuardrail("g-1", "acc-1", "gpt-4o"))
+
+	res, err := mgr.SelectPolicyForRequest(context.Background(), PolicySelectionInput{
+		AccountID:  "acc-1",
+		UserID:     "user-1",
+		GroupIDs:   []string{"grp-eng"},
+		ProviderID: "prov-1",
+		Model:      "claude-opus-4",
+	})
+	require.NoError(t, err)
+	assert.False(t, res.Allow, "a model outside the only applicable policy's allowlist must be denied")
+	assert.Equal(t, denyCodeModelBlocked, res.DenyCode, "deny code must be model_blocked")
+	assert.NotEmpty(t, res.DenyReason, "deny reason must be populated")
+}
+
+// TestSelectPolicy_ModelAllowedByAllowlist is the allow counterpart: the model
+// is on the applicable policy's allowlist, so selection proceeds normally.
+func TestSelectPolicy_ModelAllowedByAllowlist(t *testing.T) {
+	ctrl := gomock.NewController(t)
+	mgr, mockStore := newSelectorMgr(t, ctrl)
+
+	policy := guardedPolicy("pol-A", "acc-1", []string{"grp-eng"}, "prov-1", "g-1")
+	expectPolicies(mockStore, "acc-1", policy)
+	expectGuardrails(mockStore, "acc-1", allowlistGuardrail("g-1", "acc-1", "gpt-4o", "claude-opus-4"))
+	expectConsumptionBatch(mockStore, nil)
+
+	res, err := mgr.SelectPolicyForRequest(context.Background(), PolicySelectionInput{
+		AccountID:  "acc-1",
+		UserID:     "user-1",
+		GroupIDs:   []string{"grp-eng"},
+		ProviderID: "prov-1",
+		Model:      "claude-opus-4",
+	})
+	require.NoError(t, err)
+	assert.True(t, res.Allow, "a model on the applicable policy's allowlist must be allowed")
+	assert.Equal(t, "pol-A", res.SelectedPolicyID)
+}
+
+// TestSelectPolicy_CaseInsensitiveModelMatch proves the compare tolerates case
+// and surrounding whitespace, matching the proxy guardrail's normalisation.
+func TestSelectPolicy_CaseInsensitiveModelMatch(t *testing.T) {
+	ctrl := gomock.NewController(t)
+	mgr, mockStore := newSelectorMgr(t, ctrl)
+
+	policy := guardedPolicy("pol-A", "acc-1", []string{"grp-eng"}, "prov-1", "g-1")
+	expectPolicies(mockStore, "acc-1", policy)
+	expectGuardrails(mockStore, "acc-1", allowlistGuardrail("g-1", "acc-1", "  GPT-4o  "))
+	expectConsumptionBatch(mockStore, nil)
+
+	res, err := mgr.SelectPolicyForRequest(context.Background(), PolicySelectionInput{
+		AccountID:  "acc-1",
+		GroupIDs:   []string{"grp-eng"},
+		ProviderID: "prov-1",
+		Model:      "gpt-4o",
+	})
+	require.NoError(t, err)
+	assert.True(t, res.Allow, "case/whitespace variants must match the allowlist entry")
+}
+
+// TestSelectPolicy_UnguardedPolicyIsUnrestricted is the false-deny fix: when two
+// policies authorise the same (provider, group) and one has no guardrail, that
+// policy makes the request unrestricted — not caught by the other's allowlist.
+func TestSelectPolicy_UnguardedPolicyIsUnrestricted(t *testing.T) {
+	ctrl := gomock.NewController(t)
+	mgr, mockStore := newSelectorMgr(t, ctrl)
+
+	restricted := guardedPolicy("pol-restricted", "acc-1", []string{"grp-eng"}, "prov-1", "g-1")
+	open := guardedPolicy("pol-open", "acc-1", []string{"grp-eng"}, "prov-1") // no guardrail
+	expectPolicies(mockStore, "acc-1", restricted, open)
+	expectGuardrails(mockStore, "acc-1", allowlistGuardrail("g-1", "acc-1", "gpt-4o"))
+	expectConsumptionBatch(mockStore, nil)
+
+	res, err := mgr.SelectPolicyForRequest(context.Background(), PolicySelectionInput{
+		AccountID:  "acc-1",
+		GroupIDs:   []string{"grp-eng"},
+		ProviderID: "prov-1",
+		Model:      "claude-opus-4",
+	})
+	require.NoError(t, err)
+	assert.True(t, res.Allow, "an un-guardrailed policy for the same (provider, group) must leave the request unrestricted")
+	assert.Equal(t, "pol-open", res.SelectedPolicyID, "the unrestricted policy must be the one that pays")
+}
+
+// TestSelectPolicy_AllowlistDoesNotLeakAcrossGroups is the false-allow fix: a
+// model allowlisted only for grp-b must not be usable by a grp-a caller. The
+// selector considers only policies applicable to the caller's groups.
+func TestSelectPolicy_AllowlistDoesNotLeakAcrossGroups(t *testing.T) {
+	ctrl := gomock.NewController(t)
+	mgr, mockStore := newSelectorMgr(t, ctrl)
+
+	polA := guardedPolicy("pol-a", "acc-1", []string{"grp-a"}, "prov-1", "g-a")
+	polB := guardedPolicy("pol-b", "acc-1", []string{"grp-b"}, "prov-1", "g-b")
+	expectPolicies(mockStore, "acc-1", polA, polB)
+	expectGuardrails(mockStore, "acc-1",
+		allowlistGuardrail("g-a", "acc-1", "gpt-4o"),
+		allowlistGuardrail("g-b", "acc-1", "claude-opus-4"),
+	)
+
+	res, err := mgr.SelectPolicyForRequest(context.Background(), PolicySelectionInput{
+		AccountID:  "acc-1",
+		GroupIDs:   []string{"grp-a"},
+		ProviderID: "prov-1",
+		Model:      "claude-opus-4", // only allowed for grp-b
+	})
+	require.NoError(t, err)
+	assert.False(t, res.Allow, "grp-b's allowlisted model must not leak to a grp-a caller")
+	assert.Equal(t, denyCodeModelBlocked, res.DenyCode)
+}
+
+// TestSelectPolicy_UndeterminedModelFailsClosed proves the fail-closed contract
+// mirrors the proxy: with a restricted applicable policy and an empty model
+// (e.g. a path-routed shape the parser couldn't map), the request is denied.
+func TestSelectPolicy_UndeterminedModelFailsClosed(t *testing.T) {
+	ctrl := gomock.NewController(t)
+	mgr, mockStore := newSelectorMgr(t, ctrl)
+
+	policy := guardedPolicy("pol-A", "acc-1", []string{"grp-eng"}, "prov-1", "g-1")
+	expectPolicies(mockStore, "acc-1", policy)
+	expectGuardrails(mockStore, "acc-1", allowlistGuardrail("g-1", "acc-1", "gpt-4o"))
+
+	res, err := mgr.SelectPolicyForRequest(context.Background(), PolicySelectionInput{
+		AccountID:  "acc-1",
+		GroupIDs:   []string{"grp-eng"},
+		ProviderID: "prov-1",
+		Model:      "", // undetermined
+	})
+	require.NoError(t, err)
+	assert.False(t, res.Allow, "an undetermined model must fail closed against a restricted policy")
+	assert.Equal(t, denyCodeModelBlocked, res.DenyCode)
+}
+
+// TestSelectPolicy_DisabledAllowlistDoesNotRestrict proves a guardrail whose
+// model allowlist is disabled imposes no model restriction, even though the
+// policy references it.
+func TestSelectPolicy_DisabledAllowlistDoesNotRestrict(t *testing.T) {
+	ctrl := gomock.NewController(t)
+	mgr, mockStore := newSelectorMgr(t, ctrl)
+
+	policy := guardedPolicy("pol-A", "acc-1", []string{"grp-eng"}, "prov-1", "g-1")
+	disabled := &types.Guardrail{
+		ID:        "g-1",
+		AccountID: "acc-1",
+		Checks: types.GuardrailChecks{
+			ModelAllowlist: types.GuardrailModelAllowlist{Enabled: false, Models: []string{"gpt-4o"}},
+		},
+	}
+	expectPolicies(mockStore, "acc-1", policy)
+	expectGuardrails(mockStore, "acc-1", disabled)
+	expectConsumptionBatch(mockStore, nil)
+
+	res, err := mgr.SelectPolicyForRequest(context.Background(), PolicySelectionInput{
+		AccountID:  "acc-1",
+		GroupIDs:   []string{"grp-eng"},
+		ProviderID: "prov-1",
+		Model:      "anything-goes",
+	})
+	require.NoError(t, err)
+	assert.True(t, res.Allow, "a disabled allowlist must not restrict the model")
+	assert.Equal(t, "pol-A", res.SelectedPolicyID)
+}
+
+// TestSelectPolicy_UnionAcrossPolicyGuardrails proves a policy with multiple
+// allowlist guardrails permits the union of their models (not just the first).
+func TestSelectPolicy_UnionAcrossPolicyGuardrails(t *testing.T) {
+	ctrl := gomock.NewController(t)
+	mgr, mockStore := newSelectorMgr(t, ctrl)
+
+	policy := guardedPolicy("pol-A", "acc-1", []string{"grp-eng"}, "prov-1", "g-1", "g-2")
+	expectPolicies(mockStore, "acc-1", policy)
+	expectGuardrails(mockStore, "acc-1",
+		allowlistGuardrail("g-1", "acc-1", "gpt-4o"),
+		allowlistGuardrail("g-2", "acc-1", "claude-opus-4"),
+	)
+	expectConsumptionBatch(mockStore, nil)
+
+	res, err := mgr.SelectPolicyForRequest(context.Background(), PolicySelectionInput{
+		AccountID:  "acc-1",
+		GroupIDs:   []string{"grp-eng"},
+		ProviderID: "prov-1",
+		Model:      "claude-opus-4", // only in the second guardrail's list
+	})
+	require.NoError(t, err)
+	assert.True(t, res.Allow, "a model in any of the policy's allowlist guardrails must be permitted")
+	assert.Equal(t, "pol-A", res.SelectedPolicyID)
+}
+
+// TestSelectPolicy_GuardrailLookupErrorPropagates proves a store failure while
+// resolving the candidate policies' guardrails surfaces as an error, not a
+// silent allow/deny.
+func TestSelectPolicy_GuardrailLookupErrorPropagates(t *testing.T) {
+	ctrl := gomock.NewController(t)
+	mgr, mockStore := newSelectorMgr(t, ctrl)
+
+	policy := guardedPolicy("pol-A", "acc-1", []string{"grp-eng"}, "prov-1", "g-1")
+	expectPolicies(mockStore, "acc-1", policy)
+	mockStore.EXPECT().
+		GetAccountAgentNetworkGuardrails(gomock.Any(), gomock.Any(), "acc-1").
+		Return(nil, errors.New("store unavailable"))
+
+	res, err := mgr.SelectPolicyForRequest(context.Background(), PolicySelectionInput{
+		AccountID:  "acc-1",
+		GroupIDs:   []string{"grp-eng"},
+		ProviderID: "prov-1",
+		Model:      "gpt-4o",
+	})
+	require.Error(t, err, "a guardrail-lookup failure must surface as an error")
+	assert.Nil(t, res)
+}
+
+// TestSelectPolicy_MissingGuardrailReferenceTreatedAsUnrestricted proves a
+// policy referencing a guardrail ID absent from the account's set (a stale
+// reference) imposes no model restriction — same as no guardrail.
+func TestSelectPolicy_MissingGuardrailReferenceTreatedAsUnrestricted(t *testing.T) {
+	ctrl := gomock.NewController(t)
+	mgr, mockStore := newSelectorMgr(t, ctrl)
+
+	policy := guardedPolicy("pol-A", "acc-1", []string{"grp-eng"}, "prov-1", "g-missing")
+	expectPolicies(mockStore, "acc-1", policy)
+	expectGuardrails(mockStore, "acc-1")
+	expectConsumptionBatch(mockStore, nil)
+
+	res, err := mgr.SelectPolicyForRequest(context.Background(), PolicySelectionInput{
+		AccountID:  "acc-1",
+		GroupIDs:   []string{"grp-eng"},
+		ProviderID: "prov-1",
+		Model:      "anything-goes",
+	})
+	require.NoError(t, err)
+	assert.True(t, res.Allow, "an orphaned guardrail reference must not restrict the model")
+	assert.Equal(t, "pol-A", res.SelectedPolicyID)
+}
+
+// TestSelectPolicy_PartialCandidatesPermittedAfterModelFilter proves the model
+// gate narrows candidates before cap scoring: the permitting policy is selected
+// even though the blocked one has a larger, more attractive cap.
+func TestSelectPolicy_PartialCandidatesPermittedAfterModelFilter(t *testing.T) {
+	ctrl := gomock.NewController(t)
+	mgr, mockStore := newSelectorMgr(t, ctrl)
+
+	polBig := guardedPolicy("pol-big", "acc-1", []string{"grp-eng"}, "prov-1", "g-restrict")
+	polBig.Limits = types.PolicyLimits{
+		TokenLimit: types.PolicyTokenLimit{Enabled: true, GroupCap: 1_000_000, WindowSeconds: 3600},
+	}
+	polSmall := guardedPolicy("pol-small", "acc-1", []string{"grp-eng"}, "prov-1", "g-permit")
+	polSmall.Limits = types.PolicyLimits{
+		TokenLimit: types.PolicyTokenLimit{Enabled: true, GroupCap: 100, WindowSeconds: 3600},
+	}
+	expectPolicies(mockStore, "acc-1", polBig, polSmall)
+	expectGuardrails(mockStore, "acc-1",
+		allowlistGuardrail("g-restrict", "acc-1", "gpt-4o"),
+		allowlistGuardrail("g-permit", "acc-1", "claude-opus-4"),
+	)
+	expectConsumptionBatch(mockStore, nil)
+
+	res, err := mgr.SelectPolicyForRequest(context.Background(), PolicySelectionInput{
+		AccountID:  "acc-1",
+		GroupIDs:   []string{"grp-eng"},
+		ProviderID: "prov-1",
+		Model:      "claude-opus-4", // only pol-small's guardrail permits this
+	})
+	require.NoError(t, err)
+	assert.True(t, res.Allow)
+	assert.Equal(t, "pol-small", res.SelectedPolicyID,
+		"the model filter must exclude pol-big before cap scoring")
+}

management/internals/modules/agentnetwork/synthesizer.go

@@ -235,7 +235,12 @@ func SynthesizeServices(ctx context.Context, s store.Store, accountID string) ([
 
 	mergedGuardrails := mergeGuardrails(enabledPolicies, guardrailsByID)
 	applyAccountCollectionControls(&mergedGuardrails, settings)
-	guardrailJSON, err := marshalGuardrailConfig(mergedGuardrails)
+	// The proxy guardrail is a per-provider fail-closed backstop; the
+	// authoritative per-policy/group decision is management's
+	// SelectPolicyForRequest. A provider lands in this map only when every
+	// authorising policy restricts models.
+	providerAllowlists := buildProviderAllowlists(enabledPolicies, guardrailsByID)
+	guardrailJSON, err := marshalGuardrailConfig(providerAllowlists, mergedGuardrails.PromptCapture)
 	if err != nil {
 		return nil, err
 	}
@@ -780,10 +785,12 @@ func buildMiddlewareChain(routerCfgJSON, identityInjectJSON, guardrailJSON []byt
 
 // guardrailConfig is the JSON shape the proxy-side llm_guardrail
 // middleware expects. Mirrors the proxy registration documented in
-// the management→proxy contract.
+// the management→proxy contract. provider_allowlists is keyed by the
+// resolved provider id llm_router stamps; a provider absent from the map is
+// unrestricted at the proxy layer.
 type guardrailConfig struct {
-	ModelAllowlist []string               `json:"model_allowlist,omitempty"`
-	PromptCapture  guardrailPromptCapture `json:"prompt_capture"`
+	ProviderAllowlists map[string][]string    `json:"provider_allowlists,omitempty"`
+	PromptCapture      guardrailPromptCapture `json:"prompt_capture"`
 }
 
 type guardrailPromptCapture struct {
@@ -828,13 +835,10 @@ func applyAccountCollectionControls(merged *MergedGuardrails, settings *types.Se
 	merged.PromptCapture.RedactPii = settings.RedactPii || merged.PromptCapture.RedactPii
 }
 
-func marshalGuardrailConfig(merged MergedGuardrails) ([]byte, error) {
+func marshalGuardrailConfig(providerAllowlists map[string][]string, capture MergedPromptCapture) ([]byte, error) {
 	cfg := guardrailConfig{
-		ModelAllowlist: merged.ModelAllowlist,
-		PromptCapture: guardrailPromptCapture{
-			Enabled:   merged.PromptCapture.Enabled,
-			RedactPii: merged.PromptCapture.RedactPii,
-		},
+		ProviderAllowlists: providerAllowlists,
+		PromptCapture:      guardrailPromptCapture(capture),
 	}
 	out, err := json.Marshal(cfg)
 	if err != nil {
@@ -843,6 +847,74 @@ func marshalGuardrailConfig(merged MergedGuardrails) ([]byte, error) {
 	return out, nil
 }
 
+// buildProviderAllowlists returns the proxy's per-provider backstop: a provider
+// is included only when every authorising policy restricts models (their union);
+// if any leaves it unrestricted it is omitted, so management decides per group.
+func buildProviderAllowlists(policies []*types.Policy, byID map[string]*types.Guardrail) map[string][]string {
+	type providerAcc struct {
+		models          map[string]struct{}
+		anyUnrestricted bool
+	}
+	accs := make(map[string]*providerAcc)
+	for _, p := range policies {
+		if p == nil {
+			continue
+		}
+		restricted, models := policyModelAllowlist(p, byID)
+		for _, providerID := range p.DestinationProviderIDs {
+			if providerID == "" {
+				continue
+			}
+			acc, ok := accs[providerID]
+			if !ok {
+				acc = &providerAcc{models: make(map[string]struct{})}
+				accs[providerID] = acc
+			}
+			if !restricted {
+				acc.anyUnrestricted = true
+				continue
+			}
+			for _, m := range models {
+				acc.models[m] = struct{}{}
+			}
+		}
+	}
+	out := make(map[string][]string, len(accs))
+	for providerID, acc := range accs {
+		if acc.anyUnrestricted {
+			continue
+		}
+		models := make([]string, 0, len(acc.models))
+		for m := range acc.models {
+			models = append(models, m)
+		}
+		sort.Strings(models)
+		out[providerID] = models
+	}
+	return out
+}
+
+// policyModelAllowlist reports whether a policy restricts models (has an
+// allowlist-enabled guardrail) and the union of allowed models. Models are
+// verbatim; the proxy factory lowercases/trims them at decode time.
+func policyModelAllowlist(p *types.Policy, byID map[string]*types.Guardrail) (bool, []string) {
+	restricted := false
+	var models []string
+	for _, gID := range p.GuardrailIDs {
+		g, ok := byID[gID]
+		if !ok || g == nil || !g.Checks.ModelAllowlist.Enabled {
+			continue
+		}
+		restricted = true
+		for _, m := range g.Checks.ModelAllowlist.Models {
+			if m != "" {
+				models = append(models, m)
+			}
+		}
+	}
+	return restricted, models
+}
+
 // buildAccountService composes the per-account gateway Service. The
 // target carries the noop placeholder URL — the router middleware
 // rewrites every request to the matched provider's upstream before the
@@ -986,38 +1058,11 @@ func unionSourceGroups(policies []*types.Policy) []string {
 	return out
 }
 
-// MergedGuardrails is the JSON shape passed to the proxy via the
-// guardrail middleware's config_json. Mirrors the proxy-side
-// expectations and is intentionally distinct from
-// types.GuardrailChecks so we can evolve either side independently.
+// MergedGuardrails is the synthesiser's fold target. Only prompt capture is
+// merged here — the model allowlist is emitted per-provider, and
+// token/budget/retention moved onto Policy.Limits and account Settings.
 type MergedGuardrails struct {
-	ModelAllowlist []string            `json:"model_allowlist,omitempty"`
-	TokenLimits    MergedTokenLimits   `json:"token_limits"`
-	Budget         MergedBudget        `json:"budget"`
-	PromptCapture  MergedPromptCapture `json:"prompt_capture"`
-	Retention      MergedRetention     `json:"retention"`
-}
-
-type MergedTokenLimits struct {
-	Hourly  *MergedTokenWindow `json:"hourly,omitempty"`
-	Daily   *MergedTokenWindow `json:"daily,omitempty"`
-	Monthly *MergedTokenWindow `json:"monthly,omitempty"`
-}
-
-type MergedTokenWindow struct {
-	MaxInputTokens  int `json:"max_input_tokens,omitempty"`
-	MaxOutputTokens int `json:"max_output_tokens,omitempty"`
-}
-
-type MergedBudget struct {
-	Hourly  *MergedBudgetWindow `json:"hourly,omitempty"`
-	Daily   *MergedBudgetWindow `json:"daily,omitempty"`
-	Monthly *MergedBudgetWindow `json:"monthly,omitempty"`
-}
-
-type MergedBudgetWindow struct {
-	SoftCapUSD float64 `json:"soft_cap_usd,omitempty"`
-	HardCapUSD float64 `json:"hard_cap_usd,omitempty"`
+	PromptCapture MergedPromptCapture
 }
 
 type MergedPromptCapture struct {
@@ -1025,64 +1070,31 @@ type MergedPromptCapture struct {
 	RedactPii bool `json:"redact_pii"`
 }
 
-type MergedRetention struct {
-	Enabled bool `json:"enabled"`
-	Days    int  `json:"days"`
-}
-
-// mergeGuardrails computes the effective guardrail spec applied at the
-// proxy, given the referencing policies and the account's guardrail
-// catalogue. Policy enabled-ness is the caller's responsibility — only
-// enabled policies should be passed in.
+// mergeGuardrails folds the referencing policies' guardrails into the
+// prompt-capture decision only. The model allowlist is enforced per-policy/group
+// in management and shipped per-provider; token/budget/retention live off
+// guardrails now.
 //
-// Merge rules:
-//   - Model allowlist:   union of allowlists across policies that enable it.
-//   - Token / Budget:    most-restrictive (min of non-zero caps) per window.
-//   - Prompt capture:    enabled if any policy enables it; redact_pii sticks
-//     if any enabling policy turns it on.
-//   - Retention:         enabled if any enables it; smallest non-zero days wins.
+// Merge rule — prompt capture: enabled if any policy enables it; redact_pii
+// sticks if any enabling policy turns it on.
 func mergeGuardrails(policies []*types.Policy, byID map[string]*types.Guardrail) MergedGuardrails {
 	merged := MergedGuardrails{}
-	allowlist := make(map[string]struct{})
-	allowlistEnabled := false
-
 	for _, policy := range policies {
 		for _, gID := range policy.GuardrailIDs {
 			g, ok := byID[gID]
 			if !ok || g == nil {
 				continue
 			}
-			mergeGuardrail(g, &merged, allowlist, &allowlistEnabled)
+			mergeGuardrail(g, &merged)
 		}
 	}
-
-	if allowlistEnabled {
-		merged.ModelAllowlist = make([]string, 0, len(allowlist))
-		for m := range allowlist {
-			merged.ModelAllowlist = append(merged.ModelAllowlist, m)
-		}
-		sort.Strings(merged.ModelAllowlist)
-	}
 	return merged
 }
 
-// mergeGuardrail folds a single guardrail's enabled checks into the
-// running merge: model-allowlist models join the shared set (and flip
-// allowlistEnabled), and prompt-capture / redact-pii stick once any
-// enabling guardrail turns them on.
-//
-// TokenLimits, Budget, and Retention have moved off guardrails — token
-// and budget caps now live on the Policy itself (Policy.Limits) and
-// retention moves to account-level Settings — so they are not merged here.
-func mergeGuardrail(g *types.Guardrail, merged *MergedGuardrails, allowlist map[string]struct{}, allowlistEnabled *bool) {
-	if g.Checks.ModelAllowlist.Enabled {
-		*allowlistEnabled = true
-		for _, m := range g.Checks.ModelAllowlist.Models {
-			if m != "" {
-				allowlist[m] = struct{}{}
-			}
-		}
-	}
+// mergeGuardrail folds a single guardrail's prompt-capture settings into the
+// running merge: enabled / redact-pii stick once any enabling guardrail turns
+// them on.
+func mergeGuardrail(g *types.Guardrail, merged *MergedGuardrails) {
 	if g.Checks.PromptCapture.Enabled {
 		merged.PromptCapture.Enabled = true
 		if g.Checks.PromptCapture.RedactPii {

management/internals/modules/agentnetwork/synthesizer_guardrail_realstore_test.go

@@ -81,8 +81,8 @@ func TestSynthesizeServices_RealStore_PromptCaptureAccountIsSoleControl(t *testi
 	require.Len(t, services, 1)
 
 	cfg := decodeServiceGuardrailConfig(t, services[0])
-	assert.Equal(t, []string{"gpt-5.4"}, cfg.ModelAllowlist,
-		"model allowlist is a pure policy guardrail and must always reach the config")
+	assert.Equal(t, map[string][]string{"prov-1": {"gpt-5.4"}}, cfg.ProviderAllowlists,
+		"model allowlist is a pure policy guardrail and must reach the per-provider config")
 	assert.False(t, cfg.PromptCapture.Enabled,
 		"prompt capture must be off when the account toggle is off, even with a capture-enabled guardrail")
 }
@@ -172,7 +172,7 @@ func TestSynthesizeServices_RealStore_NoGuardrail_CaptureOff(t *testing.T) {
 	require.Len(t, services, 1, "exactly one synth service expected")
 
 	cfg := decodeServiceGuardrailConfig(t, services[0])
-	assert.Empty(t, cfg.ModelAllowlist, "no guardrail → no allowlist")
+	assert.Empty(t, cfg.ProviderAllowlists, "no guardrail → provider unrestricted (absent from map)")
 	assert.False(t, cfg.PromptCapture.Enabled, "no guardrail → prompt capture off by default")
 	assert.False(t, cfg.PromptCapture.RedactPii, "no guardrail → redact off by default")
 }

management/internals/modules/agentnetwork/synthesizer_provider_allowlist_test.go

@@ -0,0 +1,95 @@
+package agentnetwork
+
+import (
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+
+	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork/types"
+)
+
+// policyForProviders builds an enabled policy authorising the given providers
+// under the given guardrails (both optional). Groups are irrelevant to
+// buildProviderAllowlists, which keys purely on destination provider.
+func policyForProviders(id string, guardrailIDs []string, providerIDs ...string) *types.Policy {
+	return &types.Policy{
+		ID:                     id,
+		Enabled:                true,
+		DestinationProviderIDs: providerIDs,
+		GuardrailIDs:           guardrailIDs,
+	}
+}
+
+func TestBuildProviderAllowlists(t *testing.T) {
+	byID := map[string]*types.Guardrail{
+		"g-4o":       allowlistGuardrail("g-4o", "acc-1", "gpt-4o"),
+		"g-opus":     allowlistGuardrail("g-opus", "acc-1", "claude-opus-4"),
+		"g-disabled": {ID: "g-disabled", Checks: types.GuardrailChecks{ModelAllowlist: types.GuardrailModelAllowlist{Enabled: false, Models: []string{"gpt-4o"}}}},
+	}
+
+	t.Run("all authorising policies restrict yields per-provider union", func(t *testing.T) {
+		policies := []*types.Policy{
+			policyForProviders("p1", []string{"g-4o"}, "prov-x"),
+			policyForProviders("p2", []string{"g-opus"}, "prov-x"),
+		}
+		got := buildProviderAllowlists(policies, byID)
+		assert.Equal(t, map[string][]string{"prov-x": {"claude-opus-4", "gpt-4o"}}, got,
+			"a provider every policy restricts carries the sorted union of their models")
+	})
+
+	t.Run("any un-guardrailed policy leaves the provider unrestricted (omitted)", func(t *testing.T) {
+		policies := []*types.Policy{
+			policyForProviders("p1", []string{"g-4o"}, "prov-x"),
+			policyForProviders("p2", nil, "prov-x"), // no guardrail
+		}
+		got := buildProviderAllowlists(policies, byID)
+		assert.NotContains(t, got, "prov-x",
+			"a provider reachable by an un-guardrailed policy must be omitted so the proxy treats it as unrestricted")
+	})
+
+	t.Run("a disabled allowlist counts as unrestricted", func(t *testing.T) {
+		policies := []*types.Policy{
+			policyForProviders("p1", []string{"g-disabled"}, "prov-x"),
+		}
+		got := buildProviderAllowlists(policies, byID)
+		assert.NotContains(t, got, "prov-x",
+			"a policy whose only guardrail has a disabled allowlist is unrestricted")
+	})
+
+	t.Run("providers are isolated from one another", func(t *testing.T) {
+		policies := []*types.Policy{
+			policyForProviders("p1", []string{"g-4o"}, "prov-x"),
+			policyForProviders("p2", []string{"g-opus"}, "prov-y"),
+		}
+		got := buildProviderAllowlists(policies, byID)
+		assert.Equal(t, []string{"gpt-4o"}, got["prov-x"], "prov-x keeps only its own model")
+		assert.Equal(t, []string{"claude-opus-4"}, got["prov-y"], "prov-y keeps only its own model")
+	})
+
+	t.Run("one policy authorising two providers restricts both", func(t *testing.T) {
+		policies := []*types.Policy{
+			policyForProviders("p1", []string{"g-4o"}, "prov-x", "prov-y"),
+		}
+		got := buildProviderAllowlists(policies, byID)
+		assert.Equal(t, []string{"gpt-4o"}, got["prov-x"])
+		assert.Equal(t, []string{"gpt-4o"}, got["prov-y"])
+	})
+
+	t.Run("union across a single policy's guardrails", func(t *testing.T) {
+		policies := []*types.Policy{
+			policyForProviders("p1", []string{"g-4o", "g-opus"}, "prov-x"),
+		}
+		got := buildProviderAllowlists(policies, byID)
+		assert.ElementsMatch(t, []string{"claude-opus-4", "gpt-4o"}, got["prov-x"],
+			"a policy's own multiple allowlist guardrails union together")
+	})
+
+	t.Run("an enabled allowlist with no models denies everything", func(t *testing.T) {
+		empty := map[string]*types.Guardrail{"g-empty": allowlistGuardrail("g-empty", "acc-1")}
+		got := buildProviderAllowlists([]*types.Policy{
+			policyForProviders("p1", []string{"g-empty"}, "prov-x"),
+		}, empty)
+		assert.Equal(t, map[string][]string{"prov-x": {}}, got,
+			"an enabled-but-empty allowlist is restricted with an empty set, not unrestricted")
+	})
+}

management/internals/modules/agentnetwork/synthesizer_test.go

@@ -1031,8 +1031,12 @@ func TestSynthesizeServices_GuardrailMerge_AllowlistUnion_LimitsRestrictive(t *t
 
 	var cfg guardrailConfig
 	require.NoError(t, json.Unmarshal(guardrailJSON, &cfg), "guardrail config must unmarshal cleanly")
-	assert.ElementsMatch(t, []string{"gpt-5.4-mini", "gpt-5.4-pro"}, cfg.ModelAllowlist,
-		"model allowlist union must keep both models")
+	// Both policies restrict the same provider, so the per-provider backstop
+	// carries the union of their models — a coarse gate that management's
+	// per-policy/group check narrows; it only blocks models outside the union
+	// when management is down.
+	assert.ElementsMatch(t, []string{"gpt-5.4-mini", "gpt-5.4-pro"}, cfg.ProviderAllowlists["prov-1"],
+		"per-provider allowlist union must keep both models")
 }
 
 func TestSynthesizeServices_BackfillsMissingSessionKeys(t *testing.T) {

management/internals/modules/reverseproxy/domain/domain.go

@@ -23,9 +23,6 @@ type Domain struct {
 	// SupportsCrowdSec is populated at query time from proxy cluster capabilities.
 	// Not persisted.
 	SupportsCrowdSec *bool `gorm:"-"`
-	// SupportsAppSec is populated at query time from proxy cluster capabilities.
-	// Not persisted.
-	SupportsAppSec *bool `gorm:"-"`
 	// SupportsPrivate is populated at query time from proxy cluster capabilities. Not persisted.
 	SupportsPrivate *bool `gorm:"-"`
 }

management/internals/modules/reverseproxy/domain/manager/api.go

@@ -49,7 +49,6 @@ func domainToApi(d *domain.Domain) api.ReverseProxyDomain {
 		SupportsCustomPorts: d.SupportsCustomPorts,
 		RequireSubdomain:    d.RequireSubdomain,
 		SupportsCrowdsec:    d.SupportsCrowdSec,
-		SupportsAppsec:      d.SupportsAppSec,
 		SupportsPrivate:     d.SupportsPrivate,
 	}
 	if d.TargetCluster != "" {

management/internals/modules/reverseproxy/domain/manager/manager.go

@@ -35,7 +35,6 @@ type proxyManager interface {
 	ClusterSupportsCustomPorts(ctx context.Context, clusterAddr string) *bool
 	ClusterRequireSubdomain(ctx context.Context, clusterAddr string) *bool
 	ClusterSupportsCrowdSec(ctx context.Context, clusterAddr string) *bool
-	ClusterSupportsAppSec(ctx context.Context, clusterAddr string) *bool
 	ClusterSupportsPrivate(ctx context.Context, clusterAddr string) *bool
 }
 
