diff --git a/admin_test.go b/admin_test.go
index 74ce70e..1645e99 100644
--- a/admin_test.go
+++ b/admin_test.go
@@ -1998,34 +1998,39 @@ func TestAdminTargets_AccessListsMatchEnforcement_AndCountersReflectTraffic(t *t
 // --- default-preserving: the live cluster's real shape (security audit, 2026-08-22) ---
 
 // TestDefaultPreserving_LiveClusterShape_LoadsAndServesIdentically models
-// the live production shape this round's five fixes must not break
-// (project brief): 5 "friend" users plus 4 "home" users (steve-cw and
-// kevinsandom are named live users; the rest are representative — real
-// names this test has no visibility into, which is exactly why finding
-// 5's fix stops short of restricting the name charset, see buildEntry's
-// own doc comment, auth.go) and 11 MCP servers. A2A agents are out of
-// scope here: none of this round's five findings touch mcp_a2a.go or any
-// A2A-specific code path, so there is no mechanism by which they could
-// regress agent handling.
+// the live production shape this round's fixes must not break (project
+// brief): 5 "friend" users plus 4 "home" users (steve-cw and kevinsandom
+// are named live users; the rest are representative — real names this
+// test has no visibility into, which is exactly why finding 5's fix stops
+// short of restricting the name charset, see buildEntry's own doc
+// comment, auth.go) and 11 MCP servers. A2A agents are out of scope here:
+// none of this round's fixes touch mcp_a2a.go or any A2A-specific code
+// path, so there is no mechanism by which they could regress agent
+// handling.
 //
 // It proves construction succeeds unchanged (finding 5's new empty/
-// duplicate-name checks accept every one of these real names), a normal
-// federated tools/list call from an ordinary user still succeeds under a
-// realistic per-minute limit (finding 1b's per-backend weighting does not
-// spuriously throttle ordinary usage — 11 servers is comfortably under a
-// 60/min budget), and the admin dashboard's usage poll still returns
-// exactly one row per user and per group in a single store round trip
-// (finding 4's chunking is a no-op at 10 users, well under
-// adminUsageChunkScopes).
+// duplicate-name checks accept every one of these real names), a
+// friends-group user's configured requests-per-minute budget still buys
+// EXACTLY that many tools/list calls against all 11 servers — not
+// budget/11 — regressing security review round 2's critical finding 3
+// (weighting tools/list by backend count was reverted specifically
+// because it was not default-preserving at this exact shape: both live
+// groups have an empty MCP allow-list, i.e. all 11 servers, so a naive
+// per-backend weight would have silently cut every configured
+// requests-per-minute budget by ~11x here), and the admin dashboard's
+// usage poll still returns exactly one row per user and per group in a
+// single store round trip (finding 4's chunking is a no-op at 10 users,
+// well under adminUsageChunkScopes).
 func TestDefaultPreserving_LiveClusterShape_LoadsAndServesIdentically(t *testing.T) {
 	friendNames := []string{"steve-cw", "kevinsandom", "friend3", "friend4", "friend5"}
 	homeNames := []string{"home1", "home2", "home3", "home4"}
+	const friendsRPM = 60
 
 	cfg := CreateConfig()
 	cfg.Providers = map[string]*ProviderConfig{"openai": {Type: "openai", APIKey: "k"}}
 	cfg.Admin = &AdminConfig{Enabled: true}
 	cfg.Groups = map[string]*GroupConfig{
-		"friends": {Limits: &LimitsConfig{RequestsPerMinute: 60}},
+		"friends": {Limits: &LimitsConfig{RequestsPerMinute: friendsRPM}},
 		"home":    {Limits: &LimitsConfig{RequestsPerMinute: 120}},
 	}
 
@@ -2049,10 +2054,16 @@ func TestDefaultPreserving_LiveClusterShape_LoadsAndServesIdentically(t *testing
 		t.Fatalf("New: want the realistic live-cluster config to load unchanged, got error: %v", err)
 	}
 
-	rec := httptest.NewRecorder()
-	h.ServeHTTP(rec, newFederatedRequest(t, "sk-steve-cw", jsonrpcRequest{JSONRPC: jsonrpcVersion, Method: "tools/list", ID: json.RawMessage("1")}))
-	if rec.Code != http.StatusOK {
-		t.Fatalf("tools/list status = %d, want 200, body=%s", rec.Code, rec.Body.String())
+	// Drives friendsRPM calls, not just one (security review round 2:
+	// a single-call assertion cannot distinguish "budget buys N calls"
+	// from "budget buys N/11 calls" — exactly the blind spot that let the
+	// weighted-charge regression ship undetected).
+	for i := 1; i <= friendsRPM; i++ {
+		rec := httptest.NewRecorder()
+		h.ServeHTTP(rec, newFederatedRequest(t, "sk-steve-cw", jsonrpcRequest{JSONRPC: jsonrpcVersion, Method: "tools/list", ID: json.RawMessage(fmt.Sprintf("%d", i))}))
+		if rec.Code != http.StatusOK {
+			t.Fatalf("tools/list call %d/%d status = %d, want 200 (budget must buy all %d calls against 11 servers), body=%s", i, friendsRPM, rec.Code, friendsRPM, rec.Body.String())
+		}
 	}
 
 	usageReq := httptest.NewRequest("GET", adminUsagePath, nil)
diff --git a/limits.go b/limits.go
index 013b6b7..61ae8c9 100644
--- a/limits.go
+++ b/limits.go
@@ -1081,39 +1081,14 @@ func buildBudgetProbes(scopes []limitScope, now time.Time) (probes []budgetProbe
 // call completes, one request may still push a scope over its budget — an
 // accepted trade-off, unchanged from before this round.
 func (l *limiter) checkAndCount(scopes []limitScope) *limitViolation {
-	return l.checkAndCountWeighted(scopes, 1)
-}
-
-// checkAndCountWeighted is checkAndCount generalized to charge `weight`
-// requests against every scope's requests-per-minute/requests-per-day
-// counters instead of exactly one (security audit finding 1b, 2026-08-22):
-// a route whose single incoming client request causes this gateway to
-// issue `weight` outbound backend calls — today, only the federated MCP
-// tools/list fan-out (mcp_federation.go's handleMCPFederated, weight =
-// len(allowedMCPServerNames(grp))) — must charge its real cost, not
-// undercount a fan-out as a single request. weight<=0 is treated as 1
-// (never zero-cost, never negative); checkAndCount(scopes) is exactly
-// checkAndCountWeighted(scopes, 1), so every other call site's behavior is
-// completely unchanged by this method existing. Semantics beyond the
-// weighted delta are identical to checkAndCount's own doc comment in
-// every other respect: one storeIncrAndGetMulti round trip, requests-
-// per-minute/requests-per-day compared against the just-incremented
-// (now weighted) count, token/cost budgets read and compared unweighted
-// (a fan-out has no tokens or USD cost of its own — those probes come
-// from buildBudgetProbes, untouched by weight), violation precedence
-// unchanged.
-func (l *limiter) checkAndCountWeighted(scopes []limitScope, weight int64) *limitViolation {
-	if weight <= 0 {
-		weight = 1
-	}
 	now := l.now()
 
