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

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

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

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

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

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

service/rpc/request_task_stale_stream_test.go

@@ -0,0 +1,46 @@
+package rpc
+
+import (
+	"context"
+	"errors"
+	"testing"
+
+	"github.com/nezhahq/nezha/model"
+	"github.com/nezhahq/nezha/service/singleton"
+)
+
+// When a server is edited mid-session, updateServer swaps a new *Server into
+// ServerShared that adopts the live stream holder. The agent's RequestTask
+// cleanup must detach the stream from whichever *Server is currently published,
+// not the stale object captured when the stream attached — otherwise the new
+// object keeps reporting the agent as online on a dead stream.
+func TestRequestTaskCleanupDetachesStreamFromCurrentServerAfterEdit(t *testing.T) {
+	reporter := requestTaskSecurityServer(7, 200, "ffffffff-ffff-ffff-ffff-ffffffffffff")
+	setupRequestTaskSecurityFixture(t, []*model.Server{reporter}, nil, map[uint64]model.UserInfo{
+		200: {Role: model.RoleMember},
+	}, map[string]uint64{"reporter-secret": 200})
+
+	old, ok := singleton.ServerShared.Get(reporter.ID)
+	if !ok {
+		t.Fatalf("server %d not found", reporter.ID)
+	}
+
+	stream := requestTaskSecurityAuthedStream("reporter-secret", reporter.UUID)
+	stream.onRecv = func() {
+		edited := &model.Server{Common: model.Common{ID: old.ID, UserID: old.UserID}, UUID: old.UUID, Name: "edited"}
+		edited.CopyFromRunningServer(old)
+		singleton.ServerShared.Update(edited, "")
+	}
+
+	if err := NewNezhaHandler().RequestTask(stream); !errors.Is(err, context.Canceled) {
+		t.Fatalf("expected RequestTask to finish after Recv error, got %v", err)
+	}
+
+	current, ok := singleton.ServerShared.Get(reporter.ID)
+	if !ok {
+		t.Fatalf("server %d not found after edit", reporter.ID)
+	}
+	if got := current.GetTaskStream(); got != nil {
+		t.Fatalf("edited server must report offline after the agent stream dropped, got %T", got)
+	}
+}