@@ -95,7 +94,6 @@ func (m Manager) GetDomains(ctx context.Context, accountID, userID string) ([]*d
 		d.SupportsCustomPorts = m.proxyManager.ClusterSupportsCustomPorts(ctx, cluster)
 		d.RequireSubdomain = m.proxyManager.ClusterRequireSubdomain(ctx, cluster)
 		d.SupportsCrowdSec = m.proxyManager.ClusterSupportsCrowdSec(ctx, cluster)
-		d.SupportsAppSec = m.proxyManager.ClusterSupportsAppSec(ctx, cluster)
 		d.SupportsPrivate = m.proxyManager.ClusterSupportsPrivate(ctx, cluster)
 		ret = append(ret, d)
 	}
@@ -113,7 +111,6 @@ func (m Manager) GetDomains(ctx context.Context, accountID, userID string) ([]*d
 		if d.TargetCluster != "" {
 			cd.SupportsCustomPorts = m.proxyManager.ClusterSupportsCustomPorts(ctx, d.TargetCluster)
 			cd.SupportsCrowdSec = m.proxyManager.ClusterSupportsCrowdSec(ctx, d.TargetCluster)
-			cd.SupportsAppSec = m.proxyManager.ClusterSupportsAppSec(ctx, d.TargetCluster)
 			cd.SupportsPrivate = m.proxyManager.ClusterSupportsPrivate(ctx, d.TargetCluster)
 		}
 		// Custom domains never require a subdomain by default since

management/internals/modules/reverseproxy/domain/manager/manager_test.go

@@ -40,10 +40,6 @@ func (m *mockProxyManager) ClusterSupportsCrowdSec(_ context.Context, _ string)
 	return nil
 }
 
-func (m *mockProxyManager) ClusterSupportsAppSec(_ context.Context, _ string) *bool {
-	return nil
-}
-
 func (m *mockProxyManager) ClusterSupportsPrivate(_ context.Context, _ string) *bool {
 	return nil
 }

management/internals/modules/reverseproxy/proxy/manager.go

@@ -19,7 +19,6 @@ type Manager interface {
 	ClusterSupportsCustomPorts(ctx context.Context, clusterAddr string) *bool
 	ClusterRequireSubdomain(ctx context.Context, clusterAddr string) *bool
 	ClusterSupportsCrowdSec(ctx context.Context, clusterAddr string) *bool
-	ClusterSupportsAppSec(ctx context.Context, clusterAddr string) *bool
 	ClusterSupportsPrivate(ctx context.Context, clusterAddr string) *bool
 	CleanupStale(ctx context.Context, inactivityDuration time.Duration) error
 	GetAccountProxy(ctx context.Context, accountID string) (*Proxy, error)

management/internals/modules/reverseproxy/proxy/manager/manager.go

@@ -21,7 +21,6 @@ type store interface {
 	GetClusterSupportsCustomPorts(ctx context.Context, clusterAddr string) *bool
 	GetClusterRequireSubdomain(ctx context.Context, clusterAddr string) *bool
 	GetClusterSupportsCrowdSec(ctx context.Context, clusterAddr string) *bool
-	GetClusterSupportsAppSec(ctx context.Context, clusterAddr string) *bool
 	GetClusterSupportsPrivate(ctx context.Context, clusterAddr string) *bool
 	CleanupStaleProxies(ctx context.Context, inactivityDuration time.Duration) error
 	GetProxyByAccountID(ctx context.Context, accountID string) (*proxy.Proxy, error)
@@ -139,13 +138,6 @@ func (m Manager) ClusterSupportsCrowdSec(ctx context.Context, clusterAddr string
 	return m.store.GetClusterSupportsCrowdSec(ctx, clusterAddr)
 }
 
-// ClusterSupportsAppSec returns whether all active proxies in the cluster have
-// a CrowdSec AppSec endpoint configured (unanimous). Returns nil when no proxy
-// has reported capabilities.
-func (m Manager) ClusterSupportsAppSec(ctx context.Context, clusterAddr string) *bool {
-	return m.store.GetClusterSupportsAppSec(ctx, clusterAddr)
-}
-
 // ClusterSupportsPrivate reports whether any active proxy claims the private capability (nil = unreported).
 func (m Manager) ClusterSupportsPrivate(ctx context.Context, clusterAddr string) *bool {
 	return m.store.GetClusterSupportsPrivate(ctx, clusterAddr)

management/internals/modules/reverseproxy/proxy/manager/manager_test.go

@@ -99,9 +99,6 @@ func (m *mockStore) GetClusterRequireSubdomain(_ context.Context, _ string) *boo
 func (m *mockStore) GetClusterSupportsCrowdSec(_ context.Context, _ string) *bool {
 	return nil
 }
-func (m *mockStore) GetClusterSupportsAppSec(_ context.Context, _ string) *bool {
-	return nil
-}
 func (m *mockStore) GetClusterSupportsPrivate(_ context.Context, _ string) *bool {
 	return nil
 }

management/internals/modules/reverseproxy/proxy/manager_mock.go

@@ -50,6 +50,20 @@ func (mr *MockManagerMockRecorder) CleanupStale(ctx, inactivityDuration interfac
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "CleanupStale", reflect.TypeOf((*MockManager)(nil).CleanupStale), ctx, inactivityDuration)
 }
 
+// ClusterSupportsCustomPorts mocks base method.
+func (m *MockManager) ClusterSupportsCustomPorts(ctx context.Context, clusterAddr string) *bool {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "ClusterSupportsCustomPorts", ctx, clusterAddr)
+	ret0, _ := ret[0].(*bool)
+	return ret0
+}
+
+// ClusterSupportsCustomPorts indicates an expected call of ClusterSupportsCustomPorts.
+func (mr *MockManagerMockRecorder) ClusterSupportsCustomPorts(ctx, clusterAddr interface{}) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "ClusterSupportsCustomPorts", reflect.TypeOf((*MockManager)(nil).ClusterSupportsCustomPorts), ctx, clusterAddr)
+}
+
 // ClusterRequireSubdomain mocks base method.
 func (m *MockManager) ClusterRequireSubdomain(ctx context.Context, clusterAddr string) *bool {
 	m.ctrl.T.Helper()
@@ -64,20 +78,6 @@ func (mr *MockManagerMockRecorder) ClusterRequireSubdomain(ctx, clusterAddr inte
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "ClusterRequireSubdomain", reflect.TypeOf((*MockManager)(nil).ClusterRequireSubdomain), ctx, clusterAddr)
 }
 
-// ClusterSupportsAppSec mocks base method.
-func (m *MockManager) ClusterSupportsAppSec(ctx context.Context, clusterAddr string) *bool {
-	m.ctrl.T.Helper()
-	ret := m.ctrl.Call(m, "ClusterSupportsAppSec", ctx, clusterAddr)
-	ret0, _ := ret[0].(*bool)
-	return ret0
-}
-
-// ClusterSupportsAppSec indicates an expected call of ClusterSupportsAppSec.
-func (mr *MockManagerMockRecorder) ClusterSupportsAppSec(ctx, clusterAddr interface{}) *gomock.Call {
-	mr.mock.ctrl.T.Helper()
-	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "ClusterSupportsAppSec", reflect.TypeOf((*MockManager)(nil).ClusterSupportsAppSec), ctx, clusterAddr)
-}
-
 // ClusterSupportsCrowdSec mocks base method.
 func (m *MockManager) ClusterSupportsCrowdSec(ctx context.Context, clusterAddr string) *bool {
 	m.ctrl.T.Helper()
@@ -92,20 +92,6 @@ func (mr *MockManagerMockRecorder) ClusterSupportsCrowdSec(ctx, clusterAddr inte
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "ClusterSupportsCrowdSec", reflect.TypeOf((*MockManager)(nil).ClusterSupportsCrowdSec), ctx, clusterAddr)
 }
 
-// ClusterSupportsCustomPorts mocks base method.
-func (m *MockManager) ClusterSupportsCustomPorts(ctx context.Context, clusterAddr string) *bool {
-	m.ctrl.T.Helper()
-	ret := m.ctrl.Call(m, "ClusterSupportsCustomPorts", ctx, clusterAddr)
-	ret0, _ := ret[0].(*bool)
-	return ret0
-}
-
-// ClusterSupportsCustomPorts indicates an expected call of ClusterSupportsCustomPorts.
-func (mr *MockManagerMockRecorder) ClusterSupportsCustomPorts(ctx, clusterAddr interface{}) *gomock.Call {
-	mr.mock.ctrl.T.Helper()
-	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "ClusterSupportsCustomPorts", reflect.TypeOf((*MockManager)(nil).ClusterSupportsCustomPorts), ctx, clusterAddr)
-}
-
 // ClusterSupportsPrivate mocks base method.
 func (m *MockManager) ClusterSupportsPrivate(ctx context.Context, clusterAddr string) *bool {
 	m.ctrl.T.Helper()
@@ -135,35 +121,6 @@ func (mr *MockManagerMockRecorder) Connect(ctx, proxyID, sessionID, clusterAddre
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "Connect", reflect.TypeOf((*MockManager)(nil).Connect), ctx, proxyID, sessionID, clusterAddress, ipAddress, accountID, capabilities)
 }
 
-// CountAccountProxies mocks base method.
-func (m *MockManager) CountAccountProxies(ctx context.Context, accountID string) (int64, error) {
-	m.ctrl.T.Helper()
-	ret := m.ctrl.Call(m, "CountAccountProxies", ctx, accountID)
-	ret0, _ := ret[0].(int64)
-	ret1, _ := ret[1].(error)
-	return ret0, ret1
-}
-
-// CountAccountProxies indicates an expected call of CountAccountProxies.
-func (mr *MockManagerMockRecorder) CountAccountProxies(ctx, accountID interface{}) *gomock.Call {
-	mr.mock.ctrl.T.Helper()
-	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "CountAccountProxies", reflect.TypeOf((*MockManager)(nil).CountAccountProxies), ctx, accountID)
-}
-
-// DeleteAccountCluster mocks base method.
-func (m *MockManager) DeleteAccountCluster(ctx context.Context, clusterAddress, accountID string) error {
-	m.ctrl.T.Helper()
-	ret := m.ctrl.Call(m, "DeleteAccountCluster", ctx, clusterAddress, accountID)
-	ret0, _ := ret[0].(error)
-	return ret0
-}
-
-// DeleteAccountCluster indicates an expected call of DeleteAccountCluster.
-func (mr *MockManagerMockRecorder) DeleteAccountCluster(ctx, clusterAddress, accountID interface{}) *gomock.Call {
-	mr.mock.ctrl.T.Helper()
-	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "DeleteAccountCluster", reflect.TypeOf((*MockManager)(nil).DeleteAccountCluster), ctx, clusterAddress, accountID)
-}
-
 // Disconnect mocks base method.
 func (m *MockManager) Disconnect(ctx context.Context, proxyID, sessionID string) error {
 	m.ctrl.T.Helper()
@@ -178,21 +135,6 @@ func (mr *MockManagerMockRecorder) Disconnect(ctx, proxyID, sessionID interface{
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "Disconnect", reflect.TypeOf((*MockManager)(nil).Disconnect), ctx, proxyID, sessionID)
 }
 
-// GetAccountProxy mocks base method.
-func (m *MockManager) GetAccountProxy(ctx context.Context, accountID string) (*Proxy, error) {
-	m.ctrl.T.Helper()
-	ret := m.ctrl.Call(m, "GetAccountProxy", ctx, accountID)
-	ret0, _ := ret[0].(*Proxy)
-	ret1, _ := ret[1].(error)
-	return ret0, ret1
-}
-
-// GetAccountProxy indicates an expected call of GetAccountProxy.
-func (mr *MockManagerMockRecorder) GetAccountProxy(ctx, accountID interface{}) *gomock.Call {
-	mr.mock.ctrl.T.Helper()
-	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetAccountProxy", reflect.TypeOf((*MockManager)(nil).GetAccountProxy), ctx, accountID)
-}
-
 // GetActiveClusterAddresses mocks base method.
 func (m *MockManager) GetActiveClusterAddresses(ctx context.Context) ([]string, error) {
 	m.ctrl.T.Helper()
@@ -208,7 +150,6 @@ func (mr *MockManagerMockRecorder) GetActiveClusterAddresses(ctx interface{}) *g
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetActiveClusterAddresses", reflect.TypeOf((*MockManager)(nil).GetActiveClusterAddresses), ctx)
 }
 
-// GetActiveClusterAddressesForAccount mocks base method.
 func (m *MockManager) GetActiveClusterAddressesForAccount(ctx context.Context, accountID string) ([]string, error) {
 	m.ctrl.T.Helper()
 	ret := m.ctrl.Call(m, "GetActiveClusterAddressesForAccount", ctx, accountID)
@@ -217,7 +158,6 @@ func (m *MockManager) GetActiveClusterAddressesForAccount(ctx context.Context, a
 	return ret0, ret1
 }
 
-// GetActiveClusterAddressesForAccount indicates an expected call of GetActiveClusterAddressesForAccount.
 func (mr *MockManagerMockRecorder) GetActiveClusterAddressesForAccount(ctx, accountID interface{}) *gomock.Call {
 	mr.mock.ctrl.T.Helper()
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetActiveClusterAddressesForAccount", reflect.TypeOf((*MockManager)(nil).GetActiveClusterAddressesForAccount), ctx, accountID)
@@ -237,6 +177,36 @@ func (mr *MockManagerMockRecorder) Heartbeat(ctx, p interface{}) *gomock.Call {
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "Heartbeat", reflect.TypeOf((*MockManager)(nil).Heartbeat), ctx, p)
 }
 
+// GetAccountProxy mocks base method.
+func (m *MockManager) GetAccountProxy(ctx context.Context, accountID string) (*Proxy, error) {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "GetAccountProxy", ctx, accountID)
+	ret0, _ := ret[0].(*Proxy)
+	ret1, _ := ret[1].(error)
+	return ret0, ret1
+}
+
+// GetAccountProxy indicates an expected call of GetAccountProxy.
+func (mr *MockManagerMockRecorder) GetAccountProxy(ctx, accountID interface{}) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetAccountProxy", reflect.TypeOf((*MockManager)(nil).GetAccountProxy), ctx, accountID)
+}
+
+// CountAccountProxies mocks base method.
+func (m *MockManager) CountAccountProxies(ctx context.Context, accountID string) (int64, error) {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "CountAccountProxies", ctx, accountID)
+	ret0, _ := ret[0].(int64)
+	ret1, _ := ret[1].(error)
+	return ret0, ret1
+}
+
+// CountAccountProxies indicates an expected call of CountAccountProxies.
+func (mr *MockManagerMockRecorder) CountAccountProxies(ctx, accountID interface{}) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "CountAccountProxies", reflect.TypeOf((*MockManager)(nil).CountAccountProxies), ctx, accountID)
+}
+
 // IsClusterAddressAvailable mocks base method.
 func (m *MockManager) IsClusterAddressAvailable(ctx context.Context, clusterAddress, accountID string) (bool, error) {
 	m.ctrl.T.Helper()
@@ -252,6 +222,20 @@ func (mr *MockManagerMockRecorder) IsClusterAddressAvailable(ctx, clusterAddress
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "IsClusterAddressAvailable", reflect.TypeOf((*MockManager)(nil).IsClusterAddressAvailable), ctx, clusterAddress, accountID)
 }
 
+// DeleteAccountCluster mocks base method.
+func (m *MockManager) DeleteAccountCluster(ctx context.Context, clusterAddress, accountID string) error {
+	m.ctrl.T.Helper()
+	ret := m.ctrl.Call(m, "DeleteAccountCluster", ctx, clusterAddress, accountID)
+	ret0, _ := ret[0].(error)
+	return ret0
+}
+
+// DeleteAccountCluster indicates an expected call of DeleteAccountCluster.
+func (mr *MockManagerMockRecorder) DeleteAccountCluster(ctx, clusterAddress, accountID interface{}) *gomock.Call {
+	mr.mock.ctrl.T.Helper()
+	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "DeleteAccountCluster", reflect.TypeOf((*MockManager)(nil).DeleteAccountCluster), ctx, clusterAddress, accountID)
+}
+
 // MockController is a mock of Controller interface.
 type MockController struct {
 	ctrl     *gomock.Controller

management/internals/modules/reverseproxy/proxy/proxy.go

@@ -20,9 +20,6 @@ type Capabilities struct {
 	RequireSubdomain *bool
 	// SupportsCrowdsec indicates whether this proxy has CrowdSec configured.
 	SupportsCrowdsec *bool
-	// SupportsAppsec indicates whether this proxy has a CrowdSec AppSec (WAF)
-	// endpoint configured.
-	SupportsAppsec *bool
 	// Private indicates whether this proxy supports inbound access via Wireguard
 	// tunnel and netbird-only authentication policies
 	Private *bool
@@ -77,6 +74,5 @@ type Cluster struct {
 	SupportsCustomPorts *bool
 	RequireSubdomain    *bool
 	SupportsCrowdSec    *bool
-	SupportsAppSec      *bool
 	Private             *bool
 }

management/internals/modules/reverseproxy/service/manager/api.go

@@ -204,7 +204,6 @@ func (h *handler) getClusters(w http.ResponseWriter, r *http.Request) {
 			SupportsCustomPorts: c.SupportsCustomPorts,
 			RequireSubdomain:    c.RequireSubdomain,
 			SupportsCrowdsec:    c.SupportsCrowdSec,
-			SupportsAppsec:      c.SupportsAppSec,
 			Private:             c.Private,
 		})
 	}

management/internals/modules/reverseproxy/service/manager/manager.go

@@ -82,7 +82,6 @@ type CapabilityProvider interface {
 	ClusterSupportsCustomPorts(ctx context.Context, clusterAddr string) *bool
 	ClusterRequireSubdomain(ctx context.Context, clusterAddr string) *bool
 	ClusterSupportsCrowdSec(ctx context.Context, clusterAddr string) *bool
-	ClusterSupportsAppSec(ctx context.Context, clusterAddr string) *bool
 	ClusterSupportsPrivate(ctx context.Context, clusterAddr string) *bool
 }
 
@@ -138,7 +137,6 @@ func (m *Manager) GetClusters(ctx context.Context, accountID, userID string) ([]
 		clusters[i].SupportsCustomPorts = m.capabilities.ClusterSupportsCustomPorts(ctx, clusters[i].Address)
 		clusters[i].RequireSubdomain = m.capabilities.ClusterRequireSubdomain(ctx, clusters[i].Address)
 		clusters[i].SupportsCrowdSec = m.capabilities.ClusterSupportsCrowdSec(ctx, clusters[i].Address)
-		clusters[i].SupportsAppSec = m.capabilities.ClusterSupportsAppSec(ctx, clusters[i].Address)
 		clusters[i].Private = m.capabilities.ClusterSupportsPrivate(ctx, clusters[i].Address)
 	}
 

management/internals/modules/reverseproxy/service/service.go

@@ -165,18 +165,6 @@ type AccessRestrictions struct {
 	AllowedCountries []string `json:"allowed_countries,omitempty" gorm:"serializer:json"`
 	BlockedCountries []string `json:"blocked_countries,omitempty" gorm:"serializer:json"`
 	CrowdSecMode     string   `json:"crowdsec_mode,omitempty" gorm:"serializer:json"`
-	// AppSecMode is the CrowdSec AppSec (WAF) request inspection mode: "",
-	// "off", "enforce", or "observe". HTTP services only.
-	AppSecMode string `json:"appsec_mode,omitempty" gorm:"serializer:json"`
-}
-
-// isEmpty reports whether no restriction is configured. Both conversions drop
-// the object entirely in that case, so a field missing from this check is
-// silently discarded on the way to the API and the proxy.
-func (r AccessRestrictions) isEmpty() bool {
-	return len(r.AllowedCIDRs) == 0 && len(r.BlockedCIDRs) == 0 &&
-		len(r.AllowedCountries) == 0 && len(r.BlockedCountries) == 0 &&
-		r.CrowdSecMode == "" && r.AppSecMode == ""
 }
 
 // Copy returns a deep copy of the AccessRestrictions.
@@ -187,7 +175,6 @@ func (r AccessRestrictions) Copy() AccessRestrictions {
 		AllowedCountries: slices.Clone(r.AllowedCountries),
 		BlockedCountries: slices.Clone(r.BlockedCountries),
 		CrowdSecMode:     r.CrowdSecMode,
-		AppSecMode:       r.AppSecMode,
 	}
 }
 
@@ -821,17 +808,13 @@ func restrictionsFromAPI(r *api.AccessRestrictions) (AccessRestrictions, error)
 		}
 		res.CrowdSecMode = string(*r.CrowdsecMode)
 	}
-	if r.AppsecMode != nil {
-		if !r.AppsecMode.Valid() {
-			return AccessRestrictions{}, fmt.Errorf("invalid appsec_mode %q", *r.AppsecMode)
-		}
-		res.AppSecMode = string(*r.AppsecMode)
-	}
 	return res, nil
 }
 
 func restrictionsToAPI(r AccessRestrictions) *api.AccessRestrictions {
-	if r.isEmpty() {
+	if len(r.AllowedCIDRs) == 0 && len(r.BlockedCIDRs) == 0 &&
+		len(r.AllowedCountries) == 0 && len(r.BlockedCountries) == 0 &&
+		r.CrowdSecMode == "" {
 		return nil
 	}
 	res := &api.AccessRestrictions{}
@@ -851,15 +834,13 @@ func restrictionsToAPI(r AccessRestrictions) *api.AccessRestrictions {
 		mode := api.AccessRestrictionsCrowdsecMode(r.CrowdSecMode)
 		res.CrowdsecMode = &mode
 	}
-	if r.AppSecMode != "" {
-		mode := api.AccessRestrictionsAppsecMode(r.AppSecMode)
-		res.AppsecMode = &mode
-	}
 	return res
 }
 
 func restrictionsToProto(r AccessRestrictions) *proto.AccessRestrictions {
-	if r.isEmpty() {
+	if len(r.AllowedCIDRs) == 0 && len(r.BlockedCIDRs) == 0 &&
+		len(r.AllowedCountries) == 0 && len(r.BlockedCountries) == 0 &&
+		r.CrowdSecMode == "" {
 		return nil
 	}
 	return &proto.AccessRestrictions{
@@ -868,7 +849,6 @@ func restrictionsToProto(r AccessRestrictions) *proto.AccessRestrictions {
 		AllowedCountries: r.AllowedCountries,
 		BlockedCountries: r.BlockedCountries,
 		CrowdsecMode:     r.CrowdSecMode,
-		AppsecMode:       r.AppSecMode,
 	}
 }
 
@@ -894,11 +874,6 @@ func (s *Service) Validate() error {
 	if err := validateAccessRestrictions(&s.Restrictions); err != nil {
 		return err
 	}
-	// AppSec inspects HTTP requests, so it cannot apply to the L4 modes, which
-	// forward opaque byte streams.
-	if appSecEnabled(s.Restrictions.AppSecMode) && s.Mode != ModeHTTP {
-		return fmt.Errorf("appsec_mode is only supported for HTTP services, got mode %q", s.Mode)
-	}
 	if err := s.validatePrivateRequirements(); err != nil {
 		return err
 	}
@@ -1267,27 +1242,10 @@ func validateCrowdSecMode(mode string) error {
 	}
 }
 
-func validateAppSecMode(mode string) error {
-	switch mode {
-	case "", "off", "enforce", "observe":
-		return nil
-	default:
-		return fmt.Errorf("appsec_mode %q is invalid", mode)
-	}
-}
-
-// appSecEnabled reports whether the mode asks for request inspection.
-func appSecEnabled(mode string) bool {
-	return mode == "enforce" || mode == "observe"
-}
-
 func validateAccessRestrictions(r *AccessRestrictions) error {
 	if err := validateCrowdSecMode(r.CrowdSecMode); err != nil {
 		return err
 	}
-	if err := validateAppSecMode(r.AppSecMode); err != nil {
-		return err
-	}
 
 	if len(r.AllowedCIDRs) > maxCIDREntries {
 		return fmt.Errorf("allowed_cidrs: exceeds maximum of %d entries", maxCIDREntries)

management/internals/modules/reverseproxy/service/service_test.go

@@ -26,17 +26,6 @@ func validProxy() *Service {
 	}
 }
 
-// validL4Proxy returns a service that passes validation in one of the L4 modes.
-func validL4Proxy(mode string) *Service {
-	rp := validProxy()
-	rp.Mode = mode
-	rp.ListenPort = 9000
-	rp.Targets = []*Target{
-		{TargetId: "peer-1", TargetType: TargetTypePeer, Host: "10.0.0.1", Port: 5432, Protocol: mode, Enabled: true},
-	}
-	return rp
-}
-
 func TestValidate_Valid(t *testing.T) {
 	require.NoError(t, validProxy().Validate())
 }
@@ -1326,68 +1315,3 @@ func TestValidate_Private_RejectsNonHTTPMode(t *testing.T) {
 	}}
 	assert.ErrorContains(t, rp.Validate(), "HTTP")
 }
-
-func TestRestrictions_AppSecMode_RoundTrip(t *testing.T) {
-	mode := api.AccessRestrictionsAppsecModeEnforce
-	apiIn := &api.AccessRestrictions{AppsecMode: &mode}
-
-	model, err := restrictionsFromAPI(apiIn)
-	require.NoError(t, err)
-	assert.Equal(t, "enforce", model.AppSecMode)
-
-	// appsec_mode alone must keep the restrictions object alive on both the API
-	// and proto legs: it is meaningful without any CIDR or country entry.
-	apiOut := restrictionsToAPI(model)
-	require.NotNil(t, apiOut, "appsec_mode alone must not collapse the restrictions to nil")
-	require.NotNil(t, apiOut.AppsecMode)
-	assert.Equal(t, api.AccessRestrictionsAppsecModeEnforce, *apiOut.AppsecMode)
-
-	protoOut := restrictionsToProto(model)
-	require.NotNil(t, protoOut, "appsec_mode alone must reach the proxy")
-	assert.Equal(t, "enforce", protoOut.AppsecMode)
-}
-
-func TestRestrictions_AppSecMode_EmptyIsOmitted(t *testing.T) {
-	model, err := restrictionsFromAPI(&api.AccessRestrictions{
-		AllowedCidrs: &[]string{"203.0.113.0/24"},
-	})
-	require.NoError(t, err)
-	assert.Empty(t, model.AppSecMode)
-
-	apiOut := restrictionsToAPI(model)
-	require.NotNil(t, apiOut)
-	assert.Nil(t, apiOut.AppsecMode, "empty appsec_mode is omitted from the API response")
-}
-
-func TestRestrictions_AppSecMode_CopyIsDeep(t *testing.T) {
-	original := AccessRestrictions{AppSecMode: "observe", CrowdSecMode: "enforce"}
-	assert.Equal(t, original, original.Copy(), "Copy must carry every mode field")
-}
-
-func TestValidate_RejectsInvalidAppSecMode(t *testing.T) {
-	rp := validProxy()
-	rp.Restrictions = AccessRestrictions{AppSecMode: "sometimes"}
-	assert.ErrorContains(t, rp.Validate(), "appsec_mode")
-}
-
-func TestValidate_RejectsAppSecOnL4Modes(t *testing.T) {
-	// AppSec inspects HTTP requests, so the L4 modes cannot honor it. Accepting
-	// the field there would report protection that never runs.
-	for _, mode := range []string{ModeTCP, ModeUDP, ModeTLS} {
-		t.Run(mode, func(t *testing.T) {
-			rp := validL4Proxy(mode)
-			rp.Restrictions = AccessRestrictions{AppSecMode: "enforce"}
-			assert.ErrorContains(t, rp.Validate(), "appsec_mode is only supported for HTTP services")
-		})
-	}
-}
-
-func TestValidate_AllowsAppSecOffOnL4Modes(t *testing.T) {
-	for _, mode := range []string{ModeTCP, ModeUDP, ModeTLS} {
-		t.Run(mode, func(t *testing.T) {
-			rp := validL4Proxy(mode)
-			rp.Restrictions = AccessRestrictions{AppSecMode: "off"}
-			require.NoError(t, rp.Validate(), "an explicit off must not be rejected on L4 services")
-		})
-	}
-}

management/internals/shared/grpc/components_envelope_response.go

@@ -50,7 +50,7 @@ func ToComponentSyncResponse(
 	// TODO (dmitri) consider using invariants?
 	//
 	enableSSH := computeSSHEnabledForPeer(components, peer)
-	peerConfig := toPeerConfig(peer, components.Network, dnsName, settings, httpConfig, deviceFlowConfig, enableSSH)
+	peerConfig := toPeerConfig(peer, components.Network, dnsName, settings, httpConfig, deviceFlowConfig, enableSSH, components.ForceRoutingPeerDNSResolution)
 
 	includeIPv6 := peer.SupportsIPv6() && peer.IPv6.IsValid()
 	useSourcePrefixes := peer.SupportsSourcePrefixes()

management/internals/shared/grpc/conversion.go

@@ -119,7 +119,7 @@ func toNetbirdConfig(config *nbconfig.Config, turnCredentials *Token, relayToken
 	return nbConfig
 }
 
-func toPeerConfig(peer *nbpeer.Peer, network *types.Network, dnsName string, settings *types.Settings, httpConfig *nbconfig.HttpServerConfig, deviceFlowConfig *nbconfig.DeviceAuthorizationFlow, enableSSH bool) *proto.PeerConfig {
+func toPeerConfig(peer *nbpeer.Peer, network *types.Network, dnsName string, settings *types.Settings, httpConfig *nbconfig.HttpServerConfig, deviceFlowConfig *nbconfig.DeviceAuthorizationFlow, enableSSH bool, forceRoutingPeerDNS bool) *proto.PeerConfig {
 	netmask, _ := network.Net.Mask.Size()
 	fqdn := peer.FQDN(dnsName)
 
@@ -135,7 +135,7 @@ func toPeerConfig(peer *nbpeer.Peer, network *types.Network, dnsName string, set
 		Address:                         fmt.Sprintf("%s/%d", peer.IP.String(), netmask),
 		SshConfig:                       sshConfig,
 		Fqdn:                            fqdn,
-		RoutingPeerDnsResolutionEnabled: settings.RoutingPeerDNSResolutionEnabled,
+		RoutingPeerDnsResolutionEnabled: settings.RoutingPeerDNSResolutionEnabled || peer.ProxyMeta.Embedded || forceRoutingPeerDNS,
 		LazyConnectionEnabled:           settings.LazyConnectionEnabled,
 		AutoUpdate: &proto.AutoUpdateSettings{
 			Version:      settings.AutoUpdateVersion,
@@ -162,12 +162,12 @@ func ToSyncResponse(ctx context.Context, config *nbconfig.Config, httpConfig *nb
 	useSourcePrefixes := peer.SupportsSourcePrefixes()
 
 	response := &proto.SyncResponse{
-		PeerConfig: toPeerConfig(peer, networkMap.Network, dnsName, settings, httpConfig, deviceFlowConfig, networkMap.EnableSSH),
+		PeerConfig: toPeerConfig(peer, networkMap.Network, dnsName, settings, httpConfig, deviceFlowConfig, networkMap.EnableSSH, networkMap.ForceRoutingPeerDNSResolution),
 		NetworkMap: &proto.NetworkMap{
 			Serial:     networkMap.Network.CurrentSerial(),
 			Routes:     networkmap.ToProtocolRoutes(networkMap.Routes),
 			DNSConfig:  networkmap.ToProtocolDNSConfig(networkMap.DNSConfig, dnsCache, dnsFwdPort),
-			PeerConfig: toPeerConfig(peer, networkMap.Network, dnsName, settings, httpConfig, deviceFlowConfig, networkMap.EnableSSH),
+			PeerConfig: toPeerConfig(peer, networkMap.Network, dnsName, settings, httpConfig, deviceFlowConfig, networkMap.EnableSSH, networkMap.ForceRoutingPeerDNSResolution),
 		},
 		Checks: toProtocolChecks(ctx, checks),
 	}

management/internals/shared/grpc/conversion_test.go

@@ -2,6 +2,7 @@ package grpc
 
 import (
 	"fmt"
+	"net"
 	"net/netip"
 	"reflect"
 	"testing"
@@ -14,6 +15,7 @@ import (
 	"github.com/netbirdio/netbird/management/internals/controllers/network_map"
 	"github.com/netbirdio/netbird/management/internals/controllers/network_map/controller/cache"
 	nbconfig "github.com/netbirdio/netbird/management/internals/server/config"
+	nbpeer "github.com/netbirdio/netbird/management/server/peer"
 	"github.com/netbirdio/netbird/management/server/types"
 	"github.com/netbirdio/netbird/shared/management/networkmap"
 )
@@ -301,3 +303,35 @@ func TestToNetbirdConfig_RelayInvariant(t *testing.T) {
 		assert.True(t, nbCfg.Metrics.Enabled, "metrics flag should carry the settings value")
 	})
 }
+
+func TestToPeerConfig_RoutingPeerDNSResolution(t *testing.T) {
+	network := &types.Network{Net: net.IPNet{IP: net.IPv4(100, 0, 0, 0), Mask: net.CIDRMask(8, 32)}}
+
+	newPeer := func(embedded bool) *nbpeer.Peer {
+		p := &nbpeer.Peer{IP: netip.MustParseAddr("100.0.0.1")}
+		p.ProxyMeta.Embedded = embedded
+		return p
+	}
+
+	tests := []struct {
+		name        string
+		globalFlag  bool
+		embedded    bool
+		forceParam  bool
+		wantEnabled bool
+	}{
+		{name: "global off, regular peer, no force", wantEnabled: false},
+		{name: "global on wins", globalFlag: true, wantEnabled: true},
+		{name: "embedded proxy peer forced", embedded: true, wantEnabled: true},
+		{name: "routing peer forced via param", forceParam: true, wantEnabled: true},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			settings := &types.Settings{RoutingPeerDNSResolutionEnabled: tt.globalFlag}
+			cfg := toPeerConfig(newPeer(tt.embedded), network, "netbird.selfhosted", settings, nil, nil, false, tt.forceParam)
+			assert.Equal(t, tt.wantEnabled, cfg.RoutingPeerDnsResolutionEnabled,
+				"RoutingPeerDnsResolutionEnabled should reflect global || embedded || forced")
+		})
+	}
+}

management/internals/shared/grpc/proxy.go

@@ -29,7 +29,6 @@ import (
 
 	"github.com/netbirdio/netbird/shared/management/domain"
 