 	entries := make([]counterIncr, 0, len(scopes)*checkAndCountKeysPerScope)
 	for _, sc := range scopes {
 		entries = append(entries,
-			newCounterIncr(sc.kind, sc.id, metricReq, windowMin, now, weight, minWindowTTL),
-			newCounterIncr(sc.kind, sc.id, metricReq, windowDay, now, weight, dayWindowTTL),
-			newCounterIncr(sc.kind, sc.id, metricReq, windowHour, now, weight, hourWindowTTL),
+			newCounterIncr(sc.kind, sc.id, metricReq, windowMin, now, 1, minWindowTTL),
+			newCounterIncr(sc.kind, sc.id, metricReq, windowDay, now, 1, dayWindowTTL),
+			newCounterIncr(sc.kind, sc.id, metricReq, windowHour, now, 1, hourWindowTTL),
 		)
 	}
 	probes, reads := buildBudgetProbes(scopes, now)
diff --git a/mcp_federation.go b/mcp_federation.go
index 45e7782..25e3a6d 100644
--- a/mcp_federation.go
+++ b/mcp_federation.go
@@ -67,20 +67,51 @@ const (
 )
 
 // mcpBackendResponseMaxBytes caps a single backend MCP server's HTTP
-// response body doBackendJSONRPC will read into memory — previously
-// maxRequestBytes (10MiB, routes_unified.go), a budget sized for a
-// CLIENT's own chat/embeddings request body, not a backend's JSON-RPC
-// reply (security audit finding 1c, 2026-08-22). 4MiB matches this
-// codebase's other "generous but bounded" read caps for a non-primary-
-// request-body read (cache.go's maxAccountingTeeBytes is also 4MiB) —
-// comfortably above any tools/list schema listing or tools/call result
-// this gateway has actually seen live (typically low KB), while bounding
-// mcpFederatedFanoutConcurrency backends' worst-case concurrent memory to
-// 4 × 4MiB = 16MiB, against the unbounded ~110MiB (11 production servers
-// × up to 10MiB each, all fired at once) the security audit measured
-// before this fix.
+// response body for the tools/list FAN-OUT path only (mcpFederatedToolsList)
+// — previously maxRequestBytes (10MiB, routes_unified.go), a budget sized
+// for a CLIENT's own chat/embeddings request body, not a backend's
+// JSON-RPC reply (security audit finding 1c, 2026-08-22). 4MiB matches
+// this codebase's other "generous but bounded" read caps for a non-
+// primary-request-body read (cache.go's maxAccountingTeeBytes is also
+// 4MiB) — measured at 547KiB for a real tools/list response describing
+// 20 fat tools (security review round 2, 2026-08-22), comfortably below
+// this cap — while bounding mcpFederatedFanoutConcurrency backends'
+// worst-case concurrent memory to a small multiple of 4MiB, against the
+// unbounded ~110MiB (11 production servers × up to 10MiB each, all fired
+// at once) the security audit measured before this fix. This cap is
+// deliberately NOT used for tools/call — see mcpBackendCallResponseMaxBytes,
+// whose own doc comment covers why the fan-out memory argument does not
+// apply there.
 const mcpBackendResponseMaxBytes = 4 << 20
 