-	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork"
 	"github.com/netbirdio/netbird/management/internals/modules/peers"
 	"github.com/netbirdio/netbird/management/internals/modules/reverseproxy/accesslogs"
 	"github.com/netbirdio/netbird/management/internals/modules/reverseproxy/proxy"
@@ -37,6 +36,7 @@ import (
 	"github.com/netbirdio/netbird/management/internals/modules/reverseproxy/sessionkey"
 	"github.com/netbirdio/netbird/management/server/idp"
 	"github.com/netbirdio/netbird/management/server/peer"
+	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork"
 	"github.com/netbirdio/netbird/management/server/types"
 	"github.com/netbirdio/netbird/management/server/users"
 	proxyauth "github.com/netbirdio/netbird/proxy/auth"
@@ -285,6 +285,7 @@ func (s *ProxyServiceServer) CheckLLMPolicyLimits(ctx context.Context, req *prot
 		UserID:     req.GetUserId(),
 		GroupIDs:   req.GetGroupIds(),
 		ProviderID: req.GetProviderId(),
+		Model:      req.GetModel(),
 	})
 	if err != nil {
 		log.WithContext(ctx).Errorf("select policy for request: %v", err)
@@ -505,7 +506,6 @@ func (s *ProxyServiceServer) registerProxyConnection(ctx context.Context, params
 			SupportsCustomPorts: c.SupportsCustomPorts,
 			RequireSubdomain:    c.RequireSubdomain,
 			SupportsCrowdsec:    c.SupportsCrowdsec,
-			SupportsAppsec:      c.SupportsAppsec,
 			Private:             c.Private,
 		}
 	}

management/internals/shared/grpc/proxy_llm_policy_limits_test.go

@@ -0,0 +1,138 @@
+package grpc
+
+import (
+	"context"
+	"errors"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+	"google.golang.org/grpc/codes"
+	grpcstatus "google.golang.org/grpc/status"
+
+	"github.com/netbirdio/netbird/management/internals/modules/agentnetwork"
+	"github.com/netbirdio/netbird/shared/management/proto"
+)
+
+// fakeAgentNetworkLimits records the PolicySelectionInput it was invoked with
+// and returns a pre-programmed result, so tests can assert what the handler
+// forwards to the selector.
+type fakeAgentNetworkLimits struct {
+	gotInput agentnetwork.PolicySelectionInput
+	result   *agentnetwork.PolicySelectionResult
+	err      error
+}
+
+func (f *fakeAgentNetworkLimits) SelectPolicyForRequest(_ context.Context, in agentnetwork.PolicySelectionInput) (*agentnetwork.PolicySelectionResult, error) {
+	f.gotInput = in
+	if f.err != nil {
+		return nil, f.err
+	}
+	return f.result, nil
+}
+
+func (f *fakeAgentNetworkLimits) RecordUsage(_ context.Context, _ agentnetwork.RecordUsageInput) error {
+	return nil
+}
+
+// TestCheckLLMPolicyLimits_ForwardsModelToSelector proves the wiring added here:
+// the model the proxy extracted must reach the selector's Model unchanged,
+// alongside the account/user/group/provider fields.
+func TestCheckLLMPolicyLimits_ForwardsModelToSelector(t *testing.T) {
+	fake := &fakeAgentNetworkLimits{result: &agentnetwork.PolicySelectionResult{Allow: true, SelectedPolicyID: "pol-1"}}
+	s := &ProxyServiceServer{}
+	s.SetAgentNetworkLimitsService(fake)
+
+	req := &proto.CheckLLMPolicyLimitsRequest{
+		AccountId:  "acc-1",
+		UserId:     "user-1",
+		GroupIds:   []string{"grp-a", "grp-b"},
+		ProviderId: "prov-1",
+		Model:      "claude-opus-4",
+	}
+
+	resp, err := s.CheckLLMPolicyLimits(context.Background(), req)
+	require.NoError(t, err)
+	require.NotNil(t, resp)
+
+	assert.Equal(t, "acc-1", fake.gotInput.AccountID)
+	assert.Equal(t, "user-1", fake.gotInput.UserID)
+	assert.Equal(t, []string{"grp-a", "grp-b"}, fake.gotInput.GroupIDs)
+	assert.Equal(t, "prov-1", fake.gotInput.ProviderID)
+	assert.Equal(t, "claude-opus-4", fake.gotInput.Model,
+		"the request's model must be forwarded to the selector")
+}
+
+// TestCheckLLMPolicyLimits_EmptyModelForwardedAsEmpty proves an undetermined
+// model (empty string) is forwarded as-is; the selector decides how to treat it.
+func TestCheckLLMPolicyLimits_EmptyModelForwardedAsEmpty(t *testing.T) {
+	fake := &fakeAgentNetworkLimits{result: &agentnetwork.PolicySelectionResult{Allow: true}}
+	s := &ProxyServiceServer{}
+	s.SetAgentNetworkLimitsService(fake)
+
+	req := &proto.CheckLLMPolicyLimitsRequest{
+		AccountId:  "acc-1",
+		ProviderId: "prov-1",
+	}
+
+	_, err := s.CheckLLMPolicyLimits(context.Background(), req)
+	require.NoError(t, err)
+	assert.Equal(t, "", fake.gotInput.Model, "an absent model must be forwarded as empty, not fabricated")
+}
+
+// TestCheckLLMPolicyLimits_DenyResponseCarriesModelBlockedCode proves the deny
+// envelope surfaces the model-allowlist deny code + reason through the response.
+func TestCheckLLMPolicyLimits_DenyResponseCarriesModelBlockedCode(t *testing.T) {
+	fake := &fakeAgentNetworkLimits{result: &agentnetwork.PolicySelectionResult{
+		Allow:      false,
+		DenyCode:   "llm_policy.model_blocked",
+		DenyReason: `model "claude-opus-4" is not permitted by any applicable policy allowlist`,
+	}}
+	s := &ProxyServiceServer{}
+	s.SetAgentNetworkLimitsService(fake)
+
+	resp, err := s.CheckLLMPolicyLimits(context.Background(), &proto.CheckLLMPolicyLimitsRequest{
+		AccountId:  "acc-1",
+		ProviderId: "prov-1",
+		Model:      "claude-opus-4",
+	})
+	require.NoError(t, err)
+	require.NotNil(t, resp)
+	assert.Equal(t, "deny", resp.Decision)
+	assert.Equal(t, "llm_policy.model_blocked", resp.DenyCode)
+	assert.NotEmpty(t, resp.DenyReason)
+	assert.Empty(t, resp.SelectedPolicyId, "a denied request must carry no selected policy")
+}
+
+// TestCheckLLMPolicyLimits_SelectorErrorSurfacesAsInternal proves a selector
+// failure surfaces as an Internal gRPC error rather than a silent allow.
+func TestCheckLLMPolicyLimits_SelectorErrorSurfacesAsInternal(t *testing.T) {
+	fake := &fakeAgentNetworkLimits{err: errors.New("boom")}
+	s := &ProxyServiceServer{}
+	s.SetAgentNetworkLimitsService(fake)
+
+	_, err := s.CheckLLMPolicyLimits(context.Background(), &proto.CheckLLMPolicyLimitsRequest{
+		AccountId:  "acc-1",
+		ProviderId: "prov-1",
+		Model:      "gpt-4o",
+	})
+	require.Error(t, err)
+	st, ok := grpcstatus.FromError(err)
+	require.True(t, ok)
+	assert.Equal(t, codes.Internal, st.Code(), "selector errors must never fail open on the hot path")
+}
+
+// TestCheckLLMPolicyLimits_UnconfiguredServiceReturnsUnimplemented locks the
+// fallback: with no limits service wired the RPC returns Unimplemented.
+func TestCheckLLMPolicyLimits_UnconfiguredServiceReturnsUnimplemented(t *testing.T) {
+	s := &ProxyServiceServer{}
+
+	_, err := s.CheckLLMPolicyLimits(context.Background(), &proto.CheckLLMPolicyLimitsRequest{
+		AccountId:  "acc-1",
+		ProviderId: "prov-1",
+	})
+	require.Error(t, err)
+	st, ok := grpcstatus.FromError(err)
+	require.True(t, ok)
+	assert.Equal(t, codes.Unimplemented, st.Code())
+}

management/internals/shared/grpc/server.go

@@ -921,7 +921,7 @@ func (s *Server) prepareLoginResponse(ctx context.Context, peer *nbpeer.Peer, ne
 	// if peer has reached this point then it has logged in
 	loginResp := &proto.LoginResponse{
 		NetbirdConfig: toNetbirdConfig(s.config, nil, relayToken, nil, settings),
-		PeerConfig:    toPeerConfig(peer, network, s.networkMapController.GetDNSDomain(settings), settings, s.config.HttpConfig, s.config.DeviceAuthorizationFlow, enableSSH),
+		PeerConfig:    toPeerConfig(peer, network, s.networkMapController.GetDNSDomain(settings), settings, s.config.HttpConfig, s.config.DeviceAuthorizationFlow, enableSSH, false),
 		Checks:        toProtocolChecks(ctx, postureChecks),
 	}
 

management/server/store/sql_store.go

@@ -6,6 +6,7 @@ import (
 	"encoding/json"
 	"errors"
 	"fmt"
+	"math"
 	"net"
 	"net/netip"
 	"net/url"
@@ -2258,125 +2259,30 @@ func (s *SqlStore) getPostureChecks(ctx context.Context, accountID string) ([]*p
 	return checks, nil
 }
 
-func (s *SqlStore) getServices(ctx context.Context, accountID string) ([]*rpservice.Service, error) {
-	const serviceQuery = `SELECT id, account_id, name, domain, enabled, auth, restrictions,
-		meta_created_at, meta_certificate_issued_at, meta_status, proxy_cluster,
-		pass_host_header, rewrite_redirects, session_private_key, session_public_key,
-		mode, listen_port, port_auto_assigned, source, source_peer, terminated,
-		private, access_groups
-		FROM services WHERE account_id = $1`
+// serviceSelectColumns and targetSelectColumns are the column lists the Postgres
+// pgx read path scans. They must stay in sync with the rpservice.Service and
+// rpservice.Target gorm models; TestPgxServiceColumnsMatchGorm enforces this.
+const serviceSelectColumns = `id, account_id, name, domain, enabled, auth, restrictions,
+	meta_created_at, meta_certificate_issued_at, meta_last_renewed_at, meta_status, proxy_cluster,
+	pass_host_header, rewrite_redirects, session_private_key, session_public_key,
+	mode, listen_port, port_auto_assigned, source, source_peer, terminated,
+	private, access_groups`
 
-	const targetsQuery = `SELECT id, account_id, service_id, path, host, port, protocol,
-		target_id, target_type, enabled
-		FROM targets WHERE service_id = ANY($1)`
+const targetSelectColumns = `id, account_id, service_id, path, host, port, protocol,
+	target_id, target_type, enabled, proxy_protocol,
+	skip_tls_verify, request_timeout, session_idle_timeout, path_rewrite, custom_headers,
+	direct_upstream, middlewares, capture_max_request_bytes, capture_max_response_bytes,
+	capture_content_types, agent_network, disable_access_log`
+
+func (s *SqlStore) getServices(ctx context.Context, accountID string) ([]*rpservice.Service, error) {
+	const serviceQuery = `SELECT ` + serviceSelectColumns + ` FROM services WHERE account_id = $1`
 
 	serviceRows, err := s.pool.Query(ctx, serviceQuery, accountID)
 	if err != nil {
 		return nil, err
 	}
 
-	services, err := pgx.CollectRows(serviceRows, func(row pgx.CollectableRow) (*rpservice.Service, error) {
-		var s rpservice.Service
-		var auth []byte
-		var restrictions []byte
-		var accessGroups []byte
-		var createdAt, certIssuedAt sql.NullTime
-		var status, proxyCluster, sessionPrivateKey, sessionPublicKey sql.NullString
-		var mode, source, sourcePeer sql.NullString
-		var terminated, portAutoAssigned, private sql.NullBool
-		var listenPort sql.NullInt64
-		err := row.Scan(
-			&s.ID,
-			&s.AccountID,
-			&s.Name,
-			&s.Domain,
-			&s.Enabled,
-			&auth,
-			&restrictions,
-			&createdAt,
-			&certIssuedAt,
-			&status,
-			&proxyCluster,
-			&s.PassHostHeader,
-			&s.RewriteRedirects,
-			&sessionPrivateKey,
-			&sessionPublicKey,
-			&mode,
-			&listenPort,
-			&portAutoAssigned,
-			&source,
-			&sourcePeer,
-			&terminated,
-			&private,
-			&accessGroups,
-		)
-		if err != nil {
-			return nil, err
-		}
-
-		if auth != nil {
-			if err := json.Unmarshal(auth, &s.Auth); err != nil {
-				return nil, err
-			}
-		}
-
-		if len(restrictions) > 0 {
-			if err := json.Unmarshal(restrictions, &s.Restrictions); err != nil {
-				return nil, fmt.Errorf("unmarshal restrictions: %w", err)
-			}
-		}
-
-		if len(accessGroups) > 0 {
-			if err := json.Unmarshal(accessGroups, &s.AccessGroups); err != nil {
-				return nil, fmt.Errorf("unmarshal access_groups: %w", err)
-			}
-		}
-
-		if private.Valid {
-			s.Private = private.Bool
-		}
-
-		s.Meta = rpservice.Meta{}
-		if createdAt.Valid {
-			s.Meta.CreatedAt = createdAt.Time
-		}
-		if certIssuedAt.Valid {
-			t := certIssuedAt.Time
-			s.Meta.CertificateIssuedAt = &t
-		}
-		if status.Valid {
-			s.Meta.Status = status.String
-		}
-		if proxyCluster.Valid {
-			s.ProxyCluster = proxyCluster.String
-		}
-		if sessionPrivateKey.Valid {
-			s.SessionPrivateKey = sessionPrivateKey.String
-		}
-		if sessionPublicKey.Valid {
-			s.SessionPublicKey = sessionPublicKey.String
-		}
-		if mode.Valid {
-			s.Mode = mode.String
-		}
-		if source.Valid {
-			s.Source = source.String
-		}
-		if sourcePeer.Valid {
-			s.SourcePeer = sourcePeer.String
-		}
-		if terminated.Valid {
-			s.Terminated = terminated.Bool
-		}
-		if portAutoAssigned.Valid {
-			s.PortAutoAssigned = portAutoAssigned.Bool
-		}
-		if listenPort.Valid {
-			s.ListenPort = uint16(listenPort.Int64)
-		}
-		s.Targets = []*rpservice.Target{}
-		return &s, nil
-	})
+	services, err := pgx.CollectRows(serviceRows, scanService)
 	if err != nil {
 		return nil, err
 	}
@@ -2387,39 +2293,12 @@ func (s *SqlStore) getServices(ctx context.Context, accountID string) ([]*rpserv
 
 	serviceIDs := make([]string, len(services))
 	serviceMap := make(map[string]*rpservice.Service)
-	for i, s := range services {
-		serviceIDs[i] = s.ID
-		serviceMap[s.ID] = s
+	for i, svc := range services {
+		serviceIDs[i] = svc.ID
+		serviceMap[svc.ID] = svc
 	}
 
-	targetRows, err := s.pool.Query(ctx, targetsQuery, serviceIDs)
-	if err != nil {
-		return nil, err
-	}
-
-	targets, err := pgx.CollectRows(targetRows, func(row pgx.CollectableRow) (*rpservice.Target, error) {
-		var t rpservice.Target
-		var path sql.NullString
-		err := row.Scan(
-			&t.ID,
-			&t.AccountID,
-			&t.ServiceID,
-			&path,
-			&t.Host,
-			&t.Port,
-			&t.Protocol,
-			&t.TargetId,
-			&t.TargetType,
-			&t.Enabled,
-		)
-		if err != nil {
-			return nil, err
-		}
-		if path.Valid {
-			t.Path = &path.String
-		}
-		return &t, nil
-	})
+	targets, err := s.getServiceTargets(ctx, serviceIDs)
 	if err != nil {
 		return nil, err
 	}
@@ -2433,6 +2312,201 @@ func (s *SqlStore) getServices(ctx context.Context, accountID string) ([]*rpserv
 	return services, nil
 }
 
+func scanService(row pgx.CollectableRow) (*rpservice.Service, error) {
+	var s rpservice.Service
+	var auth []byte
+	var restrictions []byte
+	var accessGroups []byte
+	var createdAt, certIssuedAt, lastRenewedAt sql.NullTime
+	var status, proxyCluster, sessionPrivateKey, sessionPublicKey sql.NullString
+	var mode, source, sourcePeer sql.NullString
+	var terminated, portAutoAssigned, private sql.NullBool
+	var listenPort sql.NullInt64
+	err := row.Scan(
+		&s.ID,
+		&s.AccountID,
+		&s.Name,
+		&s.Domain,
+		&s.Enabled,
+		&auth,
+		&restrictions,
+		&createdAt,
+		&certIssuedAt,
+		&lastRenewedAt,
+		&status,
+		&proxyCluster,
+		&s.PassHostHeader,
+		&s.RewriteRedirects,
+		&sessionPrivateKey,
+		&sessionPublicKey,
+		&mode,
+		&listenPort,
+		&portAutoAssigned,
+		&source,
+		&sourcePeer,
+		&terminated,
+		&private,
+		&accessGroups,
+	)
+	if err != nil {
+		return nil, err
+	}
+
+	if auth != nil {
+		if err := json.Unmarshal(auth, &s.Auth); err != nil {
+			return nil, err
+		}
+	}
+
+	if len(restrictions) > 0 {
+		if err := json.Unmarshal(restrictions, &s.Restrictions); err != nil {
+			return nil, fmt.Errorf("unmarshal restrictions: %w", err)
+		}
+	}
+
+	if len(accessGroups) > 0 {
+		if err := json.Unmarshal(accessGroups, &s.AccessGroups); err != nil {
+			return nil, fmt.Errorf("unmarshal access_groups: %w", err)
+		}
+	}
+
+	if private.Valid {
+		s.Private = private.Bool
+	}
+
+	s.Meta = serviceMetaFromRow(createdAt, certIssuedAt, lastRenewedAt, status)
+	if proxyCluster.Valid {
+		s.ProxyCluster = proxyCluster.String
+	}
+	if sessionPrivateKey.Valid {
+		s.SessionPrivateKey = sessionPrivateKey.String
+	}
+	if sessionPublicKey.Valid {
+		s.SessionPublicKey = sessionPublicKey.String
+	}
+	if mode.Valid {
+		s.Mode = mode.String
+	}
+	if source.Valid {
+		s.Source = source.String
+	}
+	if sourcePeer.Valid {
+		s.SourcePeer = sourcePeer.String
+	}
+	if terminated.Valid {
+		s.Terminated = terminated.Bool
+	}
+	if portAutoAssigned.Valid {
+		s.PortAutoAssigned = portAutoAssigned.Bool
+	}
+	if listenPort.Valid {
+		if listenPort.Int64 < 0 || listenPort.Int64 > math.MaxUint16 {
+			return nil, fmt.Errorf("listen_port %d out of range", listenPort.Int64)
+		}
+		s.ListenPort = uint16(listenPort.Int64)
+	}
+	s.Targets = []*rpservice.Target{}
+	return &s, nil
+}
+
+func serviceMetaFromRow(createdAt, certIssuedAt, lastRenewedAt sql.NullTime, status sql.NullString) rpservice.Meta {
+	meta := rpservice.Meta{}
+	if createdAt.Valid {
+		meta.CreatedAt = createdAt.Time
+	}
+	if certIssuedAt.Valid {
+		t := certIssuedAt.Time
+		meta.CertificateIssuedAt = &t
+	}
+	if lastRenewedAt.Valid {
+		t := lastRenewedAt.Time
+		meta.LastRenewedAt = &t
+	}
+	if status.Valid {
+		meta.Status = status.String
+	}
+	return meta
+}
+
+func (s *SqlStore) getServiceTargets(ctx context.Context, serviceIDs []string) ([]*rpservice.Target, error) {
+	const targetsQuery = `SELECT ` + targetSelectColumns + ` FROM targets WHERE service_id = ANY($1)`
+
+	rows, err := s.pool.Query(ctx, targetsQuery, serviceIDs)
+	if err != nil {
+		return nil, err
+	}
+
+	return pgx.CollectRows(rows, scanTarget)
+}
+
+func scanTarget(row pgx.CollectableRow) (*rpservice.Target, error) {
+	var t rpservice.Target
+	var path sql.NullString
+	var pathRewrite sql.NullString
+	var proxyProtocol, skipTLSVerify, directUpstream, agentNetwork, disableAccessLog sql.NullBool
+	var requestTimeout, sessionIdleTimeout, captureMaxRequestBytes, captureMaxResponseBytes sql.NullInt64
+	var customHeaders, middlewares, captureContentTypes []byte
+	err := row.Scan(
+		&t.ID,
+		&t.AccountID,
+		&t.ServiceID,
+		&path,
+		&t.Host,
+		&t.Port,
+		&t.Protocol,
+		&t.TargetId,
+		&t.TargetType,
+		&t.Enabled,
+		&proxyProtocol,
+		&skipTLSVerify,
+		&requestTimeout,
+		&sessionIdleTimeout,
+		&pathRewrite,
+		&customHeaders,
+		&directUpstream,
+		&middlewares,
+		&captureMaxRequestBytes,
+		&captureMaxResponseBytes,
+		&captureContentTypes,
+		&agentNetwork,
+		&disableAccessLog,
+	)
+	if err != nil {
+		return nil, err
+	}
+	if path.Valid {
+		t.Path = &path.String
+	}
+
+	t.ProxyProtocol = proxyProtocol.Bool
+	t.Options.SkipTLSVerify = skipTLSVerify.Bool
+	t.Options.RequestTimeout = time.Duration(requestTimeout.Int64)
+	t.Options.SessionIdleTimeout = time.Duration(sessionIdleTimeout.Int64)
+	t.Options.PathRewrite = rpservice.PathRewriteMode(pathRewrite.String)
+	t.Options.DirectUpstream = directUpstream.Bool
+	t.Options.CaptureMaxRequestBytes = captureMaxRequestBytes.Int64
+	t.Options.CaptureMaxResponseBytes = captureMaxResponseBytes.Int64
+	t.Options.AgentNetwork = agentNetwork.Bool
+	t.Options.DisableAccessLog = disableAccessLog.Bool
+
+	if len(customHeaders) > 0 {
+		if err := json.Unmarshal(customHeaders, &t.Options.CustomHeaders); err != nil {
+			return nil, fmt.Errorf("unmarshal custom_headers: %w", err)
+		}
+	}
+	if len(middlewares) > 0 {
+		if err := json.Unmarshal(middlewares, &t.Options.Middlewares); err != nil {
+			return nil, fmt.Errorf("unmarshal middlewares: %w", err)
+		}
+	}
+	if len(captureContentTypes) > 0 {
+		if err := json.Unmarshal(captureContentTypes, &t.Options.CaptureContentTypes); err != nil {
+			return nil, fmt.Errorf("unmarshal capture_content_types: %w", err)
+		}
+	}
+	return &t, nil
+}
+
 func (s *SqlStore) getNetworks(ctx context.Context, accountID string) ([]*networkTypes.Network, error) {
 	const query = `SELECT id, account_id, public_id, name, description FROM networks WHERE account_id = $1`
 	rows, err := s.pool.Query(ctx, query, accountID)
@@ -6365,7 +6439,6 @@ var validCapabilityColumns = map[string]struct{}{
 	"supports_custom_ports": {},
 	"require_subdomain":     {},
 	"supports_crowdsec":     {},
-	"supports_appsec":       {},
 	"private":               {},
 }
 
@@ -6396,14 +6469,6 @@ func (s *SqlStore) GetClusterSupportsCrowdSec(ctx context.Context, clusterAddr s
 	return s.getClusterUnanimousCapability(ctx, clusterAddr, "supports_crowdsec")
 }
 
-// GetClusterSupportsAppSec returns whether all active proxies in the cluster
-// have a CrowdSec AppSec endpoint configured. Returns nil when no proxy
-// reported the capability. Unanimous for the same reason as CrowdSec: a single
-// proxy without AppSec would let requests through uninspected.
-func (s *SqlStore) GetClusterSupportsAppSec(ctx context.Context, clusterAddr string) *bool {
-	return s.getClusterUnanimousCapability(ctx, clusterAddr, "supports_appsec")
-}
-
 // getClusterUnanimousCapability returns an aggregated boolean capability
 // requiring all active proxies in the cluster to report true.
 func (s *SqlStore) getClusterUnanimousCapability(ctx context.Context, clusterAddr, column string) *bool {

management/server/store/sql_store_pgx_parity_test.go

@@ -0,0 +1,74 @@
+package store
+
+import (
+	"strings"
+	"sync"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+	"gorm.io/gorm/schema"
+
+	rpservice "github.com/netbirdio/netbird/management/internals/modules/reverseproxy/service"
+)
+
+// TestPgxServiceColumnsMatchGorm guards the Postgres pgx read path against
+// drifting from the gorm model. The SQLite/MySQL gorm path loads rows by struct,
+// so a new column on a model is picked up automatically, but the hand-written
+// pgx SELECT in sql_store.go must be updated by hand. This test fails when a
+// gorm column is missing from the pgx column list, which otherwise silently
+// returns zero-valued on Postgres with no compile error.
+func TestPgxServiceColumnsMatchGorm(t *testing.T) {
+	tests := []struct {
+		name          string
+		model         any
+		selectColumns string
+		// excluded lists gorm columns intentionally not loaded by the pgx path.
+		excluded map[string]struct{}
+	}{
+		{
+			name:          "service",
+			model:         &rpservice.Service{},
+			selectColumns: serviceSelectColumns,
+		},
+		{
+			name:          "target",
+			model:         &rpservice.Target{},
+			selectColumns: targetSelectColumns,
+		},
+	}
+
+	for _, tc := range tests {
+		t.Run(tc.name, func(t *testing.T) {
+			selected := parseColumnList(tc.selectColumns)
+			for _, col := range gormColumnNames(t, tc.model) {
+				if _, ok := tc.excluded[col]; ok {
+					continue
+				}
+				_, ok := selected[col]
+				assert.Truef(t, ok,
+					"gorm column %q is not read by the Postgres pgx SELECT; add it to %sSelectColumns in sql_store.go (or to the test's excluded set if it is intentionally not loaded)",
+					col, tc.name)
+			}
+		})
+	}
+}
+
+func parseColumnList(cols string) map[string]struct{} {
+	set := make(map[string]struct{})
+	for _, c := range strings.Split(cols, ",") {
+		if c = strings.TrimSpace(c); c != "" {
+			set[c] = struct{}{}
+		}
+	}
+	return set
+}
+
+// gormColumnNames returns the DB column names gorm would migrate for the model,
+// using the same default naming strategy the store configures.
+func gormColumnNames(t *testing.T, model any) []string {
+	t.Helper()
+	sch, err := schema.Parse(model, &sync.Map{}, schema.NamingStrategy{})
+	require.NoError(t, err)
+	return sch.DBNames
+}

management/server/store/sql_store_proxy_capability_test.go

@@ -1,119 +0,0 @@
-package store
-
-import (
-	"context"
-	"fmt"
-	"os"
-	"runtime"
-	"testing"
-	"time"
-
-	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
-
-	"github.com/netbirdio/netbird/management/internals/modules/reverseproxy/proxy"
-)
-
-// Capabilities travel proxy → gRPC → embedded gorm columns → aggregation → API.
-// A field dropped at any of those hops reads as "capability absent", which is
-// indistinguishable from a proxy that never reported it: the dashboard simply
-// hides the feature and nothing fails. These assertions cover the persistence
-// and aggregation hops.
-func TestSqlStore_ClusterCapabilityAggregation(t *testing.T) {
-	if os.Getenv("CI") == "true" && (runtime.GOOS == "darwin" || runtime.GOOS == "windows") {
-		t.Skip("skip CI tests on darwin and windows")
-	}
-
-	yes, no := true, false
-
-	tests := []struct {
-		name          string
-		reported      []*bool // one entry per connected proxy in the cluster
-		wantAppSec    *bool
-		wantAssertion string
-	}{
-		{
-			name:          "unreported stays unknown",
-			reported:      []*bool{nil},
-			wantAppSec:    nil,
-			wantAssertion: "an unreported capability must not read as false",
-		},
-		{
-			name:          "single proxy reporting true",
-			reported:      []*bool{&yes},
-			wantAppSec:    &yes,
-			wantAssertion: "a reported capability must survive persistence",
-		},
-		{
-			name:          "one proxy without it disables the cluster",
-			reported:      []*bool{&yes, &no},
-			wantAppSec:    &no,
-			wantAssertion: "capability must be unanimous, so a rolling upgrade cannot leave traffic uninspected",
-		},
-		{
-			name:          "one proxy yet to report disables the cluster",
-			reported:      []*bool{&yes, nil},
-			wantAppSec:    &no,
-			wantAssertion: "a proxy that has not reported must not count as capable",
-		},
-	}
-
-	runTestForAllEngines(t, "", func(t *testing.T, store Store) {
-		ctx := context.Background()
-		for i, tt := range tests {
-			t.Run(tt.name, func(t *testing.T) {
-				cluster := fmt.Sprintf("cluster-%d.proxy.example", i)
-				for j, reported := range tt.reported {
-					require.NoError(t, store.SaveProxy(ctx, &proxy.Proxy{
-						ID:             fmt.Sprintf("proxy-%d-%d", i, j),
-						ClusterAddress: cluster,
-						Status:         proxy.StatusConnected,
-						LastSeen:       time.Now(),
-						Capabilities:   proxy.Capabilities{SupportsAppsec: reported},
-					}))
-				}
-
-				got := store.GetClusterSupportsAppSec(ctx, cluster)
-				if tt.wantAppSec == nil {
-					assert.Nil(t, got, tt.wantAssertion)
-					return
-				}
-				require.NotNil(t, got, tt.wantAssertion)
-				assert.Equal(t, *tt.wantAppSec, *got, tt.wantAssertion)
-			})
-		}
-	})
-}
-
-// AppSec and IP reputation are separate endpoints, so a cluster can have either
-// without the other. Gating one on the other would silently disable a feature
-// the operator configured.
-func TestSqlStore_ClusterCapabilitiesAreIndependent(t *testing.T) {
-	if os.Getenv("CI") == "true" && (runtime.GOOS == "darwin" || runtime.GOOS == "windows") {
-		t.Skip("skip CI tests on darwin and windows")
-	}
-
-	runTestForAllEngines(t, "", func(t *testing.T, store Store) {
-		ctx := context.Background()
-		const cluster = "independent.proxy.example"
-		yes, no := true, false
-
-		require.NoError(t, store.SaveProxy(ctx, &proxy.Proxy{
-			ID:             "proxy-independent",
-			ClusterAddress: cluster,
-			Status:         proxy.StatusConnected,
-			LastSeen:       time.Now(),
-			Capabilities: proxy.Capabilities{
-				SupportsAppsec:   &yes,
-				SupportsCrowdsec: &no,
-			},
-		}))
-
-		appsec := store.GetClusterSupportsAppSec(ctx, cluster)
-		crowdsec := store.GetClusterSupportsCrowdSec(ctx, cluster)
-		require.NotNil(t, appsec)
-		require.NotNil(t, crowdsec)
-		assert.True(t, *appsec, "AppSec must not be gated on CrowdSec")
-		assert.False(t, *crowdsec, "CrowdSec must not be implied by AppSec")
-	})
-}

management/server/store/sql_store_service_test.go

@@ -5,6 +5,7 @@ import (
 	"os"
 	"runtime"
 	"testing"
+	"time"
 
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
@@ -45,30 +46,62 @@ func TestSqlStore_GetAccount_PrivateServiceRoundtrip(t *testing.T) {
 	})
 }
 
-// Restrictions are stored as a JSON blob, and the Postgres read path lists
-// columns by hand: a mode that is not read there is silently off on Postgres
-// while working in SQLite dev.
-func TestSqlStore_GetAccount_ServiceRestrictionsRoundtrip(t *testing.T) {
+// TestSqlStore_GetAccount_ServiceTargetOptionsRoundtrip guards the Postgres pgx
+// read path (getServices) against silently dropping columns present on the gorm
+// model. Before the fix these fields loaded correctly on SQLite but came back
+// zero-valued on Postgres because the hand-written SELECT and scan omitted them.
+func TestSqlStore_GetAccount_ServiceTargetOptionsRoundtrip(t *testing.T) {
 	if os.Getenv("CI") == "true" && (runtime.GOOS == "darwin" || runtime.GOOS == "windows") {
 		t.Skip("skip CI tests on darwin and windows")
 	}
 
 	runTestForAllEngines(t, "", func(t *testing.T, store Store) {
 		ctx := context.Background()
-		account := newAccountWithId(ctx, "account_svc_restrictions", "testuser", "")
+		account := newAccountWithId(ctx, "account_svc_opts", "testuser", "")
 		require.NoError(t, store.SaveAccount(ctx, account))
 
+		renewedAt := time.Now().UTC().Truncate(time.Second)
+		targetPath := "/api"
 		svc := &rpservice.Service{
-			ID:        "svc-restrictions",
+			ID:        "svc-opts",
 			AccountID: account.Id,
-			Name:      "restricted-svc",
-			Domain:    "restricted.example",
+			Name:      "opts-svc",
+			Domain:    "opts.example",
 			Enabled:   true,
 			Mode:      rpservice.ModeHTTP,
 			Restrictions: rpservice.AccessRestrictions{
-				AllowedCIDRs: []string{"203.0.113.0/24"},
-				CrowdSecMode: "observe",
-				AppSecMode:   "enforce",
+				AllowedCIDRs:     []string{"10.0.0.0/8"},
+				BlockedCountries: []string{"XX"},
+				CrowdSecMode:     "block",
+			},
+			Meta: rpservice.Meta{
+				LastRenewedAt: &renewedAt,
+			},
+			Targets: []*rpservice.Target{
+				{
+					AccountID:     account.Id,
+					ServiceID:     "svc-opts",
+					Path:          &targetPath,
+					Host:          "backend.internal",
+					Port:          8080,
+					Protocol:      "http",
+					TargetId:      "tgt-1",
+					Enabled:       true,
+					ProxyProtocol: true,
+					Options: rpservice.TargetOptions{
+						SkipTLSVerify:           true,
+						RequestTimeout:          30 * time.Second,
+						SessionIdleTimeout:      5 * time.Minute,
+						PathRewrite:             rpservice.PathRewritePreserve,
+						CustomHeaders:           map[string]string{"X-Foo": "bar"},
+						DirectUpstream:          true,
+						CaptureMaxRequestBytes:  1024,
+						CaptureMaxResponseBytes: 2048,
+						CaptureContentTypes:     []string{"application/json"},
+						AgentNetwork:            true,
+						DisableAccessLog:        true,
+					},
+				},
 			},
 		}
 		require.NoError(t, store.CreateService(ctx, svc))
@@ -77,9 +110,26 @@ func TestSqlStore_GetAccount_ServiceRestrictionsRoundtrip(t *testing.T) {
 		require.NoError(t, err)
 		require.Len(t, loaded.Services, 1)
 
-		got := loaded.Services[0].Restrictions
-		assert.Equal(t, []string{"203.0.113.0/24"}, got.AllowedCIDRs)
-		assert.Equal(t, "observe", got.CrowdSecMode)
-		assert.Equal(t, "enforce", got.AppSecMode)
+		got := loaded.Services[0]
+		assert.Equal(t, []string{"10.0.0.0/8"}, got.Restrictions.AllowedCIDRs, "restrictions allowed CIDRs")
+		assert.Equal(t, []string{"XX"}, got.Restrictions.BlockedCountries, "restrictions blocked countries")
+		assert.Equal(t, "block", got.Restrictions.CrowdSecMode, "restrictions crowdsec mode")
+		require.NotNil(t, got.Meta.LastRenewedAt, "meta last renewed at")
+		assert.WithinDuration(t, renewedAt, *got.Meta.LastRenewedAt, time.Second, "meta last renewed at")
+
+		require.Len(t, got.Targets, 1)
+		tg := got.Targets[0]
+		assert.True(t, tg.ProxyProtocol, "target proxy protocol")
+		assert.True(t, tg.Options.SkipTLSVerify, "options skip TLS verify")
+		assert.Equal(t, 30*time.Second, tg.Options.RequestTimeout, "options request timeout")
+		assert.Equal(t, 5*time.Minute, tg.Options.SessionIdleTimeout, "options session idle timeout")
+		assert.Equal(t, rpservice.PathRewritePreserve, tg.Options.PathRewrite, "options path rewrite")
+		assert.Equal(t, map[string]string{"X-Foo": "bar"}, tg.Options.CustomHeaders, "options custom headers")
+		assert.True(t, tg.Options.DirectUpstream, "options direct upstream")
+		assert.Equal(t, int64(1024), tg.Options.CaptureMaxRequestBytes, "options capture max request bytes")
+		assert.Equal(t, int64(2048), tg.Options.CaptureMaxResponseBytes, "options capture max response bytes")
+		assert.Equal(t, []string{"application/json"}, tg.Options.CaptureContentTypes, "options capture content types")
+		assert.True(t, tg.Options.AgentNetwork, "options agent network")
+		assert.True(t, tg.Options.DisableAccessLog, "options disable access log")
 	})
 }

management/server/store/store.go

@@ -321,7 +321,6 @@ type Store interface {
 	GetClusterSupportsCustomPorts(ctx context.Context, clusterAddr string) *bool
 	GetClusterRequireSubdomain(ctx context.Context, clusterAddr string) *bool
 	GetClusterSupportsCrowdSec(ctx context.Context, clusterAddr string) *bool
-	GetClusterSupportsAppSec(ctx context.Context, clusterAddr string) *bool
 	GetClusterSupportsPrivate(ctx context.Context, clusterAddr string) *bool
 	CleanupStaleProxies(ctx context.Context, inactivityDuration time.Duration) error
 	GetAllProxies(ctx context.Context) ([]*proxy.Proxy, error)

management/server/store/store_mock.go

@@ -1835,20 +1835,6 @@ func (mr *MockStoreMockRecorder) GetClusterRequireSubdomain(ctx, clusterAddr int
 	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetClusterRequireSubdomain", reflect.TypeOf((*MockStore)(nil).GetClusterRequireSubdomain), ctx, clusterAddr)
 }
 
-// GetClusterSupportsAppSec mocks base method.
-func (m *MockStore) GetClusterSupportsAppSec(ctx context.Context, clusterAddr string) *bool {
-	m.ctrl.T.Helper()
-	ret := m.ctrl.Call(m, "GetClusterSupportsAppSec", ctx, clusterAddr)
-	ret0, _ := ret[0].(*bool)
-	return ret0
-}
-
-// GetClusterSupportsAppSec indicates an expected call of GetClusterSupportsAppSec.
-func (mr *MockStoreMockRecorder) GetClusterSupportsAppSec(ctx, clusterAddr interface{}) *gomock.Call {
-	mr.mock.ctrl.T.Helper()
-	return mr.mock.ctrl.RecordCallWithMethodType(mr.mock, "GetClusterSupportsAppSec", reflect.TypeOf((*MockStore)(nil).GetClusterSupportsAppSec), ctx, clusterAddr)
-}
-
 // GetClusterSupportsCrowdSec mocks base method.
 func (m *MockStore) GetClusterSupportsCrowdSec(ctx context.Context, clusterAddr string) *bool {
 	m.ctrl.T.Helper()

management/server/types/account.go

@@ -1517,6 +1517,54 @@ func (a *Account) GetResourceRoutersMap() map[string]map[string]*routerTypes.Net
 	return routers
 }
 
+// forcesRoutingPeerDNSResolution reports whether the given peer must run
+// routing-peer DNS resolution regardless of the account-global
+// RoutingPeerDNSResolutionEnabled setting. It returns true when the peer is a
+// router for a domain network resource that is targeted by an enabled
+// reverse-proxy service, so the peer's DNS forwarder starts and can resolve
+// the target for the embedded proxy peers. Embedded proxy peers themselves are
+// handled at PeerConfig build time.
+func (a *Account) forcesRoutingPeerDNSResolution(peerID string, routers map[string]map[string]*routerTypes.NetworkRouter) bool {
+	targeted := a.proxyTargetedDomainResourceIDs()
+	if len(targeted) == 0 {
+		return false
+	}
+
+	for _, resource := range a.NetworkResources {
+		if resource == nil || !resource.Enabled || resource.Type != resourceTypes.Domain {
+			continue
+		}
+		if _, ok := targeted[resource.ID]; !ok {
+			continue
+		}
+		if _, isRouter := routers[resource.NetworkID][peerID]; isRouter {
+			return true
+		}
+	}
+
+	return false
+}
+
+// proxyTargetedDomainResourceIDs returns the set of domain network resource IDs
+// targeted by an enabled, non-terminated reverse-proxy service.
+func (a *Account) proxyTargetedDomainResourceIDs() map[string]struct{} {
+	ids := make(map[string]struct{})
+	for _, svc := range a.Services {
+		if svc == nil || !svc.Enabled || svc.Terminated {
+			continue
+		}
+		for _, target := range svc.Targets {
+			if target == nil || !target.Enabled {
+				continue
+			}
+			if target.TargetType == service.TargetTypeDomain {
+				ids[target.TargetId] = struct{}{}
+			}
+		}
+	}
+	return ids
+}
+
 // getPoliciesSourcePeers collects all unique peers from the source groups defined in the given policies.
 func getPoliciesSourcePeers(policies []*Policy, groups map[string]*Group) map[string]struct{} {
 	sourcePeers := make(map[string]struct{})

management/server/types/account_components.go

@@ -140,6 +140,8 @@ func (a *Account) GetPeerNetworkMapComponents(
 		RouterPeers:               make(map[string]*ComponentPeer),
 		NetworkXIDToPublicID:      make(map[string]string, len(a.Networks)),
 		PostureCheckXIDToPublicID: make(map[string]string, len(a.PostureChecks)),
+
+		ForceRoutingPeerDNSResolution: a.forcesRoutingPeerDNSResolution(peerID, routers),
 	}
 	for _, n := range a.Networks {
 		if n != nil {

management/server/types/account_test.go

@@ -1751,3 +1751,71 @@ func hasPrivateAccessPolicy(account *Account, serviceID string) bool {
 	}
 	return false
 }
+
+func TestForcesRoutingPeerDNSResolution(t *testing.T) {
+	buildAccountRes := func(serviceEnabled, targetEnabled, resourceEnabled bool, targetType service.TargetType, resType resourceTypes.NetworkResourceType) *Account {
+		return &Account{
+			Id: "accountID",
+			Groups: map[string]*Group{
+				"router-group": {ID: "router-group", Peers: []string{"router-peer-grp"}},
+			},
+			NetworkRouters: []*routerTypes.NetworkRouter{
+				{ID: "r1", NetworkID: "net-1", AccountID: "accountID", Peer: "router-peer", Enabled: true},
+				{ID: "r2", NetworkID: "net-1", AccountID: "accountID", PeerGroups: []string{"router-group"}, Enabled: true},
+			},
+			NetworkResources: []*resourceTypes.NetworkResource{
+				{ID: "res-domain", AccountID: "accountID", NetworkID: "net-1", Type: resType, Domain: "example.org", Enabled: resourceEnabled},
+			},
+			Services: []*service.Service{
+				{
+					ID: "svc-1", AccountID: "accountID", Enabled: serviceEnabled,
+					Targets: []*service.Target{
+						{TargetId: "res-domain", TargetType: targetType, Enabled: targetEnabled},
+					},
+				},
+			},
+		}
+	}
+
+	buildAccount := func(serviceEnabled, targetEnabled, resourceEnabled bool, targetType service.TargetType) *Account {
+		return buildAccountRes(serviceEnabled, targetEnabled, resourceEnabled, targetType, resourceTypes.Domain)
+	}
+
+	t.Run("router peer for RP-targeted domain resource is forced", func(t *testing.T) {
+		account := buildAccount(true, true, true, service.TargetTypeDomain)
+		routers := account.GetResourceRoutersMap()
+		assert.True(t, account.forcesRoutingPeerDNSResolution("router-peer", routers), "direct router peer should be forced")
+		assert.True(t, account.forcesRoutingPeerDNSResolution("router-peer-grp", routers), "group-member router peer should be forced")
+	})
+
+	t.Run("non-router peer is not forced", func(t *testing.T) {
+		account := buildAccount(true, true, true, service.TargetTypeDomain)
+		assert.False(t, account.forcesRoutingPeerDNSResolution("other-peer", account.GetResourceRoutersMap()))
+	})
+
+	t.Run("not forced when service disabled", func(t *testing.T) {
+		account := buildAccount(false, true, true, service.TargetTypeDomain)
+		assert.False(t, account.forcesRoutingPeerDNSResolution("router-peer", account.GetResourceRoutersMap()))
+	})
+
+	t.Run("not forced when target disabled", func(t *testing.T) {
+		account := buildAccount(true, false, true, service.TargetTypeDomain)
+		assert.False(t, account.forcesRoutingPeerDNSResolution("router-peer", account.GetResourceRoutersMap()))
+	})
+
+	t.Run("not forced when resource disabled", func(t *testing.T) {
+		account := buildAccount(true, true, false, service.TargetTypeDomain)
+		assert.False(t, account.forcesRoutingPeerDNSResolution("router-peer", account.GetResourceRoutersMap()))
+	})
+
+	t.Run("not forced for non-domain target type", func(t *testing.T) {
+		account := buildAccount(true, true, true, service.TargetTypePeer)
+		assert.False(t, account.forcesRoutingPeerDNSResolution("router-peer", account.GetResourceRoutersMap()))
+	})
+
+	t.Run("not forced when targeted resource is not a domain", func(t *testing.T) {
+		account := buildAccountRes(true, true, true, service.TargetTypeDomain, resourceTypes.Host)
+		assert.False(t, account.forcesRoutingPeerDNSResolution("router-peer", account.GetResourceRoutersMap()),
+			"a domain target pointing at a non-domain resource must not force resolution")
+	})
+}

proxy/cmd/proxy/cmd/root.go

@@ -79,11 +79,6 @@ var (
 	geoDataDir            string
 	crowdsecAPIURL        string
 	crowdsecAPIKey        string
-	appsecURL             string
-	appsecTimeout         time.Duration
-	appsecMaxBodyBytes    int64
-	captureBudgetBytes    int64
-	appsecMaxConcurrent   int
 )
 
 var rootCmd = &cobra.Command{
@@ -130,11 +125,6 @@ func init() {
 	rootCmd.Flags().StringVar(&geoDataDir, "geo-data-dir", envStringOrDefault("NB_PROXY_GEO_DATA_DIR", "/var/lib/netbird/geolocation"), "Directory for the GeoLite2 MMDB file (auto-downloaded if missing)")
 	rootCmd.Flags().StringVar(&crowdsecAPIURL, "crowdsec-api-url", envStringOrDefault("NB_PROXY_CROWDSEC_API_URL", ""), "CrowdSec LAPI URL for IP reputation checks")
 	rootCmd.Flags().StringVar(&crowdsecAPIKey, "crowdsec-api-key", envStringOrDefault("NB_PROXY_CROWDSEC_API_KEY", ""), "CrowdSec bouncer API key")
-	rootCmd.Flags().StringVar(&appsecURL, "crowdsec-appsec-url", envStringOrDefault("NB_PROXY_CROWDSEC_APPSEC_URL", ""), "CrowdSec AppSec (WAF) endpoint for HTTP request inspection, e.g. http://127.0.0.1:7422/ (reuses the bouncer API key)")
-	rootCmd.Flags().DurationVar(&appsecTimeout, "crowdsec-appsec-timeout", envDurationOrDefault("NB_PROXY_CROWDSEC_APPSEC_TIMEOUT", 0), "Timeout for a single AppSec inspection call (0 = 200ms)")
-	rootCmd.Flags().IntVar(&appsecMaxConcurrent, "crowdsec-appsec-max-concurrent", int(envInt64OrDefault("NB_PROXY_CROWDSEC_APPSEC_MAX_CONCURRENT", 0)), "Cap on AppSec inspections in flight; further requests are denied in enforce mode rather than queued (0 = 256, negative = no cap)")
-	rootCmd.Flags().Int64Var(&captureBudgetBytes, "capture-budget-bytes", envInt64OrDefault("NB_PROXY_CAPTURE_BUDGET_BYTES", 0), "Total in-flight request-body buffering across the proxy, shared by AppSec inspection and agent-network capture (0 = 256MiB)")
-	rootCmd.Flags().Int64Var(&appsecMaxBodyBytes, "crowdsec-appsec-max-body-bytes", envInt64OrDefault("NB_PROXY_CROWDSEC_APPSEC_MAX_BODY_BYTES", 0), "Cap on the request body mirrored to AppSec (0 = 64KiB, negative = headers and URI only)")
 }
 
 // Execute runs the root command.
@@ -228,59 +218,47 @@ func runServer(cmd *cobra.Command, args []string) error {
 	ctx, stop := signal.NotifyContext(context.Background(), syscall.SIGTERM, syscall.SIGINT)
 	defer stop()
 
-	srv := proxy.New(ctx, serverConfig(logger, proxyToken, parsedTrustedProxies, perf))
+	srv := proxy.New(ctx, proxy.Config{
+		ListenAddr:               addr,
+		Logger:                   logger,
+		Version:                  Version,
+		ManagementAddress:        mgmtAddr,
+		ProxyURL:                 proxyDomain,
+		ProxyToken:               proxyToken,
+		CertificateDirectory:     certDir,
+		CertificateFile:          certFile,
+		CertificateKeyFile:       certKeyFile,
+		GenerateACMECertificates: acmeCerts,
+		ACMEChallengeAddress:     acmeAddr,
+		ACMEDirectory:            acmeDir,
+		ACMEEABKID:               acmeEABKID,
+		ACMEEABHMACKey:           acmeEABHMACKey,
+		ACMEChallengeType:        acmeChallengeType,
+		DebugEndpointEnabled:     debugEndpoint,
+		DebugEndpointAddress:     debugEndpointAddr,
+		HealthAddr:               healthAddr,
+		ForwardedProto:           forwardedProto,
+		TrustedProxies:           parsedTrustedProxies,
+		CertLockMethod:           nbacme.CertLockMethod(certLockMethod),
+		WildcardCertDir:          wildcardCertDir,
+		WireguardPort:            wgPort,
+		Performance:              perf,
+		ProxyProtocol:            proxyProtocol,
+		PreSharedKey:             preSharedKey,
+		SupportsCustomPorts:      supportsCustomPorts,
+		RequireSubdomain:         requireSubdomain,
+		Private:                  private,
+		MaxDialTimeout:           maxDialTimeout,
+		MaxSessionIdleTimeout:    maxSessionIdleTimeout,
+		MappingBatchWatchdog:     envDurationOrDefault("NB_PROXY_MAPPING_BATCH_WATCHDOG", 0),
+		GeoDataDir:               geoDataDir,
+		CrowdSecAPIURL:           crowdsecAPIURL,
+		CrowdSecAPIKey:           crowdsecAPIKey,
+	})
 
 	return srv.ListenAndServe(ctx, addr)
 }
 
-// serverConfig maps the parsed flags and environment onto the proxy config.
-// Kept separate from runServer so registering a new flag does not grow the
-// startup path.
-func serverConfig(logger *log.Logger, proxyToken string, trustedProxyList *trustedproxy.List, perf embed.Performance) proxy.Config {
-	return proxy.Config{
-		ListenAddr:                   addr,
-		Logger:                       logger,
-		Version:                      Version,
-		ManagementAddress:            mgmtAddr,
-		ProxyURL:                     proxyDomain,
-		ProxyToken:                   proxyToken,
-		CertificateDirectory:         certDir,
-		CertificateFile:              certFile,
-		CertificateKeyFile:           certKeyFile,
-		GenerateACMECertificates:     acmeCerts,
-		ACMEChallengeAddress:         acmeAddr,
-		ACMEDirectory:                acmeDir,
-		ACMEEABKID:                   acmeEABKID,
-		ACMEEABHMACKey:               acmeEABHMACKey,
-		ACMEChallengeType:            acmeChallengeType,
-		DebugEndpointEnabled:         debugEndpoint,
-		DebugEndpointAddress:         debugEndpointAddr,
-		HealthAddr:                   healthAddr,
-		ForwardedProto:               forwardedProto,
-		TrustedProxies:               trustedProxyList,
-		CertLockMethod:               nbacme.CertLockMethod(certLockMethod),
-		WildcardCertDir:              wildcardCertDir,
-		WireguardPort:                wgPort,
-		Performance:                  perf,
-		ProxyProtocol:                proxyProtocol,
-		PreSharedKey:                 preSharedKey,
-		SupportsCustomPorts:          supportsCustomPorts,
-		RequireSubdomain:             requireSubdomain,
-		Private:                      private,
-		MaxDialTimeout:               maxDialTimeout,
-		MaxSessionIdleTimeout:        maxSessionIdleTimeout,
-		MappingBatchWatchdog:         envDurationOrDefault("NB_PROXY_MAPPING_BATCH_WATCHDOG", 0),
-		GeoDataDir:                   geoDataDir,
-		CrowdSecAPIURL:               crowdsecAPIURL,
-		CrowdSecAPIKey:               crowdsecAPIKey,
-		CrowdSecAppSecURL:            appsecURL,
-		CrowdSecAppSecTimeout:        appsecTimeout,
-		CrowdSecAppSecMaxBodyBytes:   appsecMaxBodyBytes,
-		CrowdSecAppSecMaxConcurrent:  appsecMaxConcurrent,
-		MiddlewareCaptureBudgetBytes: captureBudgetBytes,
-	}
-}
-
 func envBoolOrDefault(key string, def bool) bool {
 	v, exists := os.LookupEnv(key)
 	if !exists {
@@ -315,19 +293,6 @@ func envUint16OrDefault(key string, def uint16) uint16 {
 	return uint16(parsed)
 }
 
-func envInt64OrDefault(key string, def int64) int64 {
-	v, exists := os.LookupEnv(key)
-	if !exists {
-		return def
-	}
-	parsed, err := strconv.ParseInt(v, 10, 64)
-	if err != nil {
-		log.Warnf("parse %s=%q: %v, using default %d", key, v, err, def)
-		return def
-	}
-	return parsed
-}
-
 func envDurationOrDefault(key string, def time.Duration) time.Duration {
 	v, exists := os.LookupEnv(key)
 	if !exists {

proxy/internal/appsec/body.go

@@ -1,163 +0,0 @@
-package appsec
-
-import (
-	"bytes"
-	"errors"
-	"io"
-	"mime"
-	"net/http"
-	"net/url"
-	"slices"
-	"strings"
-)
-
-// bufferBody reads up to limit+1 bytes from r.Body and always restores r.Body so
-// the request stays forwardable. oversize reports that the body exceeded limit, in
-// which case the returned prefix must not be used for inspection: the bytes are
-// only read so they can be replayed to the backend.
-func bufferBody(r *http.Request, limit int64) (body []byte, oversize bool, err error) {
-	original := r.Body
-	buf, readErr := io.ReadAll(io.LimitReader(original, limit+1))
-	if readErr != nil && !errors.Is(readErr, io.EOF) {
-		// Restore what was read so a downstream retry sees a consistent stream,
-		// then surface the failure.
-		r.Body = replay(buf, original)
-		return nil, false, readErr
-	}
-
-	if int64(len(buf)) > limit {
-		r.Body = replay(buf, original)
-		return nil, true, nil
-	}
-
-	// The whole body is buffered, so the original is drained and can be closed.
-	// A close error on a drained read-only body does not invalidate the bytes.
-	_ = original.Close()
-	r.Body = io.NopCloser(bytes.NewReader(buf))
-	// Framing is deliberately left as the client sent it. Rewriting a chunked
-	// request to a fixed Content-Length here would be invisible to the client
-	// but not to the rest of the chain: a later body capture with a smaller cap
-	// sees a known length over its cap and skips capture entirely, where an
-	// unknown length would have given it a truncated prefix. Inspecting a
-	// request must not change what any other layer gets to inspect.
-	return buf, false, nil
-}
-
-// replay returns a ReadCloser that yields the already-read prefix followed by
-// the remainder of the original stream, and closes the original.
-func replay(prefix []byte, rest io.ReadCloser) io.ReadCloser {
-	return struct {
-		io.Reader
-		io.Closer
-	}{
-		Reader: io.MultiReader(bytes.NewReader(prefix), rest),
-		Closer: rest,
-	}
-}
-
-// redactedPlaceholder replaces a credential value in the mirrored body. It is
-// inert for rule matching, and its fixed length leaks nothing about the secret.
-const redactedPlaceholder = "redacted"
-
-// redactFormFields returns the body to mirror for a URL-encoded form, with the
-// values of the named fields replaced. The proxy's own password / PIN login
-// form posts to the service path itself, so without this the plaintext
-// credential would reach the Security Engine.
-//
-// Only the credential values are removed, never the whole body: dropping the
-// body outright would let a caller exempt any payload from inspection just by
-// appending a field named "password". Everything else in the form stays
-// inspectable, which is the point.
-//
-// Returns body unchanged when it is not a URL-encoded form or carries none of
-// the fields.
-//
-// Substitution happens on the raw bytes rather than by re-encoding parsed
-// values. Re-encoding would drop pairs that url.ParseQuery rejects, so a
-// payload hidden in a malformed pair alongside a credential-named field would
-// never be inspected while a tolerant backend parser still acted on it. Working
-// byte-wise also avoids reordering keys and normalizing escapes, so the engine
-// sees the same bytes the backend will.
-//
-// Field names match case-sensitively, on purpose: the caller passes the exact
-// names the login handler reads via r.FormValue, and that lookup is itself
-// case-sensitive. A "Password" field is therefore never a credential as far as
-// the proxy is concerned, and redacting it would only blind the WAF to a value
-// the proxy does not own.
-func redactFormFields(contentType string, body []byte, fields []string) []byte {
-	if len(fields) == 0 || len(body) == 0 {
-		return body
-	}
-	media, _, err := mime.ParseMediaType(contentType)
-	if err != nil || media != "application/x-www-form-urlencoded" {
-		return body
-	}
-	return redactURLEncoded(body, fields)
-}
-
-// redactURLEncoded replaces the values of the named keys in a URL-encoded
-// key/value sequence, the shared syntax of a query string and a form body.
-func redactURLEncoded(raw []byte, fields []string) []byte {
-	// Split on "&" only, matching how Go's form parser delimits pairs.
-	segments := bytes.Split(raw, []byte("&"))
-	redacted := false
-	for i, segment := range segments {
-		rawKey, _, hasValue := bytes.Cut(segment, []byte("="))
-		if !hasValue {
-			continue
-		}
-		// Compare the decoded name, so an escaped spelling of the field
-		// ("pass%77ord") is redacted too: the reader decodes before looking it
-		// up. A key that fails to decode never reaches that reader either,
-		// since the parser drops the pair.
-		name, err := url.QueryUnescape(string(rawKey))
-		if err != nil || !slices.Contains(fields, name) {
-			continue
-		}
-		// Keep the key bytes as sent and replace only the value. Assigning a
-		// fresh slice leaves raw untouched, which matters: the caller restored
-		// the request body from the same buffer.
-		segments[i] = []byte(string(rawKey) + "=" + redactedPlaceholder)
-		redacted = true
-	}
-	if !redacted {
-		return raw
-	}
-	return bytes.Join(segments, []byte("&"))
-}
-
-// redactQuery replaces the values of the named query parameters in a raw query
-// string, leaving every other byte as sent.
-func redactQuery(rawQuery string, params []string) string {
-	if len(params) == 0 || rawQuery == "" {
-		return rawQuery
-	}
-	return string(redactURLEncoded([]byte(rawQuery), params))
-}
-
-// redactCookieHeader replaces the values of the named cookies in a Cookie
-// header, keeping the others intact: cookies are a zone WAF rules match on, so
-// dropping the whole header would cost real coverage.
-func redactCookieHeader(value string, names []string) string {
-	if len(names) == 0 || value == "" {
-		return value
-	}
-	parts := strings.Split(value, ";")
-	redacted := false
-	for i, part := range parts {
-		name, _, hasValue := strings.Cut(part, "=")
-		if !hasValue {
-			continue
-		}
-		// Cookie names are case-sensitive and are not percent-decoded.
-		if !slices.Contains(names, strings.TrimSpace(name)) {
-			continue
-		}
-		parts[i] = name + "=" + redactedPlaceholder
-		redacted = true
-	}
-	if !redacted {
-		return value
-	}
-	return strings.Join(parts, ";")
-}