+// mcpBackendCallResponseMaxBytes caps a single backend's response for the
+// tools/call path (mcpFederatedToolsCall) — kept at maxRequestBytes
+// (10MiB), the SAME budget this file used everywhere before finding 1c,
+// deliberately NOT shrunk to mcpBackendResponseMaxBytes (security review
+// round 2, 2026-08-22, important finding 4): tools/call always resolves
+// to exactly ONE backend (resolveFederatedTool), so the fan-out memory-
+// amplification argument that justifies the fan-out path's smaller cap
+// does not hold here at all — a single tool result can legitimately be
+// much larger than a tools/list schema listing (an extracted document, an
+// image or other blob a tool returns), and silently truncating one into
+// a generic parse failure would be a real, avoidable regression for any
+// tool whose legitimate results sit between 4MiB and 10MiB.
+const mcpBackendCallResponseMaxBytes = maxRequestBytes
+
+// mcpResponseTooLargeMarker is a fixed, unique substring doBackendJSONRPC
+// embeds in its own error message when a backend response hits its
+// call's maxBytes cap (security review round 2, 2026-08-22, important
+// finding 4) — detected via strings.Contains, deliberately NOT an
+// errors.Is-checked sentinel. matchesSentinel's own doc comment
+// (limits.go) proves errors.Is safe under Yaegi only for chains reaching
+// a call site this package's own yaegi-check harness actually exercises
+// at runtime under the real interpreter — federation's tools/call path
+// is not one of those exercised paths today, so a plain string check
+// trades a little idiom for zero interpreter risk here, rather than
+// relying on reasoning about compiled-vs-interpreted-origin values that
+// nothing in the gate actually verifies for this specific path.
+const mcpResponseTooLargeMarker = "mcp backend response too large"
+
 // jsonrpcRequest is one JSON-RPC 2.0 request/notification, decoded from
 // the client's POST body and re-encoded (with a synthetic id) for this
 // gateway's own outbound calls to a backend MCP server (mcpBackendCall).
@@ -249,27 +280,32 @@ func allowedMCPServerNames(cfg *Config, grp *group) []string {
 //
 // checkAndCount runs once per incoming request, before method dispatch —
 // covering the caller's own user/group/total scopes exactly like every
-// other metered route — with per-target attribution added separately
+// other metered route, charging the plain weight of one request
+// regardless of method — with per-target attribution added separately
 // inside mcpFederatedToolsList/mcpFederatedToolsCall (their own doc
 // comments cover exactly what "attributed" means for each).
 //
-// The weight charged to checkAndCount is 1 for every method EXCEPT
-// tools/list (security audit finding 1b, 2026-08-22): a tools/list call
-// fans out to every server allowedMCPServerNames(grp) returns — up to the
-// whole MCP catalog, concurrently (mcpFederatedToolsList's own doc
-// comment) — so one incoming client request can cost this gateway many
-// real outbound backend calls. Charging it as a single request against
-// the caller's requests-per-minute/requests-per-day budget undercounts
-// that real cost; weighting it by len(names) instead makes the budget
-// reflect what the request actually does. This is computed and charged
-// exactly once, right here, before dispatch — never doubled by anything
-// mcpFederatedToolsList itself does (its own countTargetRequests call is a
-// SEPARATE scope kind, "mcp"/target-id, entirely disjoint from the
-// user/group/total scopes checkAndCountWeighted evaluates here; see
-// countTargetRequests' own doc comment, limits.go). tools/call, by
-// contrast, always resolves to exactly ONE backend (resolveFederatedTool),
-// so it keeps the plain weight of 1 — checkAndCountWeighted(scopes, 1) is
-// byte-for-byte checkAndCount's own existing behavior.
+// tools/list's fan-out is deliberately NOT charged against the caller's
+// own requests-per-minute/requests-per-day budget (security review round
+// 2, 2026-08-22, critical finding 3 — reverting security audit finding
+// 1b's weighted-charge attempt): weighting checkAndCount by the number of
+// allowed MCP servers is not default-preserving — on this operator's own
+// live fleet, both configured groups have an empty MCP allow-list (every
+// server allowed), so EVERY tools/list call would have weighed 11,
+// silently turning a configured requests-per-minute: 60 into an effective
+// budget of 5 successful tools/list calls per minute, and
+// examples/kubernetes.yaml's shipped requests-per-minute: 10 would have
+// 429'd the very FIRST tools/list call ever made — with no config change
+// on the operator's part. The fan-out's real amplification is still
+// fully visible to an operator without silently consuming tenant quota:
+// mcpFederatedToolsList's own countTargetRequests call attributes one
+// request to EVERY server actually attempted, at the "mcp"/target-id
+// scope (limits.go) — a SEPARATE scope kind from the user/group/total
+// scopes checkAndCount evaluates here — so the dashboard's targets view
+// already shows the true per-server cost. The fan-out's outbound HTTP/
+// memory cost is bounded instead by mcpFederatedFanoutConcurrency's
+// semaphore (mcpFederatedToolsList's own doc comment), which is the
+// actual protection against the amplification finding 1 described.
 func (g *Gateway) handleMCPFederated(w http.ResponseWriter, r *http.Request, u *user, grp *group) {
 	if r.Header.Get("Upgrade") != "" {
 		writeOAIError(w, http.StatusNotImplemented, "invalid_request_error", "websocket/upgrade passthrough not supported")
@@ -292,17 +328,7 @@ func (g *Gateway) handleMCPFederated(w http.ResponseWriter, r *http.Request, u *
 	}
 
 	scopes := withTotalScope(buildLimitScopes(u, grp))
-
-	var toolsListNames []string
-	weight := int64(1)
-	if req.Method == "tools/list" {
-		toolsListNames = allowedMCPServerNames(g.cfg, grp)
-		if n := int64(len(toolsListNames)); n > weight {
-			weight = n
-		}
-	}
-
-	if violation := g.limiter.checkAndCountWeighted(scopes, weight); violation != nil {
+	if violation := g.limiter.checkAndCount(scopes); violation != nil {
 		writeLimitViolation(w, violation)
 		return
 	}
@@ -313,7 +339,7 @@ func (g *Gateway) handleMCPFederated(w http.ResponseWriter, r *http.Request, u *
 	case req.Method == "ping":
 		writeJSONRPCResult(w, req.ID, map[string]any{})
 	case req.Method == "tools/list":
-		g.mcpFederatedToolsList(w, r, req, toolsListNames)
+		g.mcpFederatedToolsList(w, r, req, allowedMCPServerNames(g.cfg, grp))
 	case req.Method == "tools/call":
 		g.mcpFederatedToolsCall(w, r, req, grp)
 	case strings.HasPrefix(req.Method, "notifications/"):
@@ -413,8 +439,16 @@ type mcpToolsListResult struct {
 // doBackendJSONRPC issues one JSON-RPC 2.0 POST to targetURL — an MCP
 // server's own configured base URL, the same single endpoint
 // handleTargetProxy reverse-proxies every method to — carrying sessionID
-// as the Mcp-Session-Id request header when non-empty. It is the one place
-// that actually builds and sends an HTTP request for this file; every
+// as the Mcp-Session-Id request header when non-empty, and reading at
+// most maxBytes of the response body (mcpBackendResponseMaxBytes for the
+// tools/list fan-out, mcpBackendCallResponseMaxBytes for tools/call —
+// their own doc comments cover why the two differ; security review round
+// 2, 2026-08-22, important finding 4). A response body that hits maxBytes
+// is reported with mcpResponseTooLargeMarker in its message, not silently truncated into
+// whatever partial (and likely invalid) JSON happened to fit.
+//
+// It is the one place that actually builds and sends an HTTP request for
+// this file; every
 // other function in the handshake-fallback machinery (mcpBackendCall,
 // mcpBackendHandshake, mcpBackendCloseSession's own DELETE aside) goes
 // through this.
@@ -439,7 +473,7 @@ type mcpToolsListResult struct {
 // that WAS valid JSON but not a real JSON-RPC response (see
 // parseBackendJSONRPC's own validity check) would have silently produced
 // an all-zero-fields response otherwise.
-func (g *Gateway) doBackendJSONRPC(ctx context.Context, targetURL, method string, params any, sessionID string) (resp *jsonrpcResponse, status int, respSessionID string, err error) {
+func (g *Gateway) doBackendJSONRPC(ctx context.Context, targetURL, method string, params any, sessionID string, maxBytes int64) (resp *jsonrpcResponse, status int, respSessionID string, err error) {
 	paramsRaw, err := json.Marshal(params)
 	if err != nil {
 		return nil, 0, "", err
@@ -476,10 +510,17 @@ func (g *Gateway) doBackendJSONRPC(ctx context.Context, targetURL, method string
 	status = httpResp.StatusCode
 	respSessionID = httpResp.Header.Get(mcpSessionHeader)
 