proxy/internal/appsec/client.go

@@ -1,571 +0,0 @@
-// Package appsec implements the CrowdSec AppSec (WAF) side of the remediation
-// component protocol: each inspected HTTP request is mirrored to the Security
-// Engine's AppSec endpoint, which replies with an allow / ban / captcha verdict
-// for that request.
-//
-// This is a separate endpoint from the LAPI decision stream used by the
-// crowdsec package: LAPI answers "is this IP known bad", AppSec answers "is
-// this request an attack". The two are configured and enabled independently.
-package appsec
-
-import (
-	"bytes"
-	"context"
-	"encoding/json"
-	"errors"
-	"fmt"
-	"io"
-	"net/http"
-	"net/netip"
-	"net/url"
-	"strings"
-	"sync"
-	"time"
-
-	log "github.com/sirupsen/logrus"
-
-	"github.com/netbirdio/netbird/proxy/internal/netutil"
-	"github.com/netbirdio/netbird/proxy/internal/restrict"
-)
-
-// Header names the AppSec component reads off the mirrored request. IP, URI and
-// Verb are mandatory: the engine answers 500 when any of them is missing.
-const (
-	headerAPIKey        = "X-Crowdsec-Appsec-Api-Key" //nolint:gosec // G101: a header name, not a credential
-	headerIP            = "X-Crowdsec-Appsec-Ip"
-	headerURI           = "X-Crowdsec-Appsec-Uri"
-	headerVerb          = "X-Crowdsec-Appsec-Verb"
-	headerHost          = "X-Crowdsec-Appsec-Host"
-	headerUserAgent     = "X-Crowdsec-Appsec-User-Agent"
-	headerHTTPVersion   = "X-Crowdsec-Appsec-Http-Version"
-	headerTransactionID = "X-Crowdsec-Appsec-Transaction-Id"
-)
-
-// headerPrefix covers every protocol header. Any client-supplied header in this
-// namespace is dropped before forwarding so a caller cannot influence the
-// engine's view of its own address, or replay an API key.
-const headerPrefix = "X-Crowdsec-Appsec-"
-
-// Remediation actions the engine can return.
-const (
-	actionAllow   = "allow"
-	actionBan     = "ban"
-	actionCaptcha = "captcha"
-)
-
-const (
-	// DefaultTimeout matches the 200ms budget CrowdSec's remediation component
-	// spec sets for the blocking AppSec call.
-	DefaultTimeout = 200 * time.Millisecond
-	// MinTimeout and MaxTimeout bound the configured inspection timeout.
-	// Inspection is synchronous, so the upper bound is what keeps a
-	// mis-set value from parking every request to an inspected service on a
-	// slow engine; the lower bound keeps the call from timing out before the
-	// engine can realistically answer. Mirrors the per-middleware bounds the
-	// proxy already applies to in-path calls.
-	MinTimeout = 10 * time.Millisecond
-	MaxTimeout = 5 * time.Second
-	// DefaultMaxBodyBytes caps the request body mirrored to the engine.
-	// Requests with a larger body are inspected on headers and URI only.
-	DefaultMaxBodyBytes int64 = 64 << 10
-	// DefaultMaxConcurrent bounds inspections in flight toward the engine. The
-	// point is to fail fast instead of parking a goroutine per request for the
-	// whole timeout once the engine is saturated: a slow engine otherwise turns
-	// a traffic burst into a pile of waiters that all time out anyway. Sized so
-	// a healthy engine (single-digit milliseconds per call) never reaches it.
-	DefaultMaxConcurrent = 256
-	// MaxConcurrentLimit is the ceiling for that bound.
-	MaxConcurrentLimit = 4096
-	// MaxBodyBytesLimit is the ceiling for that cap. A single request can hold
-	// this much in memory; the shared Budget is what bounds the total across
-	// concurrent requests. Matches the proxy-wide body-capture ceiling.
-	MaxBodyBytesLimit int64 = 8 << 20
-	// maxResponseBytes bounds how much of a verdict response is read. The
-	// engine answers with a two-field JSON object, so anything beyond this is
-	// not a response we can act on.
-	maxResponseBytes int64 = 4 << 10
-)
-
-// Reasons the request body was not mirrored. Reported so an access-log reader
-// can distinguish "inspected and clean" from "never inspected", and so an
-// oversize opt-out is visible rather than silent.
-const (
-	BypassOversize = "oversize"
-	BypassUpgrade  = "upgrade"
-	BypassDisabled = "disabled"
-	BypassBudget   = "budget_exhausted"
-)
-
-// ErrUnavailable reports that the engine could not produce a verdict: the call
-// failed, timed out, or the engine rejected it (401 bad key, 500 malformed).
-// Distinguished from a block verdict so the caller can apply the per-service
-// mode: enforce fails closed, observe allows.
-var ErrUnavailable = errors.New("appsec engine unavailable")
-
-// Config configures a Client.
-type Config struct {
-	// URL is the AppSec endpoint, e.g. http://127.0.0.1:7422/.
-	URL string
-	// APIKey is the CrowdSec bouncer API key. The AppSec component validates it
-	// against LAPI, so the same key used for the decision stream works here.
-	APIKey string
-	// Timeout bounds a single inspection call. Zero means DefaultTimeout.
-	Timeout time.Duration
-	// MaxBodyBytes caps the mirrored request body. Zero means
-	// DefaultMaxBodyBytes; negative disables body forwarding entirely.
-	MaxBodyBytes int64
-	// MaxConcurrent bounds inspections in flight toward the engine. Zero means
-	// DefaultMaxConcurrent; negative disables the bound.
-	MaxConcurrent int
-	// Budget bounds the total body buffering in flight across all inspected
-	// requests. Nil disables that ceiling, which leaves the worst case at
-	// MaxBodyBytes times the concurrent request count; callers serving
-	// untrusted traffic should share the proxy-wide capture budget here.
-	Budget Budget
-	Logger *log.Entry
-}
-
-// Budget is the shared allowance for in-flight body buffering. Acquire reports
-// whether n bytes could be reserved; every successful Acquire is matched by a
-// Release of the same n. Satisfied by the proxy's capture budget, so AppSec and
-// the middleware body tap draw down one pool rather than two independent ones.
-type Budget interface {
-	Acquire(n int64) bool
-	Release(n int64)
-}
-
-// Client mirrors HTTP requests to a CrowdSec AppSec endpoint. It holds no
-// per-service state and is safe for concurrent use.
-type Client struct {
-	url          string
-	apiKey       string
-	maxBodyBytes int64
-	// sem bounds in-flight inspections. Nil when the bound is disabled.
-	sem    chan struct{}
-	budget Budget
-	http   *http.Client
-	logger *log.Entry
-}
-
-// New validates the config and returns a Client. The endpoint is not contacted
-// here: the engine may come up after the proxy.
-func New(cfg Config) (*Client, error) {
-	if cfg.URL == "" {
-		return nil, errors.New("appsec url is empty")
-	}
-	if cfg.APIKey == "" {
-		return nil, errors.New("appsec api key is empty")
-	}
-	parsed, err := url.Parse(cfg.URL)
-	if err != nil {
-		return nil, fmt.Errorf("parse appsec url: %w", err)
-	}
-	if parsed.Scheme != "http" && parsed.Scheme != "https" {
-		return nil, fmt.Errorf("appsec url scheme %q is not http(s)", parsed.Scheme)
-	}
-	if parsed.Host == "" {
-		return nil, errors.New("appsec url has no host")
-	}
-
-	logger := cfg.Logger
-	if logger == nil {
-		logger = log.NewEntry(log.StandardLogger())
-	}
-
-	timeout := cfg.Timeout
-	switch {
-	case timeout <= 0:
-		timeout = DefaultTimeout
-	case timeout < MinTimeout:
-		logger.Warnf("appsec timeout %s is below the minimum, using %s", timeout, MinTimeout)
-		timeout = MinTimeout
-	case timeout > MaxTimeout:
-		logger.Warnf("appsec timeout %s exceeds the maximum, using %s", timeout, MaxTimeout)
-		timeout = MaxTimeout
-	}
-
-	// A negative cap is meaningful: forward no body at all.
-	maxBody := cfg.MaxBodyBytes
-	switch {
-	case maxBody == 0:
-		maxBody = DefaultMaxBodyBytes
-	case maxBody > MaxBodyBytesLimit:
-		logger.Warnf("appsec max body %d exceeds the maximum, using %d", maxBody, MaxBodyBytesLimit)
-		maxBody = MaxBodyBytesLimit
-	}
-
-	maxConcurrent := cfg.MaxConcurrent
-	switch {
-	case maxConcurrent == 0:
-		maxConcurrent = DefaultMaxConcurrent
-	case maxConcurrent > MaxConcurrentLimit:
-		logger.Warnf("appsec max concurrent %d exceeds the maximum, using %d", maxConcurrent, MaxConcurrentLimit)
-		maxConcurrent = MaxConcurrentLimit
-	}
-	var sem chan struct{}
-	if maxConcurrent > 0 {
-		sem = make(chan struct{}, maxConcurrent)
-	}
-
-	return &Client{
-		url:          cfg.URL,
-		apiKey:       cfg.APIKey,
-		maxBodyBytes: maxBody,
-		sem:          sem,
-		budget:       cfg.Budget,
-		logger:       logger,
-		http: &http.Client{
-			Timeout: timeout,
-			Transport: &http.Transport{
-				MaxIdleConns:        100,
-				MaxIdleConnsPerHost: 32,
-				IdleConnTimeout:     90 * time.Second,
-			},
-		},
-	}, nil
-}
-
-// Request is one inspection request.
-type Request struct {
-	// HTTP is the in-flight client request. Inspect buffers and restores its
-	// body, so the request stays forwardable afterwards.
-	HTTP *http.Request
-	// ClientIP is the resolved client address (after trusted-proxy handling).
-	ClientIP netip.Addr
-	// TransactionID correlates the engine's alert with the proxy's access log
-	// entry. Empty lets the engine generate its own UUID.
-	TransactionID string
-	// RedactBodyFields lists form fields whose values are replaced before the
-	// body is mirrored. Used to keep credentials submitted to the proxy's own
-	// login form out of the engine while still inspecting the rest.
-	RedactBodyFields []string
-	// RedactHeaders, RedactCookies and RedactQueryParams name the credentials
-	// the proxy already withholds from backends: the header-auth values, its
-	// session cookie, and the OIDC session token. The engine logs and alerts on
-	// what it inspects, so mirroring them there would reintroduce the leak the
-	// upstream strippers exist to prevent. Only the values are replaced, so the
-	// surrounding headers, cookies and query stay inspectable.
-	RedactHeaders     []string
-	RedactCookies     []string
-	RedactQueryParams []string
-}
-
-// Result is the outcome of an inspection.
-type Result struct {
-	Verdict restrict.Verdict
-	// BodyBypass names why the request body was not mirrored, empty when it
-	// was (or when the request had none). The engine still saw the headers and
-	// URI, so this is a coverage note, not a failure.
-	BodyBypass string
-	// Release returns the buffered body's budget reservation. Never nil, so it
-	// is always safe to defer. It must run only once the request has been
-	// served, not when Inspect returns: the buffer stays alive as r.Body for
-	// the backend to read, so releasing earlier would let the budget admit
-	// buffering that is still resident.
-	Release func()
-}
-
-// noopRelease is the Release for inspections that reserved no budget.
-func noopRelease() {}
-
-// Inspect mirrors r to the AppSec engine and returns its verdict. A nil error
-// with restrict.Allow means the request passed. On failure it returns
-// DenyAppSecUnavailable wrapped with ErrUnavailable; the caller decides whether
-// that blocks, based on the per-service mode.
-func (c *Client) Inspect(ctx context.Context, req Request) (Result, error) {
-	if c == nil {
-		return Result{Verdict: restrict.DenyAppSecUnavailable, Release: noopRelease}, ErrUnavailable
-	}
-	if req.HTTP == nil {
-		return Result{Verdict: restrict.DenyAppSecUnavailable, Release: noopRelease}, fmt.Errorf("%w: nil request", ErrUnavailable)
-	}
-
-	// release is carried out to the caller rather than deferred here: the
-	// buffered body outlives this call as r.Body.
-	if !c.acquireSlot() {
-		// Deny rather than wave through: a flood must not be a way to switch
-		// inspection off. Enforce blocks, observe logs and allows, exactly as
-		// for an unreachable engine.
-		return Result{Verdict: restrict.DenyAppSecUnavailable, Release: noopRelease},
-			fmt.Errorf("%w: %d inspections already in flight", ErrUnavailable, cap(c.sem))
-	}
-	defer c.releaseSlot()
-
-	body, bypass, release, err := c.readBody(req)
-	if err != nil {
-		return Result{Verdict: restrict.DenyAppSecUnavailable, Release: release}, fmt.Errorf("%w: read body: %w", ErrUnavailable, err)
-	}
-
-	outbound, err := c.buildRequest(ctx, req, body)
-	if err != nil {
-		return Result{Verdict: restrict.DenyAppSecUnavailable, BodyBypass: bypass, Release: release}, fmt.Errorf("%w: %w", ErrUnavailable, err)
-	}
-
-	resp, err := c.http.Do(outbound)
-	if err != nil {
-		return Result{Verdict: restrict.DenyAppSecUnavailable, BodyBypass: bypass, Release: release}, fmt.Errorf("%w: %w", ErrUnavailable, err)
-	}
-	defer func() {
-		// Drain before closing. net/http only returns a connection to the idle
-		// pool once its body is read to EOF; closing with bytes outstanding
-		// discards it. Every verdict carries a JSON body, so skipping this
-		// would cost a fresh handshake per inspected request, inside the
-		// timeout budget.
-		if _, err := io.Copy(io.Discard, io.LimitReader(resp.Body, maxResponseBytes)); err != nil {
-			c.logger.Tracef("drain appsec response body: %v", err)
-		}
-		if err := resp.Body.Close(); err != nil {
-			c.logger.Tracef("close appsec response body: %v", err)
-		}
-	}()
-
-	verdict, err := c.verdict(resp)
-	return Result{Verdict: verdict, BodyBypass: bypass, Release: release}, err
-}
-
-// acquireSlot takes an in-flight slot without blocking, reporting false when
-// the engine is already at capacity.
-func (c *Client) acquireSlot() bool {
-	if c.sem == nil {
-		return true
-	}
-	select {
-	case c.sem <- struct{}{}:
-		return true
-	default:
-		return false
-	}
-}
-
-// releaseSlot returns the slot. Scoped to the engine call, not the request: the
-// buffered body outlives the call but the engine's attention does not.
-func (c *Client) releaseSlot() {
-	if c.sem == nil {
-		return
-	}
-	select {
-	case <-c.sem:
-	default:
-	}
-}
-
-// readBody buffers the body so it can be mirrored, always restoring it on the
-// original request. Returns nil when there is no body to forward: no body at
-// all, an upgrade request, or a body over the cap. A login form is forwarded
-// with its credential values redacted rather than suppressed.
-// release is never nil; the caller invokes it once the request has been served.
-func (c *Client) readBody(req Request) (body []byte, bypass string, release func(), err error) {
-	r := req.HTTP
-	if r.Body == nil || r.Body == http.NoBody {
-		return nil, "", noopRelease, nil
-	}
-	if c.maxBodyBytes < 0 {
-		return nil, BypassDisabled, noopRelease, nil
-	}
-	// A genuine upgrade request carries no body to inspect (net/http hands us
-	// http.NoBody, caught above); the hijacked stream is reached through
-	// Hijacker, never r.Body. The test has to be the forwarder's own, because a
-	// looser one would skip inspection for requests the forwarder still
-	// delivers to the backend with their body intact.
-	if netutil.IsUpgradeRequest(r.Header) {
-		return nil, BypassUpgrade, noopRelease, nil
-	}
-	// A Content-Length over the cap is known to be too large before reading.
-	if r.ContentLength > c.maxBodyBytes {
-		return nil, BypassOversize, noopRelease, nil
-	}
-
-	// Reserve the whole cap rather than the eventual length: the reservation
-	// has to be made before the body is read, and until then the only bound
-	// known is the cap. Skipping inspection when the pool is drained keeps a
-	// burst of large bodies from being an out-of-memory lever; the bypass is
-	// recorded so the gap in coverage is visible.
-	release = noopRelease
-	if c.budget != nil {
-		if !c.budget.Acquire(c.maxBodyBytes) {
-			c.logger.Debugf("appsec buffer budget exhausted, inspecting headers and URI only")
-			return nil, BypassBudget, noopRelease, nil
-		}
-		var once sync.Once
-		release = func() { once.Do(func() { c.budget.Release(c.maxBodyBytes) }) }
-	}
-
-	buffered, oversize, err := bufferBody(r, c.maxBodyBytes)
-	if err != nil {
-		// bufferBody restored r.Body from the bytes it did read, so the
-		// reservation stays held until the caller releases it.
-		return nil, "", release, err
-	}
-	// An oversize body was only partially read: a truncated prefix changes the
-	// engine's verdict in both directions, so inspect headers and URI only.
-	if oversize {
-		return nil, BypassOversize, release, nil
-	}
-	return redactFormFields(r.Header.Get("Content-Type"), buffered, req.RedactBodyFields), "", release, nil
-}
-
-// buildRequest assembles the mirrored request. Per the protocol it is a GET
-// when there is no body and a POST otherwise; bytes.Reader gives the outbound
-// request an accurate Content-Length, which the engine relies on to read the
-// body at all.
-func (c *Client) buildRequest(ctx context.Context, req Request, body []byte) (*http.Request, error) {
-	method := http.MethodGet
-	var payload io.Reader
-	if len(body) > 0 {
-		method = http.MethodPost
-		payload = bytes.NewReader(body)
-	}
-
-	outbound, err := http.NewRequestWithContext(ctx, method, c.url, payload)
-	if err != nil {
-		return nil, fmt.Errorf("build appsec request: %w", err)
-	}
-
-	r := req.HTTP
-	copyInspectableHeaders(outbound.Header, r.Header)
-	redactSecrets(outbound.Header, req)
-
-	outbound.Header.Set(headerAPIKey, c.apiKey)
-	outbound.Header.Set(headerIP, req.ClientIP.Unmap().String())
-	outbound.Header.Set(headerURI, mirroredURI(r.URL, req.RedactQueryParams))
-	outbound.Header.Set(headerVerb, r.Method)
-	outbound.Header.Set(headerHost, r.Host)
-	if ua := r.UserAgent(); ua != "" {
-		outbound.Header.Set(headerUserAgent, ua)
-	}
-	outbound.Header.Set(headerHTTPVersion, httpVersion(r))
-	if req.TransactionID != "" {
-		outbound.Header.Set(headerTransactionID, req.TransactionID)
-	}
-	return outbound, nil
-}
-
-// verdict maps the engine's response to a restrict.Verdict. 200 is a pass and
-// 401/500 are engine-side failures; every other status carries a remediation in
-// the body. The blocked status code is operator-configurable
-// (blocked_http_code), so the action field decides, not the status.
-func (c *Client) verdict(resp *http.Response) (restrict.Verdict, error) {
-	switch resp.StatusCode {
-	case http.StatusUnauthorized:
-		return restrict.DenyAppSecUnavailable, fmt.Errorf("%w: rejected api key", ErrUnavailable)
-	case http.StatusInternalServerError:
-		return restrict.DenyAppSecUnavailable, fmt.Errorf("%w: engine error", ErrUnavailable)
-	}
-
-	// Every status, 200 included, has to carry a decodable remediation. Taking a
-	// bare 200 as a pass would mean a URL pointing at anything that answers 200
-	// (a health endpoint, a load balancer's default page) silently allows every
-	// request while the service reports itself as enforcing.
-
-	var decoded struct {
-		Action string `json:"action"`
-	}
-	if err := json.NewDecoder(io.LimitReader(resp.Body, maxResponseBytes)).Decode(&decoded); err != nil {
-		// Every remediation carries a decodable action, so a response without
-		// one is not a verdict: most often the URL points at something that is
-		// not the AppSec endpoint, which answers 404 with HTML. Reported as
-		// unavailable rather than a ban so the access log names the real fault
-		// instead of sending an operator hunting for a rule that never fired.
-		// Enforce still blocks either way; only the recorded reason differs.
-		return restrict.DenyAppSecUnavailable, fmt.Errorf("%w: undecodable response (status %d): %w", ErrUnavailable, resp.StatusCode, err)
-	}
-
-	switch decoded.Action {
-	case actionAllow:
-		return restrict.Allow, nil
-	case actionCaptcha:
-		return restrict.DenyAppSecCaptcha, nil
-	case actionBan:
-		return restrict.DenyAppSecBan, nil
-	case "":
-		// Decodable JSON without a remediation is not a verdict either: the
-		// endpoint answered, but not as the engine. Same reasoning as an
-		// undecodable body, and the same reason to point at configuration.
-		return restrict.DenyAppSecUnavailable,
-			fmt.Errorf("%w: response carried no remediation (status %d)", ErrUnavailable, resp.StatusCode)
-	default:
-		// A remediation we do not implement still means the engine flagged the
-		// request, so deny.
-		c.logger.Debugf("unknown appsec action %q (status %d), treating as ban", decoded.Action, resp.StatusCode)
-		return restrict.DenyAppSecBan, nil
-	}
-}
-
-// copyInspectableHeaders copies the client's headers, which are what the WAF
-// rules actually match on, dropping hop-by-hop headers that describe the
-// proxy-to-engine connection rather than the client request, and any header in
-// the AppSec protocol namespace.
-func copyInspectableHeaders(dst, src http.Header) {
-	for name, values := range src {
-		if hopByHopHeaders[http.CanonicalHeaderKey(name)] {
-			continue
-		}
-		if strings.HasPrefix(http.CanonicalHeaderKey(name), headerPrefix) {
-			continue
-		}
-		dst[http.CanonicalHeaderKey(name)] = append([]string(nil), values...)
-	}
-	// Content-Length describes the mirrored payload, not the client's: net/http
-	// sets it from the body we actually attach. Content-Type is kept either way
-	// so rules matching on it still fire when the body was not forwarded.
-	dst.Del("Content-Length")
-}
-
-// redactSecrets replaces the credential values the proxy withholds from
-// backends, so the mirrored copy does not carry them either.
-func redactSecrets(dst http.Header, req Request) {
-	for _, name := range req.RedactHeaders {
-		// Presence, not Get: a header whose first value is empty still carries
-		// its later values to the engine, while the upstream strip deletes the
-		// name outright. Set collapses every value into the placeholder.
-		if len(dst.Values(name)) > 0 {
-			dst.Set(name, redactedPlaceholder)
-		}
-	}
-	// Every Cookie line, not just the first: a client may send several, and Get
-	// would leave the session cookie in any later one mirrored in the clear.
-	if cookies := dst.Values("Cookie"); len(cookies) > 0 {
-		redacted := make([]string, len(cookies))
-		for i, cookie := range cookies {
-			redacted[i] = redactCookieHeader(cookie, req.RedactCookies)
-		}
-		dst["Cookie"] = redacted
-	}
-}
-
-// mirroredURI renders the request target for the URI header, with the named
-// query parameter values replaced.
-func mirroredURI(u *url.URL, redactParams []string) string {
-	uri := u.RequestURI()
-	if u.RawQuery == "" || len(redactParams) == 0 {
-		return uri
-	}
-	redacted := redactQuery(u.RawQuery, redactParams)
-	if redacted == u.RawQuery {
-		return uri
-	}
-	// RequestURI is path + "?" + RawQuery; swap only the query part so the
-	// path keeps its original encoding.
-	return strings.TrimSuffix(uri, u.RawQuery) + redacted
-}
-
-var hopByHopHeaders = map[string]bool{
-	"Connection":          true,
-	"Keep-Alive":          true,
-	"Proxy-Authenticate":  true,
-	"Proxy-Authorization": true,
-	"Proxy-Connection":    true,
-	"Te":                  true,
-	"Trailer":             true,
-	"Transfer-Encoding":   true,
-	"Upgrade":             true,
-}
-
-// httpVersion renders the two-digit form the engine parses ("11", "20").
-func httpVersion(r *http.Request) string {
-	major, minor := r.ProtoMajor, r.ProtoMinor
-	if major < 0 || major > 9 || minor < 0 || minor > 9 {
-		return ""
-	}
-	return fmt.Sprintf("%d%d", major, minor)
-}

proxy/internal/auth/appsec_test.go

@@ -1,306 +0,0 @@
-package auth
-
-import (
-	"crypto/ed25519"
-	"encoding/base64"
-	"io"
-	"net/http"
-	"net/http/httptest"
-	"strings"
-	"sync"
-	"testing"
-	"time"
-
-	log "github.com/sirupsen/logrus"
-	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
-
-	"github.com/netbirdio/netbird/proxy/internal/appsec"
-	"github.com/netbirdio/netbird/proxy/internal/proxy"
-	"github.com/netbirdio/netbird/proxy/internal/restrict"
-	"github.com/netbirdio/netbird/proxy/internal/types"
-)
-
-// appsecEngine is a stub AppSec component returning a fixed remediation.
-func appsecEngine(t *testing.T, status int, body string) *httptest.Server {
-	t.Helper()
-	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
-		w.WriteHeader(status)
-		if body != "" {
-			_, _ = w.Write([]byte(body))
-		}
-	}))
-	t.Cleanup(srv.Close)
-	return srv
-}
-
-// serveWithAppSec runs a request through the middleware for a domain configured
-// with the given AppSec mode, returning the response and the captured metadata.
-func serveWithAppSec(t *testing.T, mode restrict.AppSecMode, client *appsec.Client, r *http.Request) (*httptest.ResponseRecorder, map[string]string, bool) {
-	t.Helper()
-
-	mw := NewMiddleware(log.StandardLogger(), nil, nil)
-	mw.SetAppSec(client)
-	require.NoError(t, mw.AddDomain("svc.example.com", DomainSettings{
-		AccountID:  types.AccountID("acct-1"),
-		ServiceID:  types.ServiceID("svc-1"),
-		AppSecMode: mode,
-	}))
-
-	reached := false
-	handler := mw.Protect(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
-		reached = true
-		w.WriteHeader(http.StatusOK)
-	}))
-
-	cd := proxy.NewCapturedData("req-1")
-	r = r.WithContext(proxy.WithCapturedData(r.Context(), cd))
-	rec := httptest.NewRecorder()
-	handler.ServeHTTP(rec, r)
-
-	return rec, cd.GetMetadata(), reached
-}
-
-func appsecRequest() *http.Request {
-	r := httptest.NewRequest(http.MethodGet, "http://svc.example.com/?x=/etc/passwd", nil)
-	r.Host = "svc.example.com"
-	r.RemoteAddr = "203.0.113.7:44444"
-	return r
-}
-
-func TestCheckAppSec_EnforceBlocksBannedRequest(t *testing.T) {
-	srv := appsecEngine(t, http.StatusForbidden, `{"action":"ban","http_status":403}`)
-	client, err := appsec.New(appsec.Config{URL: srv.URL, APIKey: "k"})
-	require.NoError(t, err)
-
-	rec, meta, reached := serveWithAppSec(t, restrict.AppSecEnforce, client, appsecRequest())
-
-	assert.Equal(t, http.StatusForbidden, rec.Code)
-	assert.False(t, reached, "a banned request must not reach the backend")
-	assert.Equal(t, "appsec_ban", meta["appsec_verdict"])
-	assert.NotContains(t, meta, "appsec_mode", "enforce is the default, only observe is annotated")
-}
-
-func TestCheckAppSec_ObserveAllowsAndRecordsVerdict(t *testing.T) {
-	srv := appsecEngine(t, http.StatusForbidden, `{"action":"ban","http_status":403}`)
-	client, err := appsec.New(appsec.Config{URL: srv.URL, APIKey: "k"})
-	require.NoError(t, err)
-
-	rec, meta, reached := serveWithAppSec(t, restrict.AppSecObserve, client, appsecRequest())
-
-	assert.Equal(t, http.StatusOK, rec.Code)
-	assert.True(t, reached, "observe mode must not block")
-	assert.Equal(t, "appsec_ban", meta["appsec_verdict"])
-	assert.Equal(t, "observe", meta["appsec_mode"])
-}
-
-func TestCheckAppSec_AllowedRequestPasses(t *testing.T) {
-	srv := appsecEngine(t, http.StatusOK, `{"action":"allow","http_status":200}`)
-	client, err := appsec.New(appsec.Config{URL: srv.URL, APIKey: "k"})
-	require.NoError(t, err)
-
-	rec, meta, reached := serveWithAppSec(t, restrict.AppSecEnforce, client, appsecRequest())
-
-	assert.Equal(t, http.StatusOK, rec.Code)
-	assert.True(t, reached)
-	assert.NotContains(t, meta, "appsec_verdict", "a clean request records no verdict")
-}
-
-func TestCheckAppSec_OffSkipsInspection(t *testing.T) {
-	// An engine that would ban everything; the mode must keep us away from it.
-	srv := appsecEngine(t, http.StatusForbidden, `{"action":"ban"}`)
-	client, err := appsec.New(appsec.Config{URL: srv.URL, APIKey: "k"})
-	require.NoError(t, err)
-
-	rec, meta, reached := serveWithAppSec(t, restrict.AppSecOff, client, appsecRequest())
-
-	assert.Equal(t, http.StatusOK, rec.Code)
-	assert.True(t, reached)
-	assert.Empty(t, meta)
-}
-
-func TestCheckAppSec_EnforceFailsClosedWithoutClient(t *testing.T) {
-	rec, meta, reached := serveWithAppSec(t, restrict.AppSecEnforce, nil, appsecRequest())
-
-	assert.Equal(t, http.StatusForbidden, rec.Code,
-		"enforce with no configured endpoint must deny rather than pass traffic uninspected")
-	assert.False(t, reached)
-	assert.Equal(t, "appsec_unavailable", meta["appsec_verdict"])
-}
-
-func TestCheckAppSec_ObserveAllowsWithoutClient(t *testing.T) {
-	rec, meta, reached := serveWithAppSec(t, restrict.AppSecObserve, nil, appsecRequest())
-
-	assert.Equal(t, http.StatusOK, rec.Code)
-	assert.True(t, reached)
-	assert.Equal(t, "appsec_unavailable", meta["appsec_verdict"])
-	assert.Equal(t, "observe", meta["appsec_mode"])
-}
-
-func TestCheckAppSec_EnforceFailsClosedWhenEngineUnreachable(t *testing.T) {
-	client, err := appsec.New(appsec.Config{URL: "http://127.0.0.1:1/", APIKey: "k"})
-	require.NoError(t, err)
-
-	rec, meta, reached := serveWithAppSec(t, restrict.AppSecEnforce, client, appsecRequest())
-
-	assert.Equal(t, http.StatusForbidden, rec.Code)
-	assert.False(t, reached)
-	assert.Equal(t, "appsec_unavailable", meta["appsec_verdict"])
-}
-
-func TestCheckAppSec_InspectsOverlayTraffic(t *testing.T) {
-	srv := appsecEngine(t, http.StatusForbidden, `{"action":"ban"}`)
-	client, err := appsec.New(appsec.Config{URL: srv.URL, APIKey: "k"})
-	require.NoError(t, err)
-
-	// Requests arriving over the WireGuard overlay skip the geo and IP-reputation
-	// checks, but request content is just as inspectable.
-	r := appsecRequest()
-	r = r.WithContext(types.WithOverlayOrigin(r.Context()))
-
-	rec, meta, reached := serveWithAppSec(t, restrict.AppSecEnforce, client, r)
-
-	assert.Equal(t, http.StatusForbidden, rec.Code, "overlay traffic must still be inspected")
-	assert.False(t, reached)
-	assert.Equal(t, "appsec_ban", meta["appsec_verdict"])
-}
-
-func TestCheckAppSec_UnresolvableClientIPFailsClosed(t *testing.T) {
-	srv := appsecEngine(t, http.StatusOK, `{"action":"allow"}`)
-	client, err := appsec.New(appsec.Config{URL: srv.URL, APIKey: "k"})
-	require.NoError(t, err)
-
-	r := appsecRequest()
-	r.RemoteAddr = "not-an-address"
-
-	rec, meta, reached := serveWithAppSec(t, restrict.AppSecEnforce, client, r)
-
-	assert.Equal(t, http.StatusForbidden, rec.Code,
-		"the engine requires a client address; a request we cannot attribute must not pass")
-	assert.False(t, reached)
-	assert.Equal(t, "appsec_unavailable", meta["appsec_verdict"])
-}
-
-// The redaction sets are resolved from the domain's schemes at registration, so
-// what AppSec withholds cannot drift from what those schemes actually accept.
-func TestAddDomain_ResolvesRedactionSetsFromSchemes(t *testing.T) {
-	mw := NewMiddleware(log.StandardLogger(), nil, nil)
-	require.NoError(t, mw.AddDomain("svc.example.com", DomainSettings{
-		Schemes: []Scheme{
-			NewPassword(nil, "svc-1", "acct-1"),
-			NewHeader(nil, "svc-1", "acct-1", "X-Api-Key"),
-		},
-		SessionPublicKey:  base64.StdEncoding.EncodeToString(make([]byte, ed25519.PublicKeySize)),
-		SessionExpiration: time.Hour,
-		AppSecMode:        restrict.AppSecEnforce,
-	}))
-
-	mw.domainsMux.RLock()
-	config := mw.domains["svc.example.com"]
-	mw.domainsMux.RUnlock()
-
-	assert.Equal(t, []string{"password"}, config.redactBodyFields)
-	assert.Equal(t, []string{"X-Api-Key"}, config.redactHeaders)
-	// r.FormValue merges the query into the form, so a credential passed there
-	// authenticates and must be redacted alongside the OIDC session token.
-	assert.Equal(t, []string{"session_token", "password"}, config.redactQueryParams)
-}
-
-// countingBudget records reservations so a test can observe when the
-// middleware hands them back.
-type countingBudget struct {
-	mu        sync.Mutex
-	total     int64
-	used      int64
-	maxAtOnce int64
-}
-
-func (b *countingBudget) Acquire(n int64) bool {
-	b.mu.Lock()
-	defer b.mu.Unlock()
-	if b.used+n > b.total {
-		return false
-	}
-	b.used += n
-	if b.used > b.maxAtOnce {
-		b.maxAtOnce = b.used
-	}
-	return true
-}
-
-func (b *countingBudget) Release(n int64) {
-	b.mu.Lock()
-	defer b.mu.Unlock()
-	b.used -= n
-}
-
-func (b *countingBudget) inUse() int64 {
-	b.mu.Lock()
-	defer b.mu.Unlock()
-	return b.used
-}
-
-// The buffered body stays alive as r.Body until the backend has read it, so
-// Protect must hold the reservation for the whole request and return it only
-// once the handler chain has unwound. Releasing inside Inspect would let the
-// budget admit buffering that is still resident.
-func TestProtect_AppSecBudgetHeldForRequestAndReleasedAfter(t *testing.T) {
-	srv := appsecEngine(t, http.StatusOK, `{"action":"allow"}`)
-	budget := &countingBudget{total: 1 << 20}
-	client, err := appsec.New(appsec.Config{
-		URL:          srv.URL,
-		APIKey:       "k",
-		MaxBodyBytes: 4096,
-		Budget:       budget,
-	})
-	require.NoError(t, err)
-
-	mw := NewMiddleware(log.StandardLogger(), nil, nil)
-	mw.SetAppSec(client)
-	require.NoError(t, mw.AddDomain("svc.example.com", DomainSettings{
-		AccountID:  types.AccountID("acct-1"),
-		ServiceID:  types.ServiceID("svc-1"),
-		AppSecMode: restrict.AppSecEnforce,
-	}))
-
-	var inHandler int64
-	handler := mw.Protect(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
-		// The backend reads the buffered body here, so the reservation must
-		// still be held at this point.
-		inHandler = budget.inUse()
-		_, _ = io.ReadAll(r.Body)
-		w.WriteHeader(http.StatusOK)
-	}))
-
-	r := httptest.NewRequest(http.MethodPost, "http://svc.example.com/", strings.NewReader("payload"))
-	r.Host = "svc.example.com"
-	cd := proxy.NewCapturedData("req-1")
-	r = r.WithContext(proxy.WithCapturedData(r.Context(), cd))
-
-	handler.ServeHTTP(httptest.NewRecorder(), r)
-
-	assert.Equal(t, int64(4096), inHandler, "the reservation must be held while the backend reads the body")
-	assert.Equal(t, int64(0), budget.inUse(), "Protect must release the reservation once the request is served")
-}
-
-// A denied request never reaches the backend, but Protect still has to hand the
-// reservation back or the pool leaks one cap per blocked request.
-func TestProtect_AppSecBudgetReleasedOnDeny(t *testing.T) {
-	srv := appsecEngine(t, http.StatusForbidden, `{"action":"ban"}`)
-	budget := &countingBudget{total: 1 << 20}
-	client, err := appsec.New(appsec.Config{
-		URL:          srv.URL,
-		APIKey:       "k",
-		MaxBodyBytes: 4096,
-		Budget:       budget,
-	})
-	require.NoError(t, err)
-
-	r := httptest.NewRequest(http.MethodPost, "http://svc.example.com/", strings.NewReader("payload"))
-	r.Host = "svc.example.com"
-	rec, _, reached := serveWithAppSec(t, restrict.AppSecEnforce, client, r)
-
-	assert.False(t, reached, "a banned request must not reach the backend")
-	assert.Equal(t, http.StatusForbidden, rec.Code)
-	assert.Equal(t, int64(0), budget.inUse(), "a blocked request must not leak its reservation")
-}

proxy/internal/auth/header.go

@@ -39,11 +39,6 @@ func (Header) Type() auth.Method {
 	return auth.MethodHeader
 }
 
-// HeaderName returns the request header this scheme reads its credential from.
-func (h Header) HeaderName() string {
-	return h.headerName
-}
-
 // Authenticate checks for the configured header in the request. If absent,
 // returns empty (unauthenticated). If present, validates via gRPC.
 func (h Header) Authenticate(r *http.Request) (string, string, error) {

proxy/internal/auth/middleware.go

@@ -18,7 +18,6 @@ import (
 	"google.golang.org/grpc"
 
 	"github.com/netbirdio/netbird/proxy/auth"
-	"github.com/netbirdio/netbird/proxy/internal/appsec"
 	"github.com/netbirdio/netbird/proxy/internal/proxy"
 	"github.com/netbirdio/netbird/proxy/internal/restrict"
 	"github.com/netbirdio/netbird/proxy/internal/types"
@@ -26,11 +25,6 @@ import (
 	"github.com/netbirdio/netbird/shared/management/proto"
 )
 
-// sessionCookieNames is the cookie set AppSec redacts before mirroring: the
-// proxy's session cookie is a bearer credential for the service, and the
-// reverse proxy already strips it before forwarding upstream.
-var sessionCookieNames = []string{auth.SessionCookieName}
-
 // errValidationUnavailable indicates that session validation failed due to
 // an infrastructure error (e.g. gRPC unavailable), not an invalid token.
 var errValidationUnavailable = errors.New("session validation unavailable")
@@ -65,14 +59,6 @@ type DomainConfig struct {
 	IPRestrictions    *restrict.Filter
 	// Private routes the domain through ValidateTunnelPeer; failure → 403.
 	Private bool
-	// AppSecMode enables CrowdSec AppSec request inspection for this domain.
-	AppSecMode restrict.AppSecMode
-	// redact* name the credentials this domain's schemes accept, resolved once
-	// at registration. AppSec replaces their values before mirroring a request,
-	// matching what the reverse proxy strips before forwarding upstream.
-	redactBodyFields  []string
-	redactHeaders     []string
-	redactQueryParams []string
 }
 
 type validationResult struct {
@@ -96,9 +82,6 @@ type Middleware struct {
 	sessionValidator SessionValidator
 	geo              restrict.GeoResolver
 	tunnelCache      *tunnelValidationCache
-	// appsec is the shared CrowdSec AppSec client, nil when the proxy has no
-	// AppSec endpoint configured. Set once during startup, before serving.
-	appsec *appsec.Client
 }
 
 // NewMiddleware creates a new authentication middleware. The sessionValidator is
@@ -116,12 +99,6 @@ func NewMiddleware(logger *log.Logger, sessionValidator SessionValidator, geo re
 	}
 }
 
-// SetAppSec installs the shared CrowdSec AppSec client. Must be called during
-// startup, before the middleware serves any request.
-func (mw *Middleware) SetAppSec(client *appsec.Client) {
-	mw.appsec = client
-}
-
 // Protect wraps next with per-domain authentication and IP restriction checks.
 // Requests whose Host is not registered pass through unchanged.
 func (mw *Middleware) Protect(next http.Handler) http.Handler {
@@ -146,14 +123,6 @@ func (mw *Middleware) Protect(next http.Handler) http.Handler {
 			return
 		}
 
-		// Deferred, not released here: the inspected body stays alive as r.Body
-		// until the backend has read it, which happens inside next.ServeHTTP.
-		appSecAllowed, releaseAppSec := mw.checkAppSec(w, r, config)
-		defer releaseAppSec()
-		if !appSecAllowed {
-			return
-		}
-
 		// Private services bypass operator schemes and gate on tunnel peer.
 		if config.Private {
 			if mw.forwardWithTunnelPeer(w, r, host, config, next) {
@@ -293,134 +262,6 @@ func (mw *Middleware) checkIPRestrictions(w http.ResponseWriter, r *http.Request
 	return false
 }
 
-// checkAppSec mirrors the request to the CrowdSec AppSec engine when the domain
-// enables inspection. Returns false when the request was blocked and a response
-// has been written.
-//
-// The returned release frees the body-buffering budget the inspection reserved
-// and is never nil. It must run only after the request has been served, since
-// the buffered body stays alive as r.Body for the backend to read.
-//
-// Every non-allow remediation blocks with 403, captcha included: the proxy has
-// no challenge flow to serve. The distinct verdict is still recorded so the
-// access log shows which remediation the engine actually chose.
-//
-// Unlike the geo and IP-reputation checks, this runs for overlay traffic too:
-// AppSec inspects request content, which is just as meaningful when the caller
-// reached the proxy through the WireGuard tunnel.
-func (mw *Middleware) checkAppSec(w http.ResponseWriter, r *http.Request, config DomainConfig) (bool, func()) {
-	if !config.AppSecMode.Enabled() {
-		return true, func() {}
-	}
-
-	verdict, release := mw.inspectAppSec(r, config)
-	if verdict == restrict.Allow {
-		return true, release
-	}
-
-	observe := config.AppSecMode == restrict.AppSecObserve
-	if cd := proxy.CapturedDataFromContext(r.Context()); cd != nil {
-		cd.SetMetadata("appsec_verdict", verdict.String())
-		if observe {
-			cd.SetMetadata("appsec_mode", "observe")
-		}
-	}
-
-	if observe {
-		mw.logger.Debugf("AppSec observe: would block %s for %s (%s)", r.RemoteAddr, r.Host, verdict)
-		return true, release
-	}
-
-	mw.markDenied(r, verdict.String())
-	mw.logger.Debugf("AppSec: %s for %s %s", verdict, r.Host, r.RemoteAddr)
-	http.Error(w, "Forbidden", http.StatusForbidden)
-	return false, release
-}
-
-// inspectAppSec runs the AppSec call and returns its verdict. Failures come
-// back as DenyAppSecUnavailable regardless of mode so observe mode still
-// records that inspection did not happen; the caller decides what blocks. The
-// returned release is never nil.
-func (mw *Middleware) inspectAppSec(r *http.Request, config DomainConfig) (restrict.Verdict, func()) {
-	// Mode requested but the proxy has no AppSec endpoint configured. Management
-	// gates this on the cluster capability; a stale mapping can still arrive.
-	if mw.appsec == nil {
-		mw.logger.Debugf("AppSec mode %q requested for %s but no AppSec endpoint is configured", config.AppSecMode, r.Host)
-		return restrict.DenyAppSecUnavailable, func() {}
-	}
-
-	clientIP := mw.resolveClientIP(r)
-	if !clientIP.IsValid() {
-		// The engine requires a client address, and a request whose source we
-		// cannot establish is exactly the kind we must not wave through.
-		mw.logger.Debugf("AppSec: cannot resolve client address for %q", r.RemoteAddr)
-		return restrict.DenyAppSecUnavailable, func() {}
-	}
-
-	req := appsec.Request{
-		HTTP:              r,
-		ClientIP:          clientIP,
-		RedactBodyFields:  config.redactBodyFields,
-		RedactHeaders:     config.redactHeaders,
-		RedactCookies:     sessionCookieNames,
-		RedactQueryParams: config.redactQueryParams,
-	}
-	if cd := proxy.CapturedDataFromContext(r.Context()); cd != nil {
-		req.TransactionID = cd.GetRequestID()
-	}
-
-	result, err := mw.appsec.Inspect(r.Context(), req)
-	if err != nil {
-		mw.logger.Debugf("AppSec inspection failed for %s: %v", r.Host, err)
-	}
-	// Record when the body went uninspected: headers and URI were still
-	// checked, but an operator reading the log should not read a clean verdict
-	// as "the payload was examined". Oversize is reachable by padding, so its
-	// absence from the log would hide a deliberate opt-out.
-	if result.BodyBypass != "" {
-		if cd := proxy.CapturedDataFromContext(r.Context()); cd != nil {
-			cd.SetMetadata("appsec_body_bypass", result.BodyBypass)
-		}
-	}
-	return result.Verdict, result.Release
-}
-
-// credentialFormFields lists the login form fields whose values are redacted
-// from the mirrored body, so a password or PIN submitted to the proxy's own
-// login form never reaches the Security Engine.
-func credentialFormFields(schemes []Scheme) []string {
-	var fields []string
-	for _, s := range schemes {
-		switch s.Type() {
-		case auth.MethodPassword:
-			fields = append(fields, passwordFormId)
-		case auth.MethodPIN:
-			fields = append(fields, pinFormId)
-		}
-	}
-	return fields
-}
-
-// credentialHeaders lists the request headers whose values are redacted from
-// the mirrored request. A header-auth scheme carries a session token the proxy
-// validates and never forwards upstream, so the engine must not see it either.
-func credentialHeaders(schemes []Scheme) []string {
-	var names []string
-	for _, s := range schemes {
-		// Structural, not a concrete Header assertion: if the scheme is ever
-		// registered as a pointer, a type assertion would quietly stop matching
-		// and the header would start reaching the engine again.
-		named, ok := s.(interface{ HeaderName() string })
-		if !ok {
-			continue
-		}
-		if name := named.HeaderName(); name != "" {
-			names = append(names, name)
-		}
-	}
-	return names
-}
-
 // resolveClientIP extracts the real client IP from CapturedData, falling back to r.RemoteAddr.
 func (mw *Middleware) resolveClientIP(r *http.Request) netip.Addr {
 	if cd := proxy.CapturedDataFromContext(r.Context()); cd != nil {
@@ -440,18 +281,12 @@ func (mw *Middleware) resolveClientIP(r *http.Request) netip.Addr {
 	return addr.Unmap()
 }
 
-// markDenied records the deny reason on the captured data so the access log
-// attributes the response to the proxy rather than the backend.
-func (mw *Middleware) markDenied(r *http.Request, reason string) {
+// blockIPRestriction sets captured data fields for an IP-restriction block event.
+func (mw *Middleware) blockIPRestriction(r *http.Request, reason string) {
 	if cd := proxy.CapturedDataFromContext(r.Context()); cd != nil {
 		cd.SetOrigin(proxy.OriginAuth)
 		cd.SetAuthMethod(reason)
 	}
-}
-
-// blockIPRestriction sets captured data fields for an IP-restriction block event.
-func (mw *Middleware) blockIPRestriction(r *http.Request, reason string) {
-	mw.markDenied(r, reason)
 	mw.logger.Debugf("IP restriction: %s for %s", reason, r.RemoteAddr)
 }
 
@@ -802,61 +637,45 @@ func wasCredentialSubmitted(r *http.Request, method auth.Method) bool {
 	case auth.MethodPassword:
 		return r.FormValue("password") != ""
 	case auth.MethodOIDC:
-		return r.URL.Query().Get(sessionTokenParam) != ""
+		return r.URL.Query().Get("session_token") != ""
 	}
 	return false
 }
 
-// DomainSettings is the per-domain configuration AddDomain applies.
-type DomainSettings struct {
-	Schemes []Scheme
-	// SessionPublicKey is the base64-encoded ed25519 key used to verify session
-	// cookies. Required when Schemes is non-empty.
-	SessionPublicKey  string
-	SessionExpiration time.Duration
-	AccountID         types.AccountID
-	ServiceID         types.ServiceID
-	IPRestrictions    *restrict.Filter
-	// Private forces ValidateTunnelPeer enforcement (403 on failure) regardless
-	// of the schemes list.
-	Private    bool
-	AppSecMode restrict.AppSecMode
-}
-
-// AddDomain registers authentication schemes for the given domain. With schemes
-// a valid session public key is required.
-func (mw *Middleware) AddDomain(domain string, settings DomainSettings) error {
-	credentialFields := credentialFormFields(settings.Schemes)
-	config := DomainConfig{
-		AccountID:        settings.AccountID,
-		ServiceID:        settings.ServiceID,
-		IPRestrictions:   settings.IPRestrictions,
-		Private:          settings.Private,
-		AppSecMode:       settings.AppSecMode,
-		redactBodyFields: credentialFields,
-		redactHeaders:    credentialHeaders(settings.Schemes),
-		// A credential can arrive in the query too: r.FormValue merges the URL
-		// query into the form, so "?password=..." authenticates just as a form
-		// post does and must not be mirrored in the clear either.
-		redactQueryParams: append([]string{sessionTokenParam}, credentialFields...),
+// AddDomain registers authentication schemes for the given domain. With schemes a valid session public key is required.
+// private=true forces ValidateTunnelPeer enforcement (403 on failure) regardless of the schemes list.
+func (mw *Middleware) AddDomain(domain string, schemes []Scheme, publicKeyB64 string, expiration time.Duration, accountID types.AccountID, serviceID types.ServiceID, ipRestrictions *restrict.Filter, private bool) error {
+	if len(schemes) == 0 {
+		mw.domainsMux.Lock()
+		defer mw.domainsMux.Unlock()
+		mw.domains[domain] = DomainConfig{
+			AccountID:      accountID,
+			ServiceID:      serviceID,
+			IPRestrictions: ipRestrictions,
+			Private:        private,
+		}
+		return nil
 	}
 
-	if len(settings.Schemes) > 0 {
-		pubKeyBytes, err := base64.StdEncoding.DecodeString(settings.SessionPublicKey)
-		if err != nil {
-			return fmt.Errorf("decode session public key for domain %s: %w", domain, err)
-		}
-		if len(pubKeyBytes) != ed25519.PublicKeySize {
-			return fmt.Errorf("invalid session public key size for domain %s: got %d, want %d", domain, len(pubKeyBytes), ed25519.PublicKeySize)
-		}
-		config.Schemes = settings.Schemes
-		config.SessionPublicKey = pubKeyBytes
-		config.SessionExpiration = settings.SessionExpiration
+	pubKeyBytes, err := base64.StdEncoding.DecodeString(publicKeyB64)
+	if err != nil {
+		return fmt.Errorf("decode session public key for domain %s: %w", domain, err)
+	}
+	if len(pubKeyBytes) != ed25519.PublicKeySize {
+		return fmt.Errorf("invalid session public key size for domain %s: got %d, want %d", domain, len(pubKeyBytes), ed25519.PublicKeySize)
 	}
 
 	mw.domainsMux.Lock()
 	defer mw.domainsMux.Unlock()
-	mw.domains[domain] = config
+	mw.domains[domain] = DomainConfig{
+		Schemes:           schemes,
+		SessionPublicKey:  pubKeyBytes,
+		SessionExpiration: expiration,
+		AccountID:         accountID,
+		ServiceID:         serviceID,
+		IPRestrictions:    ipRestrictions,
+		Private:           private,
+	}
 	return nil
 }
 
@@ -911,10 +730,10 @@ func (mw *Middleware) validateSessionToken(ctx context.Context, host, token stri
 // parameter removed so it doesn't linger in the browser's address bar or history.
 func stripSessionTokenParam(u *url.URL) string {
 	q := u.Query()
-	if !q.Has(sessionTokenParam) {
+	if !q.Has("session_token") {
 		return u.RequestURI()
 	}
-	q.Del(sessionTokenParam)
+	q.Del("session_token")
 	clean := *u
 	clean.RawQuery = q.Encode()
 	return clean.RequestURI()

proxy/internal/auth/middleware_test.go

@@ -64,7 +64,7 @@ func TestAddDomain_ValidKey(t *testing.T) {
 	kp := generateTestKeyPair(t)
 
 	scheme := &stubScheme{method: auth.MethodPIN, promptID: "pin"}
-	err := mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour})
+	err := mw.AddDomain("example.com", []Scheme{scheme}, kp.PublicKey, time.Hour, "", "", nil, false)
 	require.NoError(t, err)
 
 	mw.domainsMux.RLock()
@@ -81,7 +81,7 @@ func TestAddDomain_EmptyKey(t *testing.T) {
 	mw := NewMiddleware(log.StandardLogger(), nil, nil)
 
 	scheme := &stubScheme{method: auth.MethodPIN, promptID: "pin"}
-	err := mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionExpiration: time.Hour})
+	err := mw.AddDomain("example.com", []Scheme{scheme}, "", time.Hour, "", "", nil, false)
 	require.Error(t, err)
 	assert.Contains(t, err.Error(), "invalid session public key size")
 
@@ -95,7 +95,7 @@ func TestAddDomain_InvalidBase64(t *testing.T) {
 	mw := NewMiddleware(log.StandardLogger(), nil, nil)
 
 	scheme := &stubScheme{method: auth.MethodPIN, promptID: "pin"}
-	err := mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: "not-valid-base64!!!", SessionExpiration: time.Hour})
+	err := mw.AddDomain("example.com", []Scheme{scheme}, "not-valid-base64!!!", time.Hour, "", "", nil, false)
 	require.Error(t, err)
 	assert.Contains(t, err.Error(), "decode session public key")
 
@@ -110,7 +110,7 @@ func TestAddDomain_WrongKeySize(t *testing.T) {
 
 	shortKey := base64.StdEncoding.EncodeToString([]byte("tooshort"))
 	scheme := &stubScheme{method: auth.MethodPIN, promptID: "pin"}
-	err := mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: shortKey, SessionExpiration: time.Hour})
+	err := mw.AddDomain("example.com", []Scheme{scheme}, shortKey, time.Hour, "", "", nil, false)
 	require.Error(t, err)
 	assert.Contains(t, err.Error(), "invalid session public key size")
 
@@ -123,7 +123,7 @@ func TestAddDomain_WrongKeySize(t *testing.T) {
 func TestAddDomain_NoSchemes_NoKeyRequired(t *testing.T) {
 	mw := NewMiddleware(log.StandardLogger(), nil, nil)
 
-	err := mw.AddDomain("example.com", DomainSettings{SessionExpiration: time.Hour})
+	err := mw.AddDomain("example.com", nil, "", time.Hour, "", "", nil, false)
 	require.NoError(t, err, "domains with no auth schemes should not require a key")
 
 	mw.domainsMux.RLock()
@@ -139,8 +139,8 @@ func TestAddDomain_OverwritesPreviousConfig(t *testing.T) {
 
 	scheme := &stubScheme{method: auth.MethodPIN, promptID: "pin"}
 
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: kp1.PublicKey, SessionExpiration: time.Hour}))
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: kp2.PublicKey, SessionExpiration: 2 * time.Hour}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{scheme}, kp1.PublicKey, time.Hour, "", "", nil, false))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{scheme}, kp2.PublicKey, 2*time.Hour, "", "", nil, false))
 
 	mw.domainsMux.RLock()
 	config := mw.domains["example.com"]
@@ -156,7 +156,7 @@ func TestRemoveDomain(t *testing.T) {
 	kp := generateTestKeyPair(t)
 
 	scheme := &stubScheme{method: auth.MethodPIN, promptID: "pin"}
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{scheme}, kp.PublicKey, time.Hour, "", "", nil, false))
 
 	mw.RemoveDomain("example.com")
 
@@ -180,7 +180,7 @@ func TestProtect_UnknownDomainPassesThrough(t *testing.T) {
 
 func TestProtect_DomainWithNoSchemesPassesThrough(t *testing.T) {
 	mw := NewMiddleware(log.StandardLogger(), nil, nil)
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{SessionExpiration: time.Hour}))
+	require.NoError(t, mw.AddDomain("example.com", nil, "", time.Hour, "", "", nil, false))
 
 	handler := mw.Protect(newPassthroughHandler())
 
@@ -197,7 +197,7 @@ func TestProtect_UnauthenticatedRequestIsBlocked(t *testing.T) {
 	kp := generateTestKeyPair(t)
 
 	scheme := &stubScheme{method: auth.MethodPIN, promptID: "pin"}
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{scheme}, kp.PublicKey, time.Hour, "", "", nil, false))
 
 	var backendCalled bool
 	backend := http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
@@ -218,7 +218,7 @@ func TestProtect_HostWithPortIsMatched(t *testing.T) {
 	kp := generateTestKeyPair(t)
 
 	scheme := &stubScheme{method: auth.MethodPIN, promptID: "pin"}
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{scheme}, kp.PublicKey, time.Hour, "", "", nil, false))
 
 	var backendCalled bool
 	backend := http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
@@ -239,7 +239,7 @@ func TestProtect_ValidSessionCookiePassesThrough(t *testing.T) {
 	kp := generateTestKeyPair(t)
 
 	scheme := &stubScheme{method: auth.MethodPIN, promptID: "pin"}
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{scheme}, kp.PublicKey, time.Hour, "", "", nil, false))
 
 	token, err := sessionkey.SignToken(kp.PrivateKey, "test-user", "", "example.com", auth.MethodPIN, nil, nil, time.Hour)
 	require.NoError(t, err)
@@ -272,7 +272,7 @@ func TestProtect_SessionCookieGroupsPropagate(t *testing.T) {
 	kp := generateTestKeyPair(t)
 
 	scheme := &stubScheme{method: auth.MethodPIN, promptID: "pin"}
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{scheme}, kp.PublicKey, time.Hour, "", "", nil, false))
 
 	groups := []string{"engineering", "sre"}
 	token, err := sessionkey.SignToken(kp.PrivateKey, "test-user", "", "example.com", auth.MethodPIN, groups, nil, time.Hour)
@@ -337,7 +337,7 @@ func TestProtect_PrivateService_TunnelPeerGroupsPropagate(t *testing.T) {
 	kp := generateTestKeyPair(t)
 
 	// Private service: no operator schemes — auth gates solely on the tunnel peer.
-	require.NoError(t, mw.AddDomain("agent.example.com", DomainSettings{SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour, AccountID: "acct-1", ServiceID: "svc-1", Private: true}))
+	require.NoError(t, mw.AddDomain("agent.example.com", nil, kp.PublicKey, time.Hour, "acct-1", "svc-1", nil, true))
 
 	cd := proxy.NewCapturedData("")
 	cd.SetClientIP(netip.MustParseAddr("100.90.1.14")) // CGNAT tunnel source
@@ -377,7 +377,7 @@ func TestProtect_PrivateService_TunnelPeerDenied(t *testing.T) {
 	}}
 	mw := NewMiddleware(log.StandardLogger(), validator, nil)
 	kp := generateTestKeyPair(t)
-	require.NoError(t, mw.AddDomain("agent.example.com", DomainSettings{SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour, AccountID: "acct-1", ServiceID: "svc-1", Private: true}))
+	require.NoError(t, mw.AddDomain("agent.example.com", nil, kp.PublicKey, time.Hour, "acct-1", "svc-1", nil, true))
 
 	cd := proxy.NewCapturedData("")
 	cd.SetClientIP(netip.MustParseAddr("100.90.1.14"))
@@ -405,7 +405,7 @@ func TestProtect_ExpiredSessionCookieIsRejected(t *testing.T) {
 	kp := generateTestKeyPair(t)
 
 	scheme := &stubScheme{method: auth.MethodPIN, promptID: "pin"}
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{scheme}, kp.PublicKey, time.Hour, "", "", nil, false))
 
 	// Sign a token that expired 1 second ago.
 	token, err := sessionkey.SignToken(kp.PrivateKey, "test-user", "", "example.com", auth.MethodPIN, nil, nil, -time.Second)
@@ -431,7 +431,7 @@ func TestProtect_WrongDomainCookieIsRejected(t *testing.T) {
 	kp := generateTestKeyPair(t)
 
 	scheme := &stubScheme{method: auth.MethodPIN, promptID: "pin"}
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{scheme}, kp.PublicKey, time.Hour, "", "", nil, false))
 
 	// Token signed for a different domain audience.
 	token, err := sessionkey.SignToken(kp.PrivateKey, "test-user", "", "other.com", auth.MethodPIN, nil, nil, time.Hour)
@@ -458,7 +458,7 @@ func TestProtect_WrongKeyCookieIsRejected(t *testing.T) {
 	kp2 := generateTestKeyPair(t)
 
 	scheme := &stubScheme{method: auth.MethodPIN, promptID: "pin"}
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: kp1.PublicKey, SessionExpiration: time.Hour}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{scheme}, kp1.PublicKey, time.Hour, "", "", nil, false))
 
 	// Token signed with a different private key.
 	token, err := sessionkey.SignToken(kp2.PrivateKey, "test-user", "", "example.com", auth.MethodPIN, nil, nil, time.Hour)
@@ -495,7 +495,7 @@ func TestProtect_SchemeAuthRedirectsWithCookie(t *testing.T) {
 			return "", "pin", nil
 		},
 	}
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{scheme}, kp.PublicKey, time.Hour, "", "", nil, false))
 
 	var backendCalled bool
 	backend := http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
@@ -548,7 +548,7 @@ func TestProtect_FailedAuthDoesNotSetCookie(t *testing.T) {
 			return "", "pin", nil
 		},
 	}
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{scheme}, kp.PublicKey, time.Hour, "", "", nil, false))
 
 	handler := mw.Protect(newPassthroughHandler())
 
@@ -584,7 +584,7 @@ func TestProtect_MultipleSchemes(t *testing.T) {
 			return "", "password", nil
 		},
 	}
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{pinScheme, passwordScheme}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{pinScheme, passwordScheme}, kp.PublicKey, time.Hour, "", "", nil, false))
 
 	var backendCalled bool
 	backend := http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
@@ -614,7 +614,7 @@ func TestProtect_InvalidTokenFromSchemeReturns400(t *testing.T) {
 			return "invalid-jwt-token", "", nil
 		},
 	}
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{scheme}, kp.PublicKey, time.Hour, "", "", nil, false))
 
 	handler := mw.Protect(newPassthroughHandler())
 
@@ -638,7 +638,7 @@ func TestAddDomain_RandomBytes32NotEd25519(t *testing.T) {
 	key := base64.StdEncoding.EncodeToString(randomBytes)
 	scheme := &stubScheme{method: auth.MethodPIN, promptID: "pin"}
 
-	err = mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: key, SessionExpiration: time.Hour})
+	err = mw.AddDomain("example.com", []Scheme{scheme}, key, time.Hour, "", "", nil, false)
 	require.NoError(t, err, "any 32-byte key should be accepted at registration time")
 }
 
@@ -647,10 +647,10 @@ func TestAddDomain_InvalidKeyDoesNotCorruptExistingConfig(t *testing.T) {
 	kp := generateTestKeyPair(t)
 
 	scheme := &stubScheme{method: auth.MethodPIN, promptID: "pin"}
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{scheme}, kp.PublicKey, time.Hour, "", "", nil, false))
 
 	// Attempt to overwrite with an invalid key.
-	err := mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: "bad", SessionExpiration: time.Hour})
+	err := mw.AddDomain("example.com", []Scheme{scheme}, "bad", time.Hour, "", "", nil, false)
 	require.Error(t, err)
 
 	// The original valid config should still be intact.
@@ -674,7 +674,7 @@ func TestProtect_FailedPinAuthCapturesAuthMethod(t *testing.T) {
 			return "", "pin", nil
 		},
 	}
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{scheme}, kp.PublicKey, time.Hour, "", "", nil, false))
 
 	capturedData := proxy.NewCapturedData("")
 	handler := mw.Protect(newPassthroughHandler())
@@ -701,7 +701,7 @@ func TestProtect_FailedPasswordAuthCapturesAuthMethod(t *testing.T) {
 			return "", "password", nil
 		},
 	}
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{scheme}, kp.PublicKey, time.Hour, "", "", nil, false))
 
 	capturedData := proxy.NewCapturedData("")
 	handler := mw.Protect(newPassthroughHandler())
@@ -728,7 +728,7 @@ func TestProtect_NoCredentialsDoesNotCaptureAuthMethod(t *testing.T) {
 			return "", "pin", nil
 		},
 	}
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{scheme}, kp.PublicKey, time.Hour, "", "", nil, false))
 
 	capturedData := proxy.NewCapturedData("")
 	handler := mw.Protect(newPassthroughHandler())
@@ -815,7 +815,8 @@ func TestWasCredentialSubmitted(t *testing.T) {
 func TestCheckIPRestrictions_UnparseableAddress(t *testing.T) {
 	mw := NewMiddleware(log.StandardLogger(), nil, nil)
 
-	err := mw.AddDomain("example.com", DomainSettings{AccountID: "acc1", ServiceID: "svc1", IPRestrictions: restrict.ParseFilter(restrict.FilterConfig{AllowedCIDRs: []string{"10.0.0.0/8"}})})
+	err := mw.AddDomain("example.com", nil, "", 0, "acc1", "svc1",
+		restrict.ParseFilter(restrict.FilterConfig{AllowedCIDRs: []string{"10.0.0.0/8"}}), false)
 	require.NoError(t, err)
 
 	handler := mw.Protect(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
@@ -850,7 +851,8 @@ func TestCheckIPRestrictions_UsesCapturedDataClientIP(t *testing.T) {
 	// trusted proxies), checkIPRestrictions should use that IP, not RemoteAddr.
 	mw := NewMiddleware(log.StandardLogger(), nil, nil)
 
-	err := mw.AddDomain("example.com", DomainSettings{AccountID: "acc1", ServiceID: "svc1", IPRestrictions: restrict.ParseFilter(restrict.FilterConfig{AllowedCIDRs: []string{"203.0.113.0/24"}})})
+	err := mw.AddDomain("example.com", nil, "", 0, "acc1", "svc1",
+		restrict.ParseFilter(restrict.FilterConfig{AllowedCIDRs: []string{"203.0.113.0/24"}}), false)
 	require.NoError(t, err)
 
 	handler := mw.Protect(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
@@ -890,7 +892,8 @@ func TestCheckIPRestrictions_NilGeoWithCountryRules(t *testing.T) {
 	// Geo is nil, country restrictions are configured: must deny (fail-close).
 	mw := NewMiddleware(log.StandardLogger(), nil, nil)
 
-	err := mw.AddDomain("example.com", DomainSettings{AccountID: "acc1", ServiceID: "svc1", IPRestrictions: restrict.ParseFilter(restrict.FilterConfig{AllowedCountries: []string{"US"}})})
+	err := mw.AddDomain("example.com", nil, "", 0, "acc1", "svc1",
+		restrict.ParseFilter(restrict.FilterConfig{AllowedCountries: []string{"US"}}), false)
 	require.NoError(t, err)
 
 	handler := mw.Protect(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
@@ -913,10 +916,11 @@ func TestCheckIPRestrictions_NilGeoWithCountryRules(t *testing.T) {
 func TestCheckIPRestrictions_OverlayOriginSkipsCountryRules(t *testing.T) {
 	mw := NewMiddleware(log.StandardLogger(), nil, nil)
 
-	err := mw.AddDomain("example.com", DomainSettings{AccountID: "acc1", ServiceID: "svc1", IPRestrictions: restrict.ParseFilter(restrict.FilterConfig{
-		AllowedCIDRs:     []string{"100.64.0.0/10"},
-		AllowedCountries: []string{"US"},
-	})})
+	err := mw.AddDomain("example.com", nil, "", 0, "acc1", "svc1",
+		restrict.ParseFilter(restrict.FilterConfig{
+			AllowedCIDRs:     []string{"100.64.0.0/10"},
+			AllowedCountries: []string{"US"},
+		}), false)
 	require.NoError(t, err)
 
 	handler := mw.Protect(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
@@ -949,7 +953,8 @@ func TestCheckIPRestrictions_OverlayOriginSkipsCountryRules(t *testing.T) {
 func TestCheckIPRestrictions_OverlayOriginRespectsCIDR(t *testing.T) {
 	mw := NewMiddleware(log.StandardLogger(), nil, nil)
 
-	err := mw.AddDomain("example.com", DomainSettings{AccountID: "acc1", ServiceID: "svc1", IPRestrictions: restrict.ParseFilter(restrict.FilterConfig{AllowedCIDRs: []string{"100.64.0.0/16"}})})
+	err := mw.AddDomain("example.com", nil, "", 0, "acc1", "svc1",
+		restrict.ParseFilter(restrict.FilterConfig{AllowedCIDRs: []string{"100.64.0.0/16"}}), false)
 	require.NoError(t, err)
 
 	handler := mw.Protect(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
@@ -977,7 +982,7 @@ func TestProtect_OIDCOnlyRedirectsDirectly(t *testing.T) {
 			return "", oidcURL, nil
 		},
 	}
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{scheme}, kp.PublicKey, time.Hour, "", "", nil, false))
 
 	handler := mw.Protect(newPassthroughHandler())
 
@@ -1006,7 +1011,7 @@ func TestProtect_OIDCWithOtherMethodShowsLoginPage(t *testing.T) {
 			return "", "pin", nil
 		},
 	}
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{oidcScheme, pinScheme}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{oidcScheme, pinScheme}, kp.PublicKey, time.Hour, "", "", nil, false))
 
 	handler := mw.Protect(newPassthroughHandler())
 
@@ -1050,7 +1055,7 @@ func TestProtect_HeaderAuth_ForwardsOnSuccess(t *testing.T) {
 	kp := generateTestKeyPair(t)
 
 	hdr := newHeaderSchemeWithToken(t, kp, "X-API-Key", "secret-key")
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{hdr}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour, AccountID: "acc1", ServiceID: "svc1"}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{hdr}, kp.PublicKey, time.Hour, "acc1", "svc1", nil, false))
 
 	var backendCalled bool
 	capturedData := proxy.NewCapturedData("")
@@ -1093,7 +1098,7 @@ func TestProtect_HeaderAuth_MissingHeaderFallsThrough(t *testing.T) {
 	hdr := newHeaderSchemeWithToken(t, kp, "X-API-Key", "secret-key")
 	// Also add a PIN scheme so we can verify fallthrough behavior.
 	pinScheme := &stubScheme{method: auth.MethodPIN, promptID: "pin"}
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{hdr, pinScheme}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour, AccountID: "acc1", ServiceID: "svc1"}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{hdr, pinScheme}, kp.PublicKey, time.Hour, "acc1", "svc1", nil, false))
 
 	handler := mw.Protect(newPassthroughHandler())
 
@@ -1113,7 +1118,7 @@ func TestProtect_HeaderAuth_WrongValueReturns401(t *testing.T) {
 		return &proto.AuthenticateResponse{Success: false}, nil
 	}}
 	hdr := NewHeader(mock, "svc1", "acc1", "X-API-Key")
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{hdr}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour, AccountID: "acc1", ServiceID: "svc1"}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{hdr}, kp.PublicKey, time.Hour, "acc1", "svc1", nil, false))
 
 	capturedData := proxy.NewCapturedData("")
 	handler := mw.Protect(newPassthroughHandler())
@@ -1136,7 +1141,7 @@ func TestProtect_HeaderAuth_InfraErrorReturns502(t *testing.T) {
 		return nil, errors.New("gRPC unavailable")
 	}}
 	hdr := NewHeader(mock, "svc1", "acc1", "X-API-Key")
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{hdr}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour, AccountID: "acc1", ServiceID: "svc1"}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{hdr}, kp.PublicKey, time.Hour, "acc1", "svc1", nil, false))
 
 	handler := mw.Protect(newPassthroughHandler())
 