-	respBody, err := io.ReadAll(io.LimitReader(httpResp.Body, mcpBackendResponseMaxBytes))
+	// Reads one byte past maxBytes so a response that hits the cap is
+	// DETECTABLE (len(respBody) > maxBytes) rather than silently truncated
+	// and handed to parseBackendJSONRPC to fail confusingly (security
+	// review round 2, 2026-08-22, important finding 4).
+	respBody, err := io.ReadAll(io.LimitReader(httpResp.Body, maxBytes+1))
 	if err != nil {
 		return nil, status, respSessionID, err
 	}
+	if int64(len(respBody)) > maxBytes {
+		return nil, status, respSessionID, fmt.Errorf("%s: exceeds %d bytes", mcpResponseTooLargeMarker, maxBytes)
+	}
 	if status < 200 || status >= 300 {
 		return nil, status, respSessionID, fmt.Errorf("upstream returned HTTP %d", status)
 	}
@@ -500,13 +541,13 @@ func (g *Gateway) doBackendJSONRPC(ctx context.Context, targetURL, method string
 // the initialize call itself failed outright (network error, non-2xx, a
 // JSON-RPC error response to initialize itself) — there is no retry of a
 // failed handshake.
-func (g *Gateway) mcpBackendHandshake(ctx context.Context, targetURL string) (sessionID string, err error) {
+func (g *Gateway) mcpBackendHandshake(ctx context.Context, targetURL string, maxBytes int64) (sessionID string, err error) {
 	initParams := map[string]any{
 		"protocolVersion": defaultMCPProtocolVersion,
 		"capabilities":    map[string]any{},
 		"clientInfo":      map[string]any{"name": "traefik-llmgateway", "version": pluginVersion},
 	}
-	resp, _, respSessionID, err := g.doBackendJSONRPC(ctx, targetURL, "initialize", initParams, "")
+	resp, _, respSessionID, err := g.doBackendJSONRPC(ctx, targetURL, "initialize", initParams, "", maxBytes)
 	if err != nil {
 		return "", err
 	}
@@ -567,9 +608,14 @@ func (g *Gateway) mcpBackendCloseSession(ctx context.Context, targetURL, session
 // except the close itself — see mcpSessionCloseTimeout's own doc
 // comment); callers derive it from context.WithTimeout(r.Context(), ...)
 // with toolsListBackendTimeout or toolsCallBackendTimeout (MF1, review
-// round 2, 2026-08-21).
-func (g *Gateway) mcpBackendCall(ctx context.Context, targetURL, method string, params any) (*jsonrpcResponse, error) {
-	resp, status, _, err := g.doBackendJSONRPC(ctx, targetURL, method, params, "")
+// round 2, 2026-08-21). maxBytes is doBackendJSONRPC's own response-size
+// cap, passed straight through to every doBackendJSONRPC call this
+// function makes (bare attempt, handshake, retry) — callers pass
+// mcpBackendResponseMaxBytes or mcpBackendCallResponseMaxBytes depending
+// on which path they are (security review round 2, 2026-08-22, important
+// finding 4).
+func (g *Gateway) mcpBackendCall(ctx context.Context, targetURL, method string, params any, maxBytes int64) (*jsonrpcResponse, error) {
+	resp, status, _, err := g.doBackendJSONRPC(ctx, targetURL, method, params, "", maxBytes)
 	if err == nil && resp.Error == nil {
 		return resp, nil
 	}
@@ -579,7 +625,7 @@ func (g *Gateway) mcpBackendCall(ctx context.Context, targetURL, method string,
 		return nil, err
 	}
 
-	sessionID, initErr := g.mcpBackendHandshake(ctx, targetURL)
+	sessionID, initErr := g.mcpBackendHandshake(ctx, targetURL, maxBytes)
 	if initErr != nil {
 		// The handshake itself failed: surface the bare attempt's own
 		// outcome when it has one (a JSON-RPC error response is more
@@ -600,7 +646,7 @@ func (g *Gateway) mcpBackendCall(ctx context.Context, targetURL, method string,
 		}()
 	}
 
-	retryResp, _, _, retryErr := g.doBackendJSONRPC(ctx, targetURL, method, params, sessionID)
+	retryResp, _, _, retryErr := g.doBackendJSONRPC(ctx, targetURL, method, params, sessionID, maxBytes)
 	if retryErr != nil {
 		return nil, retryErr
 	}
@@ -685,19 +731,28 @@ func lastSSEDataLine(body []byte) []byte {
 // per allowed server with NO cap at all: with the 11 production servers,
 // one client call could hold up to 11 concurrent connections and up to
 // 11× mcpBackendResponseMaxBytes of backend response bodies in memory at
-// once, entirely outside this gateway's own admission control. 4 keeps
-// every server eventually contacted — a goroutine blocks on the
-// semaphore, it is never dropped or skipped — while capping the worst
-// case to 4 concurrent backends' cost, not the whole catalog's.
-const mcpFederatedFanoutConcurrency = 4
+// once, entirely outside this gateway's own admission control. Every
+// server is eventually contacted regardless of this bound — a goroutine
+// blocks on the semaphore, it is never dropped or skipped. 8 (raised from
+// an initial 4, security review round 2, 2026-08-22, important finding
+// 3): 4 measurably tripled worst-case wall-clock latency for an
+// unrestricted group against all 11 servers (907ms measured at 300ms/
+// server, versus ~300ms pre-bound) and, against hung backends, pushed the
+// worst case to ceil(11/4)×toolsListBackendTimeout = 60s — well past the
+// ~30s timeout many real MCP clients use, holding the ServeHTTP goroutine
+// the whole time. 8 nearly halves both: ceil(11/8)=2 batches. The fanoutCtx
+// deadline below (not this constant alone) is what actually re-bounds the
+// worst case back down near toolsListBackendTimeout regardless of
+// concurrency or backend count — see its own doc comment.
+const mcpFederatedFanoutConcurrency = 8
 
 // mcpFederatedToolsList implements the federated "tools/list": it fans
 // out one mcpBackendCall per server in names (handleMCPFederated's own
-// allowedMCPServerNames(grp) call, made once there so the SAME set that
-// was weighed into checkAndCountWeighted is the set actually contacted),
-// at most mcpFederatedFanoutConcurrency at a time — each under its own
-// toolsListBackendTimeout budget (context.WithTimeout off r.Context(),
-// MF1) — then merges every reachable server's own tools into one result,
+// allowedMCPServerNames(grp) call, made once there so the exact set
+// contacted is always the one the client actually asked to reach), at
+// most mcpFederatedFanoutConcurrency at a time — every backend sharing
+// ONE overall toolsListBackendTimeout deadline (fanoutCtx, below) — then
+// merges every reachable server's own tools into one result,
 // each tool's name prefixed "<serverName>_<toolName>", the exact
 // underscore-separator convention the old agentgateway this plugin
 // replaces used, and the one pugbot's mcpclient and agentkit have already
@@ -725,6 +780,24 @@ func (g *Gateway) mcpFederatedToolsList(w http.ResponseWriter, r *http.Request,
 		merged = make([]mcpTool, 0, len(names))
 		failed []string
 	)
+	// fanoutCtx bounds the WHOLE fan-out to toolsListBackendTimeout from
+	// when THIS call started — not each individual backend's own start
+	// (security review round 2, 2026-08-22, important finding 3). With
+	// mcpFederatedFanoutConcurrency backends running at once, a later
+	// batch's goroutines only begin once an earlier one's semaphore slot
+	// frees; giving each one its own FRESH toolsListBackendTimeout budget
+	// (this function's pre-bound design, and its own initial post-
+	// semaphore revision) let total wall-clock grow with
+	// ceil(len(names)/mcpFederatedFanoutConcurrency) batches. Every
+	// backend below shares this ONE already-ticking context instead —
+	// context.WithTimeout always resolves to the earlier of a parent's
+	// existing deadline and its own duration (proven by
+	// TestHandleMCPFederated_ToolsList_SlowBackend_BoundedByRequestContext),
+	// so deriving fanoutCtx from r.Context() here preserves that same
+	// property for the request's own deadline, on top of this one.
+	fanoutCtx, fanoutCancel := context.WithTimeout(r.Context(), toolsListBackendTimeout)
+	defer fanoutCancel()
+
 	sem := make(chan struct{}, mcpFederatedFanoutConcurrency)
 	for _, name := range names {
 		wg.Add(1)
@@ -752,10 +825,8 @@ func (g *Gateway) mcpFederatedToolsList(w http.ResponseWriter, r *http.Request,
 			sem <- struct{}{}
 			defer func() { <-sem }()
 
-			ctx, cancel := context.WithTimeout(r.Context(), toolsListBackendTimeout)
-			defer cancel()
 			targetURL := g.cfg.MCPServers[name].URL
-			resp, err := g.mcpBackendCall(ctx, targetURL, "tools/list", struct{}{})
+			resp, err := g.mcpBackendCall(fanoutCtx, targetURL, "tools/list", struct{}{}, mcpBackendResponseMaxBytes)
 
 			mu.Lock()
 			defer mu.Unlock()
@@ -867,11 +938,24 @@ func (g *Gateway) mcpFederatedToolsCall(w http.ResponseWriter, r *http.Request,
 	ctx, cancel := context.WithTimeout(r.Context(), toolsCallBackendTimeout)
 	defer cancel()
 	targetURL := g.cfg.MCPServers[serverName].URL
-	resp, err := g.mcpBackendCall(ctx, targetURL, "tools/call", mcpToolCallParams{Name: toolName, Arguments: params.Arguments})
+	// mcpBackendCallResponseMaxBytes (10MiB), not the fan-out's smaller
+	// mcpBackendResponseMaxBytes (security review round 2, 2026-08-22,
+	// important finding 4): tools/call always resolves to exactly ONE
+	// backend, so the fan-out memory-amplification argument does not
+	// apply here — see mcpBackendCallResponseMaxBytes' own doc comment.
+	resp, err := g.mcpBackendCall(ctx, targetURL, "tools/call", mcpToolCallParams{Name: toolName, Arguments: params.Arguments}, mcpBackendCallResponseMaxBytes)
 	g.limiter.countTargetRequest(targetKindMCP, serverName)
 	if err != nil {
 		g.logf("federated tools/call: server %q: %v", serverName, err)
-		writeJSONRPCErrorResponse(w, req.ID, jsonrpcInternalError, "upstream error")
+		// A response that hit mcpBackendCallResponseMaxBytes gets its own
+		// specific message, not the generic "upstream error" a truncated-
+		// then-unparsable body would otherwise produce (security review
+		// round 2, 2026-08-22, important finding 4).
+		if strings.Contains(err.Error(), mcpResponseTooLargeMarker) {
+			writeJSONRPCErrorResponse(w, req.ID, jsonrpcInternalError, "response too large")
+		} else {
+			writeJSONRPCErrorResponse(w, req.ID, jsonrpcInternalError, "upstream error")
+		}
 		return
 	}
 
diff --git a/mcp_federation_test.go b/mcp_federation_test.go
index dd19b17..5bf0582 100644
--- a/mcp_federation_test.go
+++ b/mcp_federation_test.go
@@ -1031,7 +1031,7 @@ func TestMcpBackendCall_HandshakeFallbackTrigger(t *testing.T) {
 				t.Fatal("handler is not *Gateway")
 			}
 
-			_, _ = gw.mcpBackendCall(context.Background(), srv.URL, "tools/list", struct{}{})
+			_, _ = gw.mcpBackendCall(context.Background(), srv.URL, "tools/list", struct{}{}, mcpBackendResponseMaxBytes)
 
 			gotCalls := callCount.Load()
 			if tc.wantHandshake && gotCalls < 2 {
@@ -1060,12 +1060,134 @@ func TestMcpBackendCall_NetworkFailure_NeverAttemptsHandshake(t *testing.T) {
 		t.Fatal("handler is not *Gateway")
 	}
 