@@ -1153,7 +1158,7 @@ func TestProtect_HeaderAuth_SubsequentRequestUsesSessionCookie(t *testing.T) {
 	kp := generateTestKeyPair(t)
 
 	hdr := newHeaderSchemeWithToken(t, kp, "X-API-Key", "secret-key")
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{hdr}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour, AccountID: "acc1", ServiceID: "svc1"}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{hdr}, kp.PublicKey, time.Hour, "acc1", "svc1", nil, false))
 
 	handler := mw.Protect(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
 		w.WriteHeader(http.StatusOK)
@@ -1213,7 +1218,7 @@ func TestProtect_HeaderAuth_MultipleValuesSameHeader(t *testing.T) {
 
 	// Single Header scheme (as if one entry existed), but the mock checks both values.
 	hdr := NewHeader(mock, "svc1", "acc1", "Authorization")
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{hdr}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour, AccountID: "acc1", ServiceID: "svc1"}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{hdr}, kp.PublicKey, time.Hour, "acc1", "svc1", nil, false))
 
 	var backendCalled bool
 	handler := mw.Protect(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
@@ -1271,7 +1276,7 @@ func TestProtect_OIDCOnPlainHTTP_BlockedWith400(t *testing.T) {
 			return "", "https://idp.example.com/authorize", nil
 		},
 	}
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{scheme}, kp.PublicKey, time.Hour, "", "", nil, false))
 
 	handler := mw.Protect(newPassthroughHandler())
 
@@ -1295,7 +1300,7 @@ func TestProtect_OIDCOverTLS_NotBlocked(t *testing.T) {
 			return "", "https://idp.example.com/authorize", nil
 		},
 	}
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{scheme}, kp.PublicKey, time.Hour, "", "", nil, false))
 
 	handler := mw.Protect(newPassthroughHandler())
 
@@ -1315,7 +1320,7 @@ func TestProtect_NonOIDCSchemes_PlainHTTP_NotBlocked(t *testing.T) {
 	kp := generateTestKeyPair(t)
 
 	scheme := &stubScheme{method: auth.MethodPIN, promptID: "pin"}
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{scheme}, kp.PublicKey, time.Hour, "", "", nil, false))
 
 	handler := mw.Protect(newPassthroughHandler())
 
@@ -1345,7 +1350,7 @@ func TestProtect_TunnelPeerFastPath_RequiresInboundMarker(t *testing.T) {
 	kp := generateTestKeyPair(t)
 
 	scheme := &stubScheme{method: auth.MethodPIN, promptID: "pin"}
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{scheme}, kp.PublicKey, time.Hour, "", "", nil, false))
 
 	handler := mw.Protect(newPassthroughHandler())
 
@@ -1380,7 +1385,7 @@ func TestProtect_TunnelPeerFastPath_TakesPathWithInboundMarker(t *testing.T) {
 	kp := generateTestKeyPair(t)
 
 	scheme := &stubScheme{method: auth.MethodPIN, promptID: "pin"}
-	require.NoError(t, mw.AddDomain("example.com", DomainSettings{Schemes: []Scheme{scheme}, SessionPublicKey: kp.PublicKey, SessionExpiration: time.Hour}))
+	require.NoError(t, mw.AddDomain("example.com", []Scheme{scheme}, kp.PublicKey, time.Hour, "", "", nil, false))
 
 	handler := mw.Protect(newPassthroughHandler())
 

proxy/internal/auth/oidc.go

@@ -13,10 +13,6 @@ import (
 	"github.com/netbirdio/netbird/shared/management/proto"
 )
 
-// sessionTokenParam is the query parameter the management server uses to hand
-// the minted session token back to the proxy after an OIDC login.
-const sessionTokenParam = "session_token"
-
 type urlGenerator interface {
 	GetOIDCURL(context.Context, *proto.GetOIDCURLRequest, ...grpc.CallOption) (*proto.GetOIDCURLResponse, error)
 }
@@ -47,7 +43,7 @@ func (o OIDC) Authenticate(r *http.Request) (string, string, error) {
 	// Check for the session_token query param (from OIDC redirects).
 	// The management server passes the token in the URL because it cannot set
 	// cookies for the proxy's domain (cookies are domain-scoped per RFC 6265).
-	if token := r.URL.Query().Get(sessionTokenParam); token != "" {
+	if token := r.URL.Query().Get("session_token"); token != "" {
 		return token, "", nil
 	}
 

proxy/internal/auth/tunnel_cache.go

@@ -3,20 +3,30 @@ package auth
 import (
 	"context"
 	"net/netip"
+	"os"
+	"strings"
 	"sync"
 	"time"
 
+	log "github.com/sirupsen/logrus"
 	"golang.org/x/sync/singleflight"
 
 	"github.com/netbirdio/netbird/proxy/internal/types"
 	"github.com/netbirdio/netbird/shared/management/proto"
 )
 
-// tunnelCacheTTL caps how long a positive ValidateTunnelPeer result is
-// reused before re-fetching from management. 5 minutes balances freshness
-// against management load on busy mesh networks.
+// tunnelCacheTTL is the default cap on how long a positive ValidateTunnelPeer
+// result is reused before re-fetching from management. 5 minutes balances
+// freshness against management load on busy mesh networks. Override it with
+// envTunnelCacheTTL when an account needs authorization changes to take effect
+// sooner (at the cost of more ValidateTunnelPeer RPCs).
 const tunnelCacheTTL = 300 * time.Second
 
+// envTunnelCacheTTL overrides tunnelCacheTTL. The value is a Go duration string
+// (e.g. "30s", "2m"); an unset, unparseable, or non-positive value keeps the
+// default.
+const envTunnelCacheTTL = "NB_PROXY_TUNNEL_CACHE_TTL"
+
 // tunnelCachePerAccount caps the number of cached identities per account.
 // Bounded eviction avoids memory growth in pathological cases (huge peer
 // roster, brief request bursts) while staying generous for normal use.
@@ -60,16 +70,35 @@ type accountBucket struct {
 	order []tunnelCacheKey
 }
 
-// newTunnelValidationCache constructs a cache with default TTL and bounds.
+// newTunnelValidationCache constructs a cache with the configured TTL
+// (envTunnelCacheTTL override or default) and default bounds.
 func newTunnelValidationCache() *tunnelValidationCache {
 	return &tunnelValidationCache{
 		entries: make(map[types.AccountID]*accountBucket),
-		ttl:     tunnelCacheTTL,
+		ttl:     tunnelCacheTTLFromEnv(),
 		maxSize: tunnelCachePerAccount,
 		now:     time.Now,
 	}
 }
 
+// tunnelCacheTTLFromEnv returns the tunnel-cache TTL, honoring the
+// envTunnelCacheTTL override. The override must be a positive Go duration
+// string (e.g. "30s", "2m"); anything unset, unparseable, or non-positive
+// falls back to tunnelCacheTTL.
+func tunnelCacheTTLFromEnv() time.Duration {
+	raw := strings.TrimSpace(os.Getenv(envTunnelCacheTTL))
+	if raw == "" {
+		return tunnelCacheTTL
+	}
+	d, err := time.ParseDuration(raw)
+	if err != nil || d <= 0 {
+		log.Warnf("ignoring invalid %s=%q (want a positive Go duration like 30s or 2m); using default %s",
+			envTunnelCacheTTL, raw, tunnelCacheTTL)
+		return tunnelCacheTTL
+	}
+	return d
+}
+
 // get returns a cached response for the key, or nil when missing or
 // expired. Expired entries are evicted lazily on read.
 func (c *tunnelValidationCache) get(key tunnelCacheKey) *proto.ValidateTunnelPeerResponse {

proxy/internal/auth/tunnel_cache_test.go

@@ -169,3 +169,32 @@ func TestTunnelCache_BoundedSizeEvictsOldest(t *testing.T) {
 	assert.NotNil(t, cache.get(keys[1]), "second-newest must remain cached")
 	assert.NotNil(t, cache.get(keys[2]), "newest must remain cached")
 }
+
+func TestTunnelCacheTTLFromEnv(t *testing.T) {
+	t.Run("unset uses default", func(t *testing.T) {
+		t.Setenv(envTunnelCacheTTL, "")
+		assert.Equal(t, tunnelCacheTTL, tunnelCacheTTLFromEnv())
+	})
+	t.Run("valid duration overrides", func(t *testing.T) {
+		t.Setenv(envTunnelCacheTTL, "45s")
+		assert.Equal(t, 45*time.Second, tunnelCacheTTLFromEnv())
+	})
+	t.Run("whitespace trimmed", func(t *testing.T) {
+		t.Setenv(envTunnelCacheTTL, "  2m ")
+		assert.Equal(t, 2*time.Minute, tunnelCacheTTLFromEnv())
+	})
+	t.Run("unparseable uses default", func(t *testing.T) {
+		t.Setenv(envTunnelCacheTTL, "nonsense")
+		assert.Equal(t, tunnelCacheTTL, tunnelCacheTTLFromEnv())
+	})
+	t.Run("non-positive uses default", func(t *testing.T) {
+		t.Setenv(envTunnelCacheTTL, "0s")
+		assert.Equal(t, tunnelCacheTTL, tunnelCacheTTLFromEnv())
+		t.Setenv(envTunnelCacheTTL, "-30s")
+		assert.Equal(t, tunnelCacheTTL, tunnelCacheTTLFromEnv())
+	})
+	t.Run("constructor honors override", func(t *testing.T) {
+		t.Setenv(envTunnelCacheTTL, "90s")
+		assert.Equal(t, 90*time.Second, newTunnelValidationCache().ttl)
+	})
+}

proxy/internal/auth/tunnel_lookup_test.go

@@ -44,7 +44,7 @@ func (s *stubSessionValidator) ValidateTunnelPeer(_ context.Context, in *proto.V
 func newTunnelMiddleware(t *testing.T, validator SessionValidator) *Middleware {
 	t.Helper()
 	mw := NewMiddleware(log.New(), validator, nil)
-	require.NoError(t, mw.AddDomain("svc.example", DomainSettings{AccountID: "acct-1", ServiceID: "svc-1"}))
+	require.NoError(t, mw.AddDomain("svc.example", nil, "", 0, "acct-1", "svc-1", nil, false))
 	return mw
 }
 
@@ -235,8 +235,8 @@ func TestForwardWithTunnelPeer_RoutesAccountIDIntoCacheKey(t *testing.T) {
 	}
 	mw := NewMiddleware(log.New(), validator, nil)
 
-	require.NoError(t, mw.AddDomain("svc-a.example", DomainSettings{AccountID: "acct-a", ServiceID: "svc-a"}))
-	require.NoError(t, mw.AddDomain("svc-b.example", DomainSettings{AccountID: "acct-b", ServiceID: "svc-b"}))
+	require.NoError(t, mw.AddDomain("svc-a.example", nil, "", 0, "acct-a", "svc-a", nil, false))
+	require.NoError(t, mw.AddDomain("svc-b.example", nil, "", 0, "acct-b", "svc-b", nil, false))
 
 	// The fast-path requires the inbound-listener marker on the context.
 	// The peerstore lookup itself is account-agnostic at this level
@@ -299,7 +299,7 @@ func TestForwardWithTunnelPeer_LocalLookupShortCircuitDoesNotPopulateCache(t *te
 
 func TestPrivateService_FailsClosedOnTunnelPeerFailure(t *testing.T) {
 	mw := NewMiddleware(log.New(), nil, nil)
-	require.NoError(t, mw.AddDomain("private.svc", DomainSettings{AccountID: "acct-1", ServiceID: "svc-1", Private: true}))
+	require.NoError(t, mw.AddDomain("private.svc", nil, "", 0, "acct-1", "svc-1", nil, true))
 
 	called := false
 	handler := mw.Protect(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
@@ -328,7 +328,7 @@ func TestPrivateService_ForwardsOnTunnelPeerSuccess(t *testing.T) {
 		},
 	}
 	mw := NewMiddleware(log.New(), validator, nil)
-	require.NoError(t, mw.AddDomain("private.svc", DomainSettings{AccountID: "acct-1", ServiceID: "svc-1", Private: true}))
+	require.NoError(t, mw.AddDomain("private.svc", nil, "", 0, "acct-1", "svc-1", nil, true))
 
 	called := false
 	handler := mw.Protect(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {

proxy/internal/llm/pricing/defaults_coverage_test.go

@@ -28,12 +28,12 @@ func TestDefaultTable_FirstPartyModelCoverage(t *testing.T) {
 			"ministral-8b-latest", "mistral-embed",
 		},
 		"anthropic": {
-			"claude-fable-5", "claude-opus-4-8", "claude-opus-4-7", "claude-opus-4-6",
+			"claude-fable-5", "claude-opus-5", "claude-opus-4-8", "claude-opus-4-7", "claude-opus-4-6",
 			"claude-opus-4-1", "claude-sonnet-4-6", "claude-sonnet-4-5", "claude-haiku-4-5",
 		},
 		// bedrock keys are the normalized ids the request parser emits.
 		"bedrock": {
-			"anthropic.claude-opus-4-8", "anthropic.claude-opus-4-7", "anthropic.claude-opus-4-6",
+			"anthropic.claude-opus-5", "anthropic.claude-opus-4-8", "anthropic.claude-opus-4-7", "anthropic.claude-opus-4-6",
 			"anthropic.claude-opus-4-1", "anthropic.claude-sonnet-4-6", "anthropic.claude-sonnet-4-5",
 			"anthropic.claude-haiku-4-5", "meta.llama3-3-70b-instruct",
 			"amazon.nova-pro", "amazon.nova-lite", "amazon.nova-micro", "amazon.nova-2-lite",

proxy/internal/llm/pricing/defaults_pricing.yaml

@@ -180,6 +180,11 @@ anthropic:
     output_per_1k: 0.050
     cache_read_per_1k: 0.001
     cache_creation_per_1k: 0.0125
+  claude-opus-5:
+    input_per_1k: 0.005
+    output_per_1k: 0.025
+    cache_read_per_1k: 0.0005
+    cache_creation_per_1k: 0.00625
   claude-opus-4-8:
     input_per_1k: 0.005
     output_per_1k: 0.025
@@ -236,6 +241,11 @@ bedrock:
   # eu.anthropic.claude-sonnet-4-5-20250929-v1:0 -> anthropic.claude-sonnet-4-5.
   # Anthropic-on-Bedrock keeps the additive cache buckets (read ≈0.1x input,
   # write ≈1.25x input); Nova / Llama report no cache, so cost is input+output.
+  anthropic.claude-opus-5:
+    input_per_1k: 0.005
+    output_per_1k: 0.025
+    cache_read_per_1k: 0.0005
+    cache_creation_per_1k: 0.00625
   anthropic.claude-opus-4-8:
     input_per_1k: 0.005
     output_per_1k: 0.025

proxy/internal/middleware/bodytap/request.go

@@ -21,8 +21,6 @@ import (
 	"strconv"
 	"strings"
 	"sync"
-
-	"github.com/netbirdio/netbird/proxy/internal/netutil"
 )
 
 // MaxRoutingScanBytes bounds how far ScanRoutingFields will read into a
@@ -36,6 +34,7 @@ const MaxRoutingScanBytes int64 = 32 << 20
 // metadata key by the chain when a request body is not surfaced.
 const (
 	BypassUpgradeHeader    = "upgrade_header"
+	BypassConnectionUpgrd  = "connection_upgrade"
 	BypassContentType      = "content_type_not_allowed"
 	BypassBudget           = "capture_budget_exhausted"
 	BypassNoConfig         = "no_capture_config"
@@ -126,13 +125,12 @@ func CaptureRequest(r *http.Request, cfg *Config, b Budget) (body []byte, trunca
 	if cfg.MaxRequestBytes <= 0 {
 		return nil, false, 0, BypassCapZero, release, nil
 	}
-	// The predicate has to be the forwarder's own: a looser one (either header
-	// on its own) skips capture for requests the forwarder still delivers to
-	// the upstream with their body intact, which hides them from every
-	// deny-capable middleware in the chain.
-	if netutil.IsUpgradeRequest(r.Header) {
+	if r.Header.Get("Upgrade") != "" {
 		return nil, false, 0, BypassUpgradeHeader, release, nil
 	}
+	if strings.EqualFold(r.Header.Get("Connection"), "upgrade") {
+		return nil, false, 0, BypassConnectionUpgrd, release, nil
+	}
 	if !contentTypeAllowed(r.Header.Get("Content-Type"), cfg.ContentTypes) {
 		return nil, false, 0, BypassContentType, release, nil
 	}

proxy/internal/middleware/builtin/llm_guardrail/factory.go

@@ -10,11 +10,15 @@ import (
 )
 
 // Config is the JSON-decoded shape accepted by the factory. The
-// runtime path consumes the normalised allowlist; raw config is not
+// runtime path consumes the normalised allowlists; raw config is not
 // retained beyond construction.
 type Config struct {
-	ModelAllowlist []string      `json:"model_allowlist"`
-	PromptCapture  PromptCapture `json:"prompt_capture"`
+	// ProviderAllowlists maps a resolved provider id (KeyLLMResolvedProviderID) to
+	// its model allowlist. A provider present is restricted to those models; one
+	// absent is unrestricted. Kept per-provider so one provider's list can't leak
+	// onto another.
+	ProviderAllowlists map[string][]string `json:"provider_allowlists,omitempty"`
+	PromptCapture      PromptCapture       `json:"prompt_capture"`
 }
 
 // PromptCapture toggles the optional prompt capture + redaction step
@@ -54,21 +58,28 @@ func isEmptyJSON(raw []byte) bool {
 	return false
 }
 
-// normaliseConfig lowercases and trims allowlist entries so the runtime
-// match is case-insensitive. Empty entries are dropped.
+// normaliseConfig lowercases and trims allowlist entries for case-insensitive
+// matching; empty entries drop. A provider whose entries all drop keeps an empty
+// (non-nil) list — "deny every model" — distinct from an absent provider
+// (unrestricted).
 func normaliseConfig(cfg Config) Config {
-	if len(cfg.ModelAllowlist) == 0 {
+	if len(cfg.ProviderAllowlists) == 0 {
+		cfg.ProviderAllowlists = nil
 		return cfg
 	}
-	cleaned := make([]string, 0, len(cfg.ModelAllowlist))
-	for _, entry := range cfg.ModelAllowlist {
-		n := normaliseModel(entry)
-		if n == "" {
-			continue
+	cleaned := make(map[string][]string, len(cfg.ProviderAllowlists))
+	for provider, models := range cfg.ProviderAllowlists {
+		list := make([]string, 0, len(models))
+		for _, entry := range models {
+			n := normaliseModel(entry)
+			if n == "" {
+				continue
+			}
+			list = append(list, n)
 		}
-		cleaned = append(cleaned, n)
+		cleaned[provider] = list
 	}
-	cfg.ModelAllowlist = cleaned
+	cfg.ProviderAllowlists = cleaned
 	return cfg
 }
 

proxy/internal/middleware/builtin/llm_guardrail/middleware.go

@@ -83,8 +83,9 @@ func (m *Middleware) MutationsSupported() bool { return false }
 // prompt capture only affects the metadata emitted alongside an allow.
 func (m *Middleware) Invoke(_ context.Context, in *middleware.Input) (*middleware.Output, error) {
 	model, modelPresent := lookupMetadata(in.Metadata, middleware.KeyLLMModel)
+	providerID, _ := lookupMetadata(in.Metadata, middleware.KeyLLMResolvedProviderID)
 
-	if denial := m.evaluateAllowlist(model, modelPresent); denial != nil {
+	if denial := m.evaluateAllowlist(providerID, model, modelPresent); denial != nil {
 		return denial, nil
 	}
 
@@ -110,20 +111,32 @@ func (m *Middleware) Invoke(_ context.Context, in *middleware.Input) (*middlewar
 // is a no-op.
 func (m *Middleware) Close() error { return nil }
 
-// evaluateAllowlist returns a deny Output when the configured allowlist
-// rejects the model. A nil return means the request should proceed.
-func (m *Middleware) evaluateAllowlist(model string, modelPresent bool) *middleware.Output {
-	if len(m.cfg.ModelAllowlist) == 0 {
+// evaluateAllowlist denies when the resolved provider's allowlist rejects the
+// model; nil means proceed. Scoped to the provider llm_router resolved, so an
+// unrestricted provider (absent from config) is never caught by another's list.
+func (m *Middleware) evaluateAllowlist(providerID, model string, modelPresent bool) *middleware.Output {
+	if len(m.cfg.ProviderAllowlists) == 0 {
 		return nil
 	}
-	// Fail closed: with an allowlist configured, a request whose model the
-	// upstream parser could not extract (absent or empty) must be denied rather
-	// than allowed. This is what enforces the allowlist for URL/path-routed
-	// providers (Bedrock, Vertex, ...) whose model lives outside the JSON body.
+	// Restrictions exist but the resolved provider is unknown, so we can't tell
+	// if this request targets a restricted provider — fail closed. llm_router
+	// normally stamps the provider first, so this is a defensive guard.
+	if providerID == "" {
+		return denyModel("", denyCodeModelUnknown, denyMessageModelUnknown, denyReasonModelUnknown)
+	}
+	allowlist, restricted := m.cfg.ProviderAllowlists[providerID]
+	if !restricted {
+		// This provider has no allowlist (some authorising policy left it
+		// unrestricted); management owns any per-policy/group decision.
+		return nil
+	}
+	// Fail closed: with an allowlist in effect for this provider, a request whose
+	// model the parser couldn't extract (absent/empty) is denied. This enforces
+	// the allowlist for path-routed providers (Bedrock, Vertex) with no body model.
 	if !modelPresent || normaliseModel(model) == "" {
 		return denyModel("", denyCodeModelUnknown, denyMessageModelUnknown, denyReasonModelUnknown)
 	}
-	if m.modelInAllowlist(model) {
+	if modelInAllowlist(allowlist, model) {
 		return nil
 	}
 	return denyModel(model, denyCodeModel, denyMessageModel, denyReasonModel)
@@ -151,14 +164,15 @@ func denyModel(model, code, message, reason string) *middleware.Output {
 	}
 }
 
-// modelInAllowlist reports whether the model matches any allowlist
-// entry under the case-insensitive, trim-tolerant comparison rule.
-func (m *Middleware) modelInAllowlist(model string) bool {
+// modelInAllowlist reports whether the model matches any entry in the supplied
+// (already-normalised) allowlist under the case-insensitive, trim-tolerant
+// comparison rule.
+func modelInAllowlist(allowlist []string, model string) bool {
 	normalised := normaliseModel(model)
 	if normalised == "" {
 		return false
 	}
-	for _, allowed := range m.cfg.ModelAllowlist {
+	for _, allowed := range allowlist {
 		if allowed == normalised {
 			return true
 		}

proxy/internal/middleware/builtin/llm_guardrail/middleware_test.go

@@ -26,6 +26,25 @@ func newInput(meta ...middleware.KV) *middleware.Input {
 	return &middleware.Input{Slot: middleware.SlotOnRequest, Metadata: meta}
 }
 
+const (
+	testProvider  = "prov-1"
+	otherProvider = "prov-2"
+)
+
+// providerCfg builds a Config restricting testProvider to the given models.
+func providerCfg(models ...string) Config {
+	return Config{ProviderAllowlists: map[string][]string{testProvider: models}}
+}
+
+// newInputProvider builds an input that carries a resolved provider id (as
+// llm_router would stamp) plus any extra metadata.
+func newInputProvider(provider string, meta ...middleware.KV) *middleware.Input {
+	all := make([]middleware.KV, 0, len(meta)+1)
+	all = append(all, middleware.KV{Key: middleware.KeyLLMResolvedProviderID, Value: provider})
+	all = append(all, meta...)
+	return &middleware.Input{Slot: middleware.SlotOnRequest, Metadata: all}
+}
+
 func TestMiddlewareIdentity(t *testing.T) {
 	mw := New(Config{})
 	assert.Equal(t, ID, mw.ID(), "middleware ID must be llm_guardrail")
@@ -47,12 +66,12 @@ func TestMiddlewareIdentity(t *testing.T) {
 
 func TestAllowlistEmptyAllowsAnyModel(t *testing.T) {
 	mw := New(Config{})
-	out, err := mw.Invoke(context.Background(), newInput(
+	out, err := mw.Invoke(context.Background(), newInputProvider(testProvider,
 		middleware.KV{Key: middleware.KeyLLMModel, Value: "gpt-4o"},
 	))
 	require.NoError(t, err)
 	require.NotNil(t, out)
-	assert.Equal(t, middleware.DecisionAllow, out.Decision, "empty allowlist must allow any model")
+	assert.Equal(t, middleware.DecisionAllow, out.Decision, "no provider allowlists must allow any model")
 	v, ok := metaValue(t, out.Metadata, middleware.KeyLLMPolicyDecision)
 	require.True(t, ok, "decision metadata must be emitted")
 	assert.Equal(t, "allow", v, "decision must be allow")
@@ -62,8 +81,8 @@ func TestAllowlistEmptyAllowsAnyModel(t *testing.T) {
 }
 
 func TestAllowlistMatchAllows(t *testing.T) {
-	mw := New(Config{ModelAllowlist: []string{"gpt-4o", "claude-opus-4"}})
-	out, err := mw.Invoke(context.Background(), newInput(
+	mw := New(providerCfg("gpt-4o", "claude-opus-4"))
+	out, err := mw.Invoke(context.Background(), newInputProvider(testProvider,
 		middleware.KV{Key: middleware.KeyLLMModel, Value: "gpt-4o"},
 	))
 	require.NoError(t, err)
@@ -71,8 +90,8 @@ func TestAllowlistMatchAllows(t *testing.T) {
 }
 
 func TestAllowlistMissDenies(t *testing.T) {
-	mw := New(Config{ModelAllowlist: []string{"gpt-4o"}})
-	out, err := mw.Invoke(context.Background(), newInput(
+	mw := New(providerCfg("gpt-4o"))
+	out, err := mw.Invoke(context.Background(), newInputProvider(testProvider,
 		middleware.KV{Key: middleware.KeyLLMModel, Value: "claude-opus-4"},
 	))
 	require.NoError(t, err)
@@ -91,10 +110,10 @@ func TestAllowlistMissDenies(t *testing.T) {
 }
 
 func TestAllowlistCaseInsensitive(t *testing.T) {
-	mw := New(Config{ModelAllowlist: []string{"  GPT-4o  ", "Claude-OPUS-4"}})
+	mw := New(providerCfg("  GPT-4o  ", "Claude-OPUS-4"))
 	cases := []string{"gpt-4o", "GPT-4O", "  claude-opus-4 "}
 	for _, model := range cases {
-		out, err := mw.Invoke(context.Background(), newInput(
+		out, err := mw.Invoke(context.Background(), newInputProvider(testProvider,
 			middleware.KV{Key: middleware.KeyLLMModel, Value: model},
 		))
 		require.NoError(t, err)
@@ -103,14 +122,15 @@ func TestAllowlistCaseInsensitive(t *testing.T) {
 }
 
 func TestAllowlistMissingModelKeyDenies(t *testing.T) {
-	// Fail closed: with an allowlist configured, a request whose model the
-	// parser could not extract (URL/path-routed providers such as Bedrock or
-	// Vertex whose shape wasn't recognised) must be denied, not allowed.
-	mw := New(Config{ModelAllowlist: []string{"gpt-4o"}})
-	out, err := mw.Invoke(context.Background(), newInput())
+	// Fail closed: with an allowlist in effect for the resolved provider, a
+	// request whose model the parser could not extract (URL/path-routed
+	// providers such as Bedrock or Vertex whose shape wasn't recognised) must be
+	// denied, not allowed.
+	mw := New(providerCfg("gpt-4o"))
+	out, err := mw.Invoke(context.Background(), newInputProvider(testProvider))
 	require.NoError(t, err)
 	require.NotNil(t, out)
-	assert.Equal(t, middleware.DecisionDeny, out.Decision, "absent model must be denied when an allowlist is set")
+	assert.Equal(t, middleware.DecisionDeny, out.Decision, "absent model must be denied when the provider is restricted")
 	assert.Equal(t, 403, out.DenyStatus, "deny status must be 403")
 	require.NotNil(t, out.DenyReason, "deny reason must be populated")
 	assert.Equal(t, "llm_policy.model_unknown", out.DenyReason.Code, "deny code must be model_unknown")
@@ -122,26 +142,101 @@ func TestAllowlistMissingModelKeyDenies(t *testing.T) {
 
 func TestAllowlistEmptyModelValueDenies(t *testing.T) {
 	// A present-but-empty model is as undeterminable as an absent one.
-	mw := New(Config{ModelAllowlist: []string{"gpt-4o"}})
-	out, err := mw.Invoke(context.Background(), newInput(
+	mw := New(providerCfg("gpt-4o"))
+	out, err := mw.Invoke(context.Background(), newInputProvider(testProvider,
 		middleware.KV{Key: middleware.KeyLLMModel, Value: "   "},
 	))
 	require.NoError(t, err)
 	require.NotNil(t, out)
-	assert.Equal(t, middleware.DecisionDeny, out.Decision, "empty model must be denied when an allowlist is set")
+	assert.Equal(t, middleware.DecisionDeny, out.Decision, "empty model must be denied when the provider is restricted")
 	require.NotNil(t, out.DenyReason, "deny reason must be populated")
 	assert.Equal(t, "llm_policy.model_unknown", out.DenyReason.Code, "deny code must be model_unknown")
 }
 
 func TestAllowlistEmptyListAllowsMissingModel(t *testing.T) {
-	// Without an allowlist there is nothing to enforce, so a missing model is
-	// still allowed — the fail-closed rule only applies when a list is set.
+	// Without any provider allowlists there is nothing to enforce, so a missing
+	// model is still allowed — the fail-closed rule only applies when a
+	// restriction is in effect.
 	mw := New(Config{})
 	out, err := mw.Invoke(context.Background(), newInput())
 	require.NoError(t, err)
 	assert.Equal(t, middleware.DecisionAllow, out.Decision, "no allowlist must allow even without a model")
 }
 
+func TestUnrestrictedProviderAllowsAnyModel(t *testing.T) {
+	// The request resolved to otherProvider, which has no allowlist, so its
+	// traffic must not be caught by testProvider's restriction — the
+	// cross-provider-leak / false-deny guard.
+	mw := New(providerCfg("gpt-4o"))
+	out, err := mw.Invoke(context.Background(), newInputProvider(otherProvider,
+		middleware.KV{Key: middleware.KeyLLMModel, Value: "claude-opus-4"},
+	))
+	require.NoError(t, err)
+	assert.Equal(t, middleware.DecisionAllow, out.Decision, "an unrestricted provider must not inherit another provider's allowlist")
+}
+
+func TestPerProviderAllowlistsAreIsolated(t *testing.T) {
+	// gpt-4o is allowed only on testProvider; claude-opus-4 only on
+	// otherProvider. A model allowlisted for one provider must not be usable on
+	// the other — the fail-closed layer never unions allowlists across providers.
+	mw := New(Config{ProviderAllowlists: map[string][]string{
+		testProvider:  {"gpt-4o"},
+		otherProvider: {"claude-opus-4"},
+	}})
+	out, err := mw.Invoke(context.Background(), newInputProvider(testProvider,
+		middleware.KV{Key: middleware.KeyLLMModel, Value: "claude-opus-4"},
+	))
+	require.NoError(t, err)
+	assert.Equal(t, middleware.DecisionDeny, out.Decision, "claude-opus-4 is allowed only on otherProvider, not testProvider")
+	require.NotNil(t, out.DenyReason)
+	assert.Equal(t, "llm_policy.model_blocked", out.DenyReason.Code, "cross-provider model must be blocked, not model_unknown")
+}
+
+func TestRestrictionsButNoResolvedProviderFailsClosed(t *testing.T) {
+	// Restrictions exist for the account but the resolved provider id is absent,
+	// so the request cannot be scoped to a provider. Fail closed rather than
+	// wave it through.
+	mw := New(providerCfg("gpt-4o"))
+	out, err := mw.Invoke(context.Background(), newInput(
+		middleware.KV{Key: middleware.KeyLLMModel, Value: "gpt-4o"},
+	))
+	require.NoError(t, err)
+	require.NotNil(t, out)
+	assert.Equal(t, middleware.DecisionDeny, out.Decision, "missing resolved provider must fail closed when restrictions exist")
+	require.NotNil(t, out.DenyReason)
+	assert.Equal(t, "llm_policy.model_unknown", out.DenyReason.Code, "deny code must be model_unknown")
+}
+
+func TestEnabledButEmptyAllowlistDeniesEveryModel(t *testing.T) {
+	// An allowlist-enabled provider with zero models is distinct from an
+	// unrestricted (absent) provider: it must deny every model.
+	mw := New(providerCfg())
+	out, err := mw.Invoke(context.Background(), newInputProvider(testProvider,
+		middleware.KV{Key: middleware.KeyLLMModel, Value: "gpt-4o"},
+	))
+	require.NoError(t, err)
+	require.NotNil(t, out)
+	assert.Equal(t, middleware.DecisionDeny, out.Decision, "an enabled-but-empty allowlist must deny every model")
+	require.NotNil(t, out.DenyReason)
+	assert.Equal(t, "llm_policy.model_blocked", out.DenyReason.Code, "deny code must be model_blocked, not model_unknown")
+}
+
+func TestFactoryAllEmptyEntriesDenyEveryModel(t *testing.T) {
+	// All the provider's entries are blank; they collapse to a non-nil empty
+	// list (deny everything for that provider), not "no restriction".
+	raw := []byte(`{"provider_allowlists":{"prov-1":["","   "]}}`)
+	mw, err := Factory{}.New(raw)
+	require.NoError(t, err)
+	out, err := mw.Invoke(context.Background(), newInputProvider(testProvider,
+		middleware.KV{Key: middleware.KeyLLMModel, Value: "gpt-4o"},
+	))
+	require.NoError(t, err)
+	require.NotNil(t, out)
+	assert.Equal(t, middleware.DecisionDeny, out.Decision, "all-blank allowlist entries must still restrict the provider")
+	require.NotNil(t, out.DenyReason)
+	assert.Equal(t, "llm_policy.model_blocked", out.DenyReason.Code, "deny code must be model_blocked")
+}
+
 func TestPromptCaptureDisabledEmitsNoPrompt(t *testing.T) {
 	mw := New(Config{})
 	out, err := mw.Invoke(context.Background(), newInput(
@@ -217,8 +312,8 @@ func TestFactoryAcceptsZeroConfigs(t *testing.T) {
 
 func TestFactoryDecodesValidConfig(t *testing.T) {
 	cfg := Config{
-		ModelAllowlist: []string{"gpt-4o"},
-		PromptCapture:  PromptCapture{Enabled: true, RedactPii: true},
+		ProviderAllowlists: map[string][]string{testProvider: {"gpt-4o"}},
+		PromptCapture:      PromptCapture{Enabled: true, RedactPii: true},
 	}
 	raw, err := json.Marshal(cfg)
 	require.NoError(t, err, "marshalling test config must succeed")
@@ -234,15 +329,15 @@ func TestFactoryRejectsMalformedJSON(t *testing.T) {
 }
 
 func TestFactoryNormalisesAllowlist(t *testing.T) {
-	raw := []byte(`{"model_allowlist":["  GPT-4o  ","",""," Claude-3 "]}`)
+	raw := []byte(`{"provider_allowlists":{"prov-1":["  GPT-4o  ","",""," Claude-3 "]}}`)
 	mw, err := Factory{}.New(raw)
 	require.NoError(t, err)
-	out, err := mw.Invoke(context.Background(), newInput(
+	out, err := mw.Invoke(context.Background(), newInputProvider(testProvider,
 		middleware.KV{Key: middleware.KeyLLMModel, Value: "gpt-4o"},
 	))
 	require.NoError(t, err)
 	assert.Equal(t, middleware.DecisionAllow, out.Decision, "factory must lowercase + trim allowlist entries")
-	out2, err := mw.Invoke(context.Background(), newInput(
+	out2, err := mw.Invoke(context.Background(), newInputProvider(testProvider,
 		middleware.KV{Key: middleware.KeyLLMModel, Value: "claude-3"},
 	))
 	require.NoError(t, err)

proxy/internal/middleware/builtin/llm_limit_check/middleware.go

@@ -175,7 +175,7 @@ func denyFromManagement(resp *proto.CheckLLMPolicyLimitsResponse) *middleware.Ou
 		DenyStatus: 403,
 		DenyReason: &middleware.DenyReason{
 			Code:    code,
-			Message: "LLM policy limit exceeded",
+			Message: denyMessageForCode(code),
 		},
 		Metadata: []middleware.KV{
 			{Key: middleware.KeyLLMPolicyDecision, Value: "deny"},
@@ -184,6 +184,21 @@ func denyFromManagement(resp *proto.CheckLLMPolicyLimitsResponse) *middleware.Ou
 	}
 }
 