-	_, err = gw.mcpBackendCall(context.Background(), "http://127.0.0.1:1", "tools/list", struct{}{})
+	_, err = gw.mcpBackendCall(context.Background(), "http://127.0.0.1:1", "tools/list", struct{}{}, mcpBackendResponseMaxBytes)
 	if err == nil {
 		t.Fatal("want an error (nothing listening on 127.0.0.1:1)")
 	}
 }
 
+// --- security review round 2, important finding 4: dual response caps ---
+
+// TestMcpBackendCall_ResponseExceedsCap_ReturnsDistinctTooLargeError is a
+// fast, small-scale unit test of doBackendJSONRPC's own maxBytes
+// mechanism, using an artificially tiny cap so the test needs no
+// multi-megabyte payload: a backend response over the cap must be
+// reported with mcpResponseTooLargeMarker in its message —
+// distinguishable via strings.Contains from an ordinary network/parse
+// failure — never silently truncated into a confusing parse error.
+func TestMcpBackendCall_ResponseExceedsCap_ReturnsDistinctTooLargeError(t *testing.T) {
+	const tinyCap = 64
+	big := strings.Repeat("x", tinyCap*4) // well past tinyCap once JSON-encoded
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		var req jsonrpcRequest
+		_ = json.NewDecoder(r.Body).Decode(&req)
+		result, _ := json.Marshal(map[string]string{"data": big})
+		w.Header().Set("Content-Type", "application/json")
+		_ = json.NewEncoder(w).Encode(jsonrpcResponse{ID: req.ID, Result: result})
+	}))
+	defer srv.Close()
+
+	cfg := newFederationTestConfig(srv.URL, "http://beta.invalid", false)
+	h, err := New(context.Background(), http.HandlerFunc(func(http.ResponseWriter, *http.Request) {}), cfg, "llmgw")
+	if err != nil {
+		t.Fatalf("New: %v", err)
+	}
+	gw, ok := h.(*Gateway)
+	if !ok {
+		t.Fatal("handler is not *Gateway")
+	}
+
+	_, err = gw.mcpBackendCall(context.Background(), srv.URL, "tools/list", struct{}{}, tinyCap)
+	if err == nil {
+		t.Fatal("want an error for a response exceeding the cap")
+	}
+	if !strings.Contains(err.Error(), mcpResponseTooLargeMarker) {
+		t.Errorf("err = %v, want it to contain %q", err, mcpResponseTooLargeMarker)
+	}
+}
+
+// TestMcpBackendCall_ResponseAtOrUnderCap_Succeeds proves the cap+1 read
+// technique does not false-positive: a response exactly AT the cap must
+// still succeed, not be mistaken for oversized.
+func TestMcpBackendCall_ResponseAtOrUnderCap_Succeeds(t *testing.T) {
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		var req jsonrpcRequest
+		_ = json.NewDecoder(r.Body).Decode(&req)
+		result, _ := json.Marshal(map[string]bool{"ok": true})
+		w.Header().Set("Content-Type", "application/json")
+		_ = json.NewEncoder(w).Encode(jsonrpcResponse{ID: req.ID, Result: result})
+	}))
+	defer srv.Close()
+
+	cfg := newFederationTestConfig(srv.URL, "http://beta.invalid", false)
+	h, err := New(context.Background(), http.HandlerFunc(func(http.ResponseWriter, *http.Request) {}), cfg, "llmgw")
+	if err != nil {
+		t.Fatalf("New: %v", err)
+	}
+	gw, ok := h.(*Gateway)
+	if !ok {
+		t.Fatal("handler is not *Gateway")
+	}
+
+	// A generous cap comfortably above this tiny response — must succeed
+	// cleanly, not be flagged as too-large.
+	resp, err := gw.mcpBackendCall(context.Background(), srv.URL, "tools/list", struct{}{}, 4096)
+	if err != nil {
+		t.Fatalf("unexpected error: %v", err)
+	}
+	if resp.Error != nil {
+		t.Fatalf("resp.Error = %+v, want nil", resp.Error)
+	}
+}
+
+// TestHandleMCPFederated_ToolsCall_ResponseBetween4And10MiB_Succeeds is
+// the end-to-end regression for important finding 4: a real tool result
+// between mcpBackendResponseMaxBytes (4MiB, the fan-out-only cap) and
+// mcpBackendCallResponseMaxBytes (10MiB, tools/call's own cap) — legal
+// under the pre-finding-1c budget, and a realistic size for an image or
+// extracted-document tool result — must succeed on the tools/call path,
+// proving that path was never shrunk to the fan-out's smaller cap.
+func TestHandleMCPFederated_ToolsCall_ResponseBetween4And10MiB_Succeeds(t *testing.T) {
+	const payloadBytes = 6 << 20 // 6MiB: over the 4MiB fan-out cap, under the 10MiB tools/call cap
+	big := strings.Repeat("y", payloadBytes)
+
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		var req jsonrpcRequest
+		_ = json.NewDecoder(r.Body).Decode(&req)
+		if req.Method == "tools/call" {
+			result, _ := json.Marshal(map[string]string{"data": big})
+			w.Header().Set("Content-Type", "application/json")
+			_ = json.NewEncoder(w).Encode(jsonrpcResponse{ID: req.ID, Result: result})
+			return
+		}
+		w.Header().Set("Content-Type", "application/json")
+		_ = json.NewEncoder(w).Encode(jsonrpcResponse{ID: req.ID, Result: json.RawMessage(`{}`)})
+	}))
+	defer srv.Close()
+
+	cfg := newFederationTestConfig(srv.URL, "http://beta.invalid", false)
+	h, err := New(context.Background(), http.HandlerFunc(func(http.ResponseWriter, *http.Request) {}), cfg, "llmgw")
+	if err != nil {
+		t.Fatalf("New: %v", err)
+	}
+
+	rec := httptest.NewRecorder()
+	h.ServeHTTP(rec, newFederatedRequest(t, "sk-alice", jsonrpcRequest{
+		JSONRPC: jsonrpcVersion, Method: "tools/call", ID: json.RawMessage("1"),
+		Params: json.RawMessage(`{"name":"alpha_lookup"}`),
+	}))
+	if rec.Code != http.StatusOK {
+		t.Fatalf("status = %d, want 200, body=%s", rec.Code, rec.Body.String())
+	}
+	var got jsonrpcResponse
+	if err := json.Unmarshal(rec.Body.Bytes(), &got); err != nil {
+		t.Fatalf("decode: %v", err)
+	}
+	if got.Error != nil {
+		t.Fatalf("error = %+v, want nil (a 6MiB result must succeed on the tools/call path)", got.Error)
+	}
+}
+
 // TestHandleMCPFederated_ToolsList_HandshakeFallback_SucceedsAfterSessionRetry
 // is MF3's core proof: a server that rejects a bare tools/list call
 // (readitall/fetch's real, live-probed behavior — sessionRequiredMockServer's
@@ -1391,100 +1513,96 @@ func betaHost(rawURL string) (string, error) {
 	return strings.TrimPrefix(rawURL, prefix), nil
 }
 
-// --- security audit finding 1b: tools/list weighted by backend count ---
+// --- security review round 2, critical finding 3: tools/list is NOT weighted ---
 
-// TestHandleMCPFederated_ToolsList_WeightedByBackendCount_RejectsWhenOverBudget
-// proves a tools/list call charges checkAndCount len(allowedMCPServerNames)
-// requests, not one: with 3 configured (and unrestricted) servers and a
-// requests-per-minute budget of 2, the very FIRST tools/list call must be
-// rejected outright — and, because checkAndCountWeighted runs and fails
-// BEFORE the fan-out ever starts, none of the three servers may be
-// contacted at all.
-func TestHandleMCPFederated_ToolsList_WeightedByBackendCount_RejectsWhenOverBudget(t *testing.T) {
-	alpha := newMockJSONRPCServer(t, nil)
-	beta := newMockJSONRPCServer(t, nil)
-	gamma := newMockJSONRPCServer(t, nil)
-	cfg := newFederationTestConfig(alpha.srv.URL, beta.srv.URL, false)
-	cfg.MCPServers["gamma"] = &TargetConfig{URL: gamma.srv.URL}
-	cfg.Groups["default"].Limits = &LimitsConfig{RequestsPerMinute: 2}
+// TestHandleMCPFederated_ToolsList_NotWeightedByBackendCount_BudgetBuysNCalls
+// is the regression for security review round 2's critical finding 3:
+// weighting checkAndCount by the number of allowed MCP servers (security
+// audit finding 1b's original fix) was not default-preserving — on a
+// config where the MCP allow-list is unrestricted (every group on the
+// live fleet), it silently turned a configured requests-per-minute budget
+// into budget/serverCount successful tools/list calls, with NO config
+// change on the operator's part (11 servers, requests-per-minute: 60 ->
+// only 5 calls/minute actually succeeded; examples/kubernetes.yaml's
+// shipped requests-per-minute: 10 would 429 the very FIRST call ever
+// made). With numServers (11, matching the live fleet) configured and a
+// requests-per-minute budget of exactly rpmBudget, all rpmBudget calls
+// must succeed and the next one must be rejected — proving the charge is
+// exactly 1 per call, regardless of how many backends that call fans out
+// to.
+func TestHandleMCPFederated_ToolsList_NotWeightedByBackendCount_BudgetBuysNCalls(t *testing.T) {
+	const numServers = 11 // matches the live fleet's own MCP server count
+	const rpmBudget = 5
+
+	cfg := CreateConfig()
+	cfg.Providers = map[string]*ProviderConfig{"openai": {Type: "openai", APIKey: "k"}}
+	cfg.MCPServers = make(map[string]*TargetConfig, numServers)
+	for i := 0; i < numServers; i++ {
+		srv := newMockJSONRPCServer(t, []mcpTool{{Name: "lookup"}})
+		cfg.MCPServers[fmt.Sprintf("srv%d", i)] = &TargetConfig{URL: srv.srv.URL}
+	}
+	cfg.Groups = map[string]*GroupConfig{"default": {Limits: &LimitsConfig{RequestsPerMinute: rpmBudget}}}
+	cfg.Users = &UsersConfig{Inline: []*UserConfig{{Name: "alice", Group: "default", APIKey: "sk-alice"}}}
 
 	h, err := New(context.Background(), http.HandlerFunc(func(http.ResponseWriter, *http.Request) {}), cfg, "llmgw")
 	if err != nil {
 		t.Fatalf("New: %v", err)
 	}
 