+// denyMessageForCode maps a management deny code to a public message.
+// Model-allowlist rejections get a model-specific message matching the
+// local guardrail; everything else keeps the generic quota wording. The
+// message stays generic so it never leaks internal quota detail.
+func denyMessageForCode(code string) string {
+	switch code {
+	case "llm_policy.model_blocked":
+		return "model is not in the policy allowlist"
+	case "llm_policy.model_unknown":
+		return "request model could not be determined for the policy allowlist"
+	default:
+		return "LLM policy limit exceeded"
+	}
+}
+
 // lookupKV returns the value associated with key, or the empty
 // string when absent.
 func lookupKV(kvs []middleware.KV, key string) string {

proxy/internal/middleware/builtin/llm_limit_check/middleware_test.go

@@ -115,6 +115,46 @@ func TestInvoke_DenyConvertsToProxyDeny(t *testing.T) {
 	assert.NotContains(t, out.DenyReason.Message, "1000", "internal cap numbers must not reach the caller")
 }
 
+// TestInvoke_ModelDenyMessages proves a model-allowlist rejection gets a
+// model-specific public message rather than the generic quota wording, so a
+// blocked or undetermined model reads consistently with the local guardrail.
+func TestInvoke_ModelDenyMessages(t *testing.T) {
+	cases := []struct {
+		name    string
+		code    string
+		message string
+	}{
+		{"blocked", "llm_policy.model_blocked", "model is not in the policy allowlist"},
+		{"unknown", "llm_policy.model_unknown", "request model could not be determined for the policy allowlist"},
+	}
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			mgmt := &fakeMgmt{
+				checkResp: &proto.CheckLLMPolicyLimitsResponse{
+					Decision: "deny",
+					DenyCode: tc.code,
+				},
+			}
+			m := New(mgmt, nil)
+
+			out := runInvoke(t, m, &middleware.Input{
+				AccountID:  "acc-1",
+				UserGroups: []string{"grp-engineers"},
+				Metadata: []middleware.KV{
+					{Key: middleware.KeyLLMResolvedProviderID, Value: "prov-1"},
+					{Key: middleware.KeyLLMModel, Value: "some-model"},
+				},
+			})
+
+			assert.Equal(t, middleware.DecisionDeny, out.Decision)
+			require.NotNil(t, out.DenyReason, "deny envelope must carry a reason payload")
+			assert.Equal(t, tc.code, out.DenyReason.Code, "canonical deny code surfaces to the caller")
+			assert.Equal(t, tc.message, out.DenyReason.Message,
+				"model denials must use a model-specific message, matching the local guardrail")
+		})
+	}
+}
+
 // TestInvoke_NoMgmtClientPassesThrough proves the partial-wiring
 // safety: a middleware constructed without a management client
 // allows every request without attribution. This makes a half-set-up

proxy/internal/middleware/builtin/llm_request_parser/guardrail_allowlist_test.go

@@ -25,10 +25,17 @@ func runParserGuardrail(t *testing.T, url string, body []byte, allowlist []strin
 	})
 	require.NoError(t, err, "parser must not error")
 
-	guard := llm_guardrail.New(llm_guardrail.Config{ModelAllowlist: allowlist})
+	const providerID = "prov-under-test"
+	guard := llm_guardrail.New(llm_guardrail.Config{
+		ProviderAllowlists: map[string][]string{providerID: allowlist},
+	})
+	// The real chain has llm_router stamp the resolved provider id before the
+	// guardrail runs; the parser doesn't, so add it here so the guardrail can
+	// scope the allowlist to this provider.
+	meta := append([]middleware.KV{{Key: middleware.KeyLLMResolvedProviderID, Value: providerID}}, parsed.Metadata...)
 	out, err := guard.Invoke(context.Background(), &middleware.Input{
 		Slot:     middleware.SlotOnRequest,
-		Metadata: parsed.Metadata,
+		Metadata: meta,
 	})
 	require.NoError(t, err, "guardrail must not error")
 	require.NotNil(t, out, "guardrail must return an output")

proxy/internal/netutil/upgrade.go

@@ -1,23 +0,0 @@
-package netutil
-
-import (
-	"net/http"
-
-	"golang.org/x/net/http/httpguts"
-)
-
-// IsUpgradeRequest reports whether r is a protocol-upgrade request, using the
-// same predicate httputil.ReverseProxy applies when it decides to hand the
-// connection over instead of proxying normally.
-//
-// Matching the forwarder exactly matters for anything that inspects a request
-// before it is proxied: a looser test (an Upgrade header on its own, say) marks
-// a request as an upgrade and skips inspection, while the forwarder still
-// delivers it to the backend as an ordinary request with its body intact. That
-// gap is a body-inspection bypass reachable by adding one header.
-func IsUpgradeRequest(h http.Header) bool {
-	if !httpguts.HeaderValuesContainsToken(h["Connection"], "Upgrade") {
-		return false
-	}
-	return h.Get("Upgrade") != ""
-}

proxy/internal/restrict/restrict.go

@@ -50,37 +50,6 @@ const (
 	CrowdSecObserve CrowdSecMode = "observe"
 )
 
-// AppSecMode is the per-service CrowdSec AppSec (WAF) enforcement mode.
-type AppSecMode string
-
-const (
-	// AppSecOff disables request inspection.
-	AppSecOff AppSecMode = ""
-	// AppSecEnforce blocks requests the engine flags, and fails closed when the
-	// engine is unreachable.
-	AppSecEnforce AppSecMode = "enforce"
-	// AppSecObserve records the verdict without blocking.
-	AppSecObserve AppSecMode = "observe"
-)
-
-// ParseAppSecMode maps a wire value to an AppSecMode. Unrecognized values map
-// to AppSecOff so a typo never turns inspection into an unintended block.
-func ParseAppSecMode(s string) AppSecMode {
-	switch AppSecMode(s) {
-	case AppSecEnforce:
-		return AppSecEnforce
-	case AppSecObserve:
-		return AppSecObserve
-	default:
-		return AppSecOff
-	}
-}
-
-// Enabled reports whether the mode asks for request inspection.
-func (m AppSecMode) Enabled() bool {
-	return m == AppSecEnforce || m == AppSecObserve
-}
-
 // Filter evaluates IP restrictions. CIDR checks are performed first
 // (cheap), followed by country lookups (more expensive) only when needed.
 type Filter struct {
@@ -177,13 +146,6 @@ const (
 	// DenyCrowdSecUnavailable indicates enforce mode but the bouncer has not
 	// completed its initial sync.
 	DenyCrowdSecUnavailable
-	// DenyAppSecBan indicates a CrowdSec AppSec "ban" remediation.
-	DenyAppSecBan
-	// DenyAppSecCaptcha indicates a CrowdSec AppSec "captcha" remediation.
-	DenyAppSecCaptcha
-	// DenyAppSecUnavailable indicates enforce mode but the AppSec engine could
-	// not produce a verdict (unreachable, timed out, or it rejected the call).
-	DenyAppSecUnavailable
 )
 
 // String returns the deny reason string matching the HTTP auth mechanism names.
@@ -205,12 +167,6 @@ func (v Verdict) String() string {
 		return "crowdsec_throttle"
 	case DenyCrowdSecUnavailable:
 		return "crowdsec_unavailable"
-	case DenyAppSecBan:
-		return "appsec_ban"
-	case DenyAppSecCaptcha:
-		return "appsec_captcha"
-	case DenyAppSecUnavailable:
-		return "appsec_unavailable"
 	default:
 		return "unknown"
 	}
@@ -226,16 +182,6 @@ func (v Verdict) IsCrowdSec() bool {
 	}
 }
 
-// IsAppSec returns true when the verdict originates from an AppSec inspection.
-func (v Verdict) IsAppSec() bool {
-	switch v {
-	case DenyAppSecBan, DenyAppSecCaptcha, DenyAppSecUnavailable:
-		return true
-	default:
-		return false
-	}
-}
-
 // IsObserveOnly returns true when v is a CrowdSec verdict and the filter is in
 // observe mode. Callers should log the verdict but not block the request.
 func (f *Filter) IsObserveOnly(v Verdict) bool {

proxy/lifecycle.go

@@ -126,23 +126,6 @@ type Config struct {
 	// CrowdSecAPIKey is the CrowdSec bouncer API key. Empty disables
 	// CrowdSec.
 	CrowdSecAPIKey string
-	// CrowdSecAppSecURL is the CrowdSec AppSec (WAF) endpoint. Empty disables
-	// HTTP request inspection.
-	CrowdSecAppSecURL string
-	// CrowdSecAppSecTimeout bounds a single AppSec inspection call. Zero falls
-	// back to the internal default.
-	CrowdSecAppSecTimeout time.Duration
-	// CrowdSecAppSecMaxBodyBytes caps the request body mirrored to AppSec.
-	// Zero falls back to the internal default; negative forwards no body.
-	CrowdSecAppSecMaxBodyBytes int64
-	// CrowdSecAppSecMaxConcurrent bounds AppSec inspections in flight toward
-	// the engine. Zero falls back to the internal default; negative removes
-	// the bound.
-	CrowdSecAppSecMaxConcurrent int
-	// MiddlewareCaptureBudgetBytes bounds the total request-body buffering in
-	// flight across the proxy, shared by AppSec inspection and the
-	// agent-network capture. Zero falls back to the internal default.
-	MiddlewareCaptureBudgetBytes int64
 }
 
 // New builds a Server from cfg without performing any I/O. No goroutines
@@ -152,47 +135,42 @@ type Config struct {
 // directly) byte-for-byte equivalent.
 func New(ctx context.Context, cfg Config) *Server {
 	return &Server{
-		ctx:                          ctx,
-		ListenAddr:                   cfg.ListenAddr,
-		ID:                           cfg.ID,
-		Logger:                       cfg.Logger,
-		Version:                      cfg.Version,
-		ProxyURL:                     cfg.ProxyURL,
-		ManagementAddress:            cfg.ManagementAddress,
-		ProxyToken:                   cfg.ProxyToken,
-		CertificateDirectory:         cfg.CertificateDirectory,
-		CertificateFile:              cfg.CertificateFile,
-		CertificateKeyFile:           cfg.CertificateKeyFile,
-		GenerateACMECertificates:     cfg.GenerateACMECertificates,
-		ACMEChallengeAddress:         cfg.ACMEChallengeAddress,
-		ACMEDirectory:                cfg.ACMEDirectory,
-		ACMEEABKID:                   cfg.ACMEEABKID,
-		ACMEEABHMACKey:               cfg.ACMEEABHMACKey,
-		ACMEChallengeType:            cfg.ACMEChallengeType,
-		CertLockMethod:               cfg.CertLockMethod,
-		WildcardCertDir:              cfg.WildcardCertDir,
-		DebugEndpointEnabled:         cfg.DebugEndpointEnabled,
-		DebugEndpointAddress:         cfg.DebugEndpointAddress,
-		HealthAddress:                cfg.HealthAddr,
-		ForwardedProto:               cfg.ForwardedProto,
-		TrustedProxies:               cfg.TrustedProxies,
-		WireguardPort:                cfg.WireguardPort,
-		ProxyProtocol:                cfg.ProxyProtocol,
-		PreSharedKey:                 cfg.PreSharedKey,
-		Performance:                  cfg.Performance,
-		SupportsCustomPorts:          cfg.SupportsCustomPorts,
-		RequireSubdomain:             cfg.RequireSubdomain,
-		Private:                      cfg.Private,
-		MaxDialTimeout:               cfg.MaxDialTimeout,
-		MaxSessionIdleTimeout:        cfg.MaxSessionIdleTimeout,
-		MappingBatchWatchdog:         cfg.MappingBatchWatchdog,
-		GeoDataDir:                   cfg.GeoDataDir,
-		CrowdSecAPIURL:               cfg.CrowdSecAPIURL,
-		CrowdSecAPIKey:               cfg.CrowdSecAPIKey,
-		CrowdSecAppSecURL:            cfg.CrowdSecAppSecURL,
-		CrowdSecAppSecTimeout:        cfg.CrowdSecAppSecTimeout,
-		CrowdSecAppSecMaxBodyBytes:   cfg.CrowdSecAppSecMaxBodyBytes,
-		CrowdSecAppSecMaxConcurrent:  cfg.CrowdSecAppSecMaxConcurrent,
-		MiddlewareCaptureBudgetBytes: cfg.MiddlewareCaptureBudgetBytes,
+		ctx:                      ctx,
+		ListenAddr:               cfg.ListenAddr,
+		ID:                       cfg.ID,
+		Logger:                   cfg.Logger,
+		Version:                  cfg.Version,
+		ProxyURL:                 cfg.ProxyURL,
+		ManagementAddress:        cfg.ManagementAddress,
+		ProxyToken:               cfg.ProxyToken,
+		CertificateDirectory:     cfg.CertificateDirectory,
+		CertificateFile:          cfg.CertificateFile,
+		CertificateKeyFile:       cfg.CertificateKeyFile,
+		GenerateACMECertificates: cfg.GenerateACMECertificates,
+		ACMEChallengeAddress:     cfg.ACMEChallengeAddress,
+		ACMEDirectory:            cfg.ACMEDirectory,
+		ACMEEABKID:               cfg.ACMEEABKID,
+		ACMEEABHMACKey:           cfg.ACMEEABHMACKey,
+		ACMEChallengeType:        cfg.ACMEChallengeType,
+		CertLockMethod:           cfg.CertLockMethod,
+		WildcardCertDir:          cfg.WildcardCertDir,
+		DebugEndpointEnabled:     cfg.DebugEndpointEnabled,
+		DebugEndpointAddress:     cfg.DebugEndpointAddress,
+		HealthAddress:            cfg.HealthAddr,
+		ForwardedProto:           cfg.ForwardedProto,
+		TrustedProxies:           cfg.TrustedProxies,
+		WireguardPort:            cfg.WireguardPort,
+		ProxyProtocol:            cfg.ProxyProtocol,
+		PreSharedKey:             cfg.PreSharedKey,
+		Performance:              cfg.Performance,
+		SupportsCustomPorts:      cfg.SupportsCustomPorts,
+		RequireSubdomain:         cfg.RequireSubdomain,
+		Private:                  cfg.Private,
+		MaxDialTimeout:           cfg.MaxDialTimeout,
+		MaxSessionIdleTimeout:    cfg.MaxSessionIdleTimeout,
+		MappingBatchWatchdog:     cfg.MappingBatchWatchdog,
+		GeoDataDir:               cfg.GeoDataDir,
+		CrowdSecAPIURL:           cfg.CrowdSecAPIURL,
+		CrowdSecAPIKey:           cfg.CrowdSecAPIKey,
 	}
 }

proxy/lifecycle_test.go

@@ -1,45 +0,0 @@
-package proxy
-
-import (
-	"context"
-	"testing"
-	"time"
-
-	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
-)
-
-// New maps Config onto Server field by field, and a field left out of that
-// literal still compiles: the knob is simply parsed and then dropped, so an
-// operator setting it sees the default with no error anywhere. These assertions
-// are the only thing standing between a new setting and that silent no-op.
-func TestNew_ForwardsOperatorTuning(t *testing.T) {
-	cfg := Config{
-		ManagementAddress:            "http://localhost:8080",
-		ProxyToken:                   "token",
-		CrowdSecAPIURL:               "http://crowdsec:8080/",
-		CrowdSecAPIKey:               "key",
-		CrowdSecAppSecURL:            "http://crowdsec:7422/",
-		CrowdSecAppSecTimeout:        321 * time.Millisecond,
-		CrowdSecAppSecMaxBodyBytes:   4321,
-		CrowdSecAppSecMaxConcurrent:  17,
-		MiddlewareCaptureBudgetBytes: 5 << 20,
-		MaxDialTimeout:               7 * time.Second,
-		MaxSessionIdleTimeout:        11 * time.Second,
-		GeoDataDir:                   "/var/lib/geo",
-	}
-
-	srv := New(context.Background(), cfg)
-	require.NotNil(t, srv)
-
-	assert.Equal(t, cfg.CrowdSecAPIURL, srv.CrowdSecAPIURL)
-	assert.Equal(t, cfg.CrowdSecAPIKey, srv.CrowdSecAPIKey)
-	assert.Equal(t, cfg.CrowdSecAppSecURL, srv.CrowdSecAppSecURL)
-	assert.Equal(t, cfg.CrowdSecAppSecTimeout, srv.CrowdSecAppSecTimeout)
-	assert.Equal(t, cfg.CrowdSecAppSecMaxBodyBytes, srv.CrowdSecAppSecMaxBodyBytes)
-	assert.Equal(t, cfg.CrowdSecAppSecMaxConcurrent, srv.CrowdSecAppSecMaxConcurrent)
-	assert.Equal(t, cfg.MiddlewareCaptureBudgetBytes, srv.MiddlewareCaptureBudgetBytes)
-	assert.Equal(t, cfg.MaxDialTimeout, srv.MaxDialTimeout)
-	assert.Equal(t, cfg.MaxSessionIdleTimeout, srv.MaxSessionIdleTimeout)
-	assert.Equal(t, cfg.GeoDataDir, srv.GeoDataDir)
-}

proxy/management_integration_test.go

@@ -240,10 +240,6 @@ func (m *testProxyManager) ClusterSupportsCrowdSec(_ context.Context, _ string)
 	return nil
 }
 
-func (m *testProxyManager) ClusterSupportsAppSec(_ context.Context, _ string) *bool {
-	return nil
-}
-
 func (m *testProxyManager) ClusterSupportsPrivate(_ context.Context, _ string) *bool {
 	return nil
 }
@@ -566,11 +562,16 @@ func TestIntegration_ProxyConnection_ReconnectDoesNotDuplicateState(t *testing.T
 				addMappingCalls.Add(1)
 
 				// Apply to real auth middleware (idempotent)
-				err := authMw.AddDomain(mapping.GetDomain(), auth.DomainSettings{
-					AccountID: proxytypes.AccountID(mapping.GetAccountId()),
-					ServiceID: proxytypes.ServiceID(mapping.GetId()),
-					Private:   mapping.GetPrivate(),
-				})
+				err := authMw.AddDomain(
+					mapping.GetDomain(),
+					nil,
+					"",
+					0,
+					proxytypes.AccountID(mapping.GetAccountId()),
+					proxytypes.ServiceID(mapping.GetId()),
+					nil,
+					mapping.GetPrivate(),
+				)
 				require.NoError(t, err)
 
 				// Apply to real proxy (idempotent)

proxy/server.go

@@ -45,7 +45,6 @@ import (
 	"github.com/netbirdio/netbird/client/embed"
 	"github.com/netbirdio/netbird/proxy/internal/accesslog"
 	"github.com/netbirdio/netbird/proxy/internal/acme"
-	"github.com/netbirdio/netbird/proxy/internal/appsec"
 	"github.com/netbirdio/netbird/proxy/internal/auth"
 	"github.com/netbirdio/netbird/proxy/internal/certwatch"
 	"github.com/netbirdio/netbird/proxy/internal/conntrack"
@@ -127,10 +126,6 @@ type Server struct {
 	crowdsecMu       sync.Mutex
 	crowdsecServices map[types.ServiceID]bool
 
-	// appsecClient is the shared CrowdSec AppSec client, nil when no AppSec
-	// endpoint is configured. Stateless, so it needs no per-service lifecycle.
-	appsecClient *appsec.Client
-
 	// routerReady is closed once mainRouter is fully initialized.
 	// The mapping worker waits on this before processing updates.
 	routerReady chan struct{}
@@ -243,20 +238,6 @@ type Server struct {
 	CrowdSecAPIURL string
 	// CrowdSecAPIKey is the CrowdSec bouncer API key. Empty disables CrowdSec.
 	CrowdSecAPIKey string
-	// CrowdSecAppSecURL is the CrowdSec AppSec (WAF) endpoint, e.g.
-	// http://127.0.0.1:7422/. Empty disables request inspection. Requires
-	// CrowdSecAPIKey, which the AppSec component validates against LAPI.
-	CrowdSecAppSecURL string
-	// CrowdSecAppSecTimeout bounds a single AppSec inspection call.
-	// Zero means appsec.DefaultTimeout.
-	CrowdSecAppSecTimeout time.Duration
-	// CrowdSecAppSecMaxBodyBytes caps the request body mirrored to the AppSec
-	// engine. Zero means appsec.DefaultMaxBodyBytes; negative disables body
-	// forwarding, leaving header and URI inspection.
-	CrowdSecAppSecMaxBodyBytes int64
-	// CrowdSecAppSecMaxConcurrent bounds AppSec inspections in flight. Zero
-	// means appsec.DefaultMaxConcurrent; negative removes the bound.
-	CrowdSecAppSecMaxConcurrent int
 	// MaxSessionIdleTimeout caps the per-service session idle timeout.
 	// Zero means no cap (the proxy honors whatever management sends).
 	// Set via NB_PROXY_MAX_SESSION_IDLE_TIMEOUT for shared deployments.
@@ -403,13 +384,6 @@ func (s *Server) Start(ctx context.Context) error {
 
 	s.crowdsecRegistry = crowdsec.NewRegistry(s.CrowdSecAPIURL, s.CrowdSecAPIKey, log.NewEntry(s.Logger))
 	s.crowdsecServices = make(map[types.ServiceID]bool)
-	// Must precede the mapping worker: the worker opens the management stream
-	// and reports proxyCapabilities, which reads appsecClient. Building it
-	// afterwards would both race the read and, when the worker won, advertise
-	// the proxy as AppSec-incapable for the lifetime of that stream.
-	if err := s.initAppSec(); err != nil {
-		return err
-	}
 
 	go s.newManagementMappingWorker(runCtx, s.mgmtClient)
 
@@ -437,7 +411,6 @@ func (s *Server) Start(ctx context.Context) error {
 	}()
 
 	s.auth = auth.NewMiddleware(s.Logger, s.mgmtClient, s.geo)
-	s.auth.SetAppSec(s.appsecClient)
 	s.accessLog = accesslog.NewLogger(s.mgmtClient, s.Logger, s.TrustedProxies)
 
 	s.startDebugEndpoint()
@@ -1301,7 +1274,6 @@ func (s *Server) newManagementMappingWorker(ctx context.Context, client proto.Pr
 
 func (s *Server) proxyCapabilities() *proto.ProxyCapabilities {
 	supportsCrowdSec := s.crowdsecRegistry.Available()
-	supportsAppSec := s.appsecClient != nil
 	privateCapability := s.Private
 	// Always true: this build enforces ProxyMapping.private via the auth middleware.
 	supportsPrivateService := true
@@ -1309,7 +1281,6 @@ func (s *Server) proxyCapabilities() *proto.ProxyCapabilities {
 		SupportsCustomPorts:    &s.SupportsCustomPorts,
 		RequireSubdomain:       &s.RequireSubdomain,
 		SupportsCrowdsec:       &supportsCrowdSec,
-		SupportsAppsec:         &supportsAppSec,
 		Private:                &privateCapability,
 		SupportsPrivateService: &supportsPrivateService,
 	}
@@ -1937,67 +1908,6 @@ func (s *Server) parseRestrictions(mapping *proto.ProxyMapping) *restrict.Filter
 	})
 }
 
-// initAppSec builds the shared AppSec client when an endpoint is configured.
-// A configured-but-invalid endpoint is a startup error rather than a silent
-// downgrade: services asking for enforce would otherwise fail closed on every
-// request with no indication why.
-//
-// Runs before the management stream opens so the reported capability is stable;
-// the auth middleware picks the client up separately once it exists.
-func (s *Server) initAppSec() error {
-	if s.CrowdSecAppSecURL == "" {
-		return nil
-	}
-	if s.CrowdSecAPIKey == "" {
-		return errors.New("crowdsec appsec url is set but the crowdsec api key is empty")
-	}
-
-	// Share the middleware capture budget rather than opening a second pool:
-	// AppSec buffers before authentication, so its ceiling has to count against
-	// the same proxy-wide allowance the body tap draws from.
-	var budget appsec.Budget
-	if s.middlewareManager != nil {
-		budget = s.middlewareManager.Budget()
-	}
-
-	client, err := appsec.New(appsec.Config{
-		URL:           s.CrowdSecAppSecURL,
-		APIKey:        s.CrowdSecAPIKey,
-		Timeout:       s.CrowdSecAppSecTimeout,
-		MaxBodyBytes:  s.CrowdSecAppSecMaxBodyBytes,
-		MaxConcurrent: s.CrowdSecAppSecMaxConcurrent,
-		Budget:        budget,
-		Logger:        log.NewEntry(s.Logger),
-	})
-	if err != nil {
-		return fmt.Errorf("init crowdsec appsec: %w", err)
-	}
-
-	s.appsecClient = client
-	s.Logger.Infof("CrowdSec AppSec inspection available at %s", s.CrowdSecAppSecURL)
-	return nil
-}
-
-// appSecMode resolves the per-service AppSec mode. A service asking for
-// inspection on a proxy with no AppSec endpoint keeps its mode so the auth
-// middleware fails closed for enforce, mirroring the CrowdSec behavior.
-func (s *Server) appSecMode(mapping *proto.ProxyMapping) restrict.AppSecMode {
-	raw := mapping.GetAccessRestrictions().GetAppsecMode()
-	mode := restrict.ParseAppSecMode(raw)
-	// An unrecognized value disables inspection, which is the safe default but a
-	// silent one: with a newer management and an older proxy, a mode this build
-	// does not know would look identical to "off" on a service the operator set
-	// to enforce. Say so rather than leaving it to be discovered.
-	if mode == restrict.AppSecOff && raw != "" && raw != "off" {
-		s.Logger.Warnf("service %s requests unrecognized AppSec mode %q; this build supports %q and %q, so inspection is disabled",
-			mapping.GetId(), raw, restrict.AppSecEnforce, restrict.AppSecObserve)
-	}
-	if mode.Enabled() && s.appsecClient == nil {
-		s.Logger.Warnf("service %s requests AppSec mode %q but proxy has no AppSec endpoint configured", mapping.GetId(), mode)
-	}
-	return mode
-}
-
 // releaseCrowdSec releases the CrowdSec bouncer reference for the given
 // service if it had one.
 func (s *Server) releaseCrowdSec(svcID types.ServiceID) {
@@ -2164,17 +2074,7 @@ func (s *Server) updateMapping(ctx context.Context, mapping *proto.ProxyMapping)
 	s.warnIfGeoUnavailable(mapping.GetDomain(), mapping.GetAccessRestrictions())
 
 	maxSessionAge := time.Duration(mapping.GetAuth().GetMaxSessionAgeSeconds()) * time.Second
-	settings := auth.DomainSettings{
-		Schemes:           schemes,
-		SessionPublicKey:  mapping.GetAuth().GetSessionKey(),
-		SessionExpiration: maxSessionAge,
-		AccountID:         accountID,
-		ServiceID:         svcID,
-		IPRestrictions:    ipRestrictions,
-		Private:           mapping.GetPrivate(),
-		AppSecMode:        s.appSecMode(mapping),
-	}
-	if err := s.auth.AddDomain(mapping.GetDomain(), settings); err != nil {
+	if err := s.auth.AddDomain(mapping.GetDomain(), schemes, mapping.GetAuth().GetSessionKey(), maxSessionAge, accountID, svcID, ipRestrictions, mapping.GetPrivate()); err != nil {
 		return fmt.Errorf("auth setup for domain %s: %w", mapping.GetDomain(), err)
 	}
 	m := s.protoToMapping(ctx, mapping)

shared/management/client/grpc.go

@@ -25,6 +25,7 @@ import (
 	"github.com/netbirdio/netbird/encryption"
 	"github.com/netbirdio/netbird/shared/management/domain"
 	nbmgmtgrpc "github.com/netbirdio/netbird/shared/management/grpc"
+	nbnetworkmap "github.com/netbirdio/netbird/shared/management/networkmap"
 	"github.com/netbirdio/netbird/shared/management/proto"
 	"github.com/netbirdio/netbird/util/wsproxy"
 )
@@ -472,11 +473,48 @@ func (c *GrpcClient) GetNetworkMap(sysInfo *system.Info) (*proto.NetworkMap, err
 		return nil, err
 	}
 
-	if decryptedResp.GetNetworkMap() == nil {
+	var nm *proto.NetworkMap
+	if version := decryptedResp.GetVersion(); version == int32(nbmgmtgrpc.ComponentNetworkMap) {
+		// Components-format peer: decode the envelope back to typed
+		// components, run Calculate() locally, and convert to the wire
+		// NetworkMap shape the rest of the engine consumes. Components are
+		// retained so future incremental updates can apply deltas instead
+		// of doing a full reconstruction.
+		envelope := decryptedResp.GetNetworkMapEnvelope()
+		if envelope == nil {
+			return nil, fmt.Errorf("received a SyncReponse indicating use of components network map, but components are missing")
+		}
+
+		localKey := c.key.String()
+		dnsName := ""
+		if pc := decryptedResp.GetPeerConfig(); pc != nil {
+			// PeerConfig.Fqdn = "<dns_label>.<dns_domain>" — extract the
+			// shared domain by stripping the peer's own label prefix. Falls
+			// back to empty if the FQDN doesn't have the expected shape.
+			dnsName = extractDNSDomainFromFQDN(pc.GetFqdn())
+		}
+		result, err := nbnetworkmap.EnvelopeToNetworkMap(c.ctx, envelope, localKey, dnsName)
+		if err != nil {
+			return nil, fmt.Errorf("decode network map envelope: %w", err)
+		}
+		nm = result.NetworkMap
+	} else {
+		nm = decryptedResp.GetNetworkMap()
+	}
+	if nm == nil {
 		return nil, fmt.Errorf("invalid msg, required network map")
 	}
 
-	return decryptedResp.GetNetworkMap(), nil
+	return nm, nil
+}
+
+func extractDNSDomainFromFQDN(fqdn string) string {
+	for i := 0; i < len(fqdn); i++ {
+		if fqdn[i] == '.' && i+1 < len(fqdn) {
+			return fqdn[i+1:]
+		}
+	}
+	return ""
 }
 
 func (c *GrpcClient) connectToSyncStream(ctx context.Context, serverPubKey wgtypes.Key, sysInfo *system.Info) (proto.ManagementService_SyncClient, error) {