+	for i := 1; i <= rpmBudget; i++ {
+		rec := httptest.NewRecorder()
+		h.ServeHTTP(rec, newFederatedRequest(t, "sk-alice", jsonrpcRequest{JSONRPC: jsonrpcVersion, Method: "tools/list", ID: json.RawMessage(fmt.Sprintf("%d", i))}))
+		if rec.Code != http.StatusOK {
+			t.Fatalf("call %d/%d status = %d, want 200 (budget must buy exactly rpmBudget calls, not rpmBudget/numServers), body=%s", i, rpmBudget, rec.Code, rec.Body.String())
+		}
+	}
+
+	rec := httptest.NewRecorder()
+	h.ServeHTTP(rec, newFederatedRequest(t, "sk-alice", jsonrpcRequest{JSONRPC: jsonrpcVersion, Method: "tools/list", ID: json.RawMessage(`"over"`)}))
+	if rec.Code != http.StatusTooManyRequests {
+		t.Fatalf("call %d status = %d, want 429 (budget exhausted), body=%s", rpmBudget+1, rec.Code, rec.Body.String())
+	}
+}
+
+// TestHandleMCPFederated_ToolsList_AmplificationVisibleAtTargetScopeOnly
+// proves the fan-out's real per-server cost is still fully visible to an
+// operator without silently consuming tenant quota (security review round
+// 2, 2026-08-22, critical finding 3's own ruling): after ONE tools/list
+// call against 2 servers, the user's own requests-per-minute counter
+// reads exactly 1 (never weighted), while EACH server's own "mcp"
+// target-scope counter (countTargetRequests, limits.go) reads exactly 1
+// — the real amplification, attributed where an operator can see it.
+func TestHandleMCPFederated_ToolsList_AmplificationVisibleAtTargetScopeOnly(t *testing.T) {
+	alpha := newMockJSONRPCServer(t, []mcpTool{{Name: "lookup"}})
+	beta := newMockJSONRPCServer(t, []mcpTool{{Name: "search"}})
+	cfg := newFederationTestConfig(alpha.srv.URL, beta.srv.URL, false)
+	h, err := New(context.Background(), http.HandlerFunc(func(http.ResponseWriter, *http.Request) {}), cfg, "llmgw")
+	if err != nil {
+		t.Fatalf("New: %v", err)
+	}
+	gw, ok := h.(*Gateway)
+	if !ok {
+		t.Fatal("handler is not *Gateway")
+	}
+
 	rec := httptest.NewRecorder()
 	h.ServeHTTP(rec, newFederatedRequest(t, "sk-alice", jsonrpcRequest{JSONRPC: jsonrpcVersion, Method: "tools/list", ID: json.RawMessage("1")}))
-	if rec.Code != http.StatusTooManyRequests {
-		t.Fatalf("status = %d, want 429 (weight=3 > limit=2), body=%s", rec.Code, rec.Body.String())
-	}
-	for name, srv := range map[string]*mockJSONRPCServer{"alpha": alpha, "beta": beta, "gamma": gamma} {
-		if srv.called.Load() != 0 {
-			t.Errorf("server %q was contacted %d times, want 0 (rejected before fan-out)", name, srv.called.Load())
-		}
-	}
-}
-
-// TestHandleMCPFederated_ToolsList_WeightedByBackendCount_AccumulatesAcrossCalls
-// proves the weighted charge actually accumulates against the same
-// requests-per-minute window: with 3 servers and a budget of 5, the first
-// tools/list call (weight 3) succeeds, and the second (weight 3 more,
-// total 6 > 5) is rejected.
-func TestHandleMCPFederated_ToolsList_WeightedByBackendCount_AccumulatesAcrossCalls(t *testing.T) {
-	alpha := newMockJSONRPCServer(t, []mcpTool{{Name: "lookup"}})
-	beta := newMockJSONRPCServer(t, []mcpTool{{Name: "search"}})
-	gamma := newMockJSONRPCServer(t, []mcpTool{{Name: "fetch"}})
-	cfg := newFederationTestConfig(alpha.srv.URL, beta.srv.URL, false)
-	cfg.MCPServers["gamma"] = &TargetConfig{URL: gamma.srv.URL}
-	cfg.Groups["default"].Limits = &LimitsConfig{RequestsPerMinute: 5}
-
-	h, err := New(context.Background(), http.HandlerFunc(func(http.ResponseWriter, *http.Request) {}), cfg, "llmgw")
-	if err != nil {
-		t.Fatalf("New: %v", err)
+	if rec.Code != http.StatusOK {
+		t.Fatalf("status = %d, want 200, body=%s", rec.Code, rec.Body.String())
 	}
 
-	rec1 := httptest.NewRecorder()
-	h.ServeHTTP(rec1, newFederatedRequest(t, "sk-alice", jsonrpcRequest{JSONRPC: jsonrpcVersion, Method: "tools/list", ID: json.RawMessage("1")}))
-	if rec1.Code != http.StatusOK {
-		t.Fatalf("first call status = %d, want 200, body=%s", rec1.Code, rec1.Body.String())
+	now := time.Now()
+	userReq, ok := gw.limiter.getCounter("user", "alice", metricReq, windowDay, now)
+	if !ok || userReq != 1 {
+		t.Errorf("user req:day counter = %d (ok=%v), want exactly 1 (never weighted by backend count)", userReq, ok)
 	}
-
-	rec2 := httptest.NewRecorder()
-	h.ServeHTTP(rec2, newFederatedRequest(t, "sk-alice", jsonrpcRequest{JSONRPC: jsonrpcVersion, Method: "tools/list", ID: json.RawMessage("2")}))
-	if rec2.Code != http.StatusTooManyRequests {
-		t.Fatalf("second call status = %d, want 429 (3+3=6 > limit=5), body=%s", rec2.Code, rec2.Body.String())
+	alphaReq, ok := gw.limiter.getCounter(targetKindMCP, "alpha", metricReq, windowDay, now)
+	if !ok || alphaReq != 1 {
+		t.Errorf("mcp/alpha req:day counter = %d (ok=%v), want 1", alphaReq, ok)
 	}
-}
-
-// TestHandleMCPFederated_ToolsCall_WeightStaysOne_EvenWithManyServersConfigured
-// proves tools/call — which only ever resolves to ONE backend — never
-// inherits tools/list's per-backend weighting: with 3 servers configured
-// and a requests-per-minute budget of 2, two sequential tools/call
-// requests must both succeed (weight 1 each, total 2, at the limit), where
-// the equivalent tools/list weighting (3 each) would already have
-// rejected the first one.
-func TestHandleMCPFederated_ToolsCall_WeightStaysOne_EvenWithManyServersConfigured(t *testing.T) {
-	alpha := newMockJSONRPCServer(t, nil)
-	beta := newMockJSONRPCServer(t, nil)
-	gamma := newMockJSONRPCServer(t, nil)
-	cfg := newFederationTestConfig(alpha.srv.URL, beta.srv.URL, false)
-	cfg.MCPServers["gamma"] = &TargetConfig{URL: gamma.srv.URL}
-	cfg.Groups["default"].Limits = &LimitsConfig{RequestsPerMinute: 2}
-
-	h, err := New(context.Background(), http.HandlerFunc(func(http.ResponseWriter, *http.Request) {}), cfg, "llmgw")
-	if err != nil {
-		t.Fatalf("New: %v", err)
-	}
-
-	for i := 1; i <= 2; i++ {
-		rec := httptest.NewRecorder()
-		h.ServeHTTP(rec, newFederatedRequest(t, "sk-alice", jsonrpcRequest{
-			JSONRPC: jsonrpcVersion, Method: "tools/call", ID: json.RawMessage(fmt.Sprintf("%d", i)),
-			Params: json.RawMessage(`{"name":"alpha_lookup"}`),
-		}))
-		if rec.Code != http.StatusOK {
-			t.Fatalf("call %d status = %d, want 200 (tools/call weight must stay 1), body=%s", i, rec.Code, rec.Body.String())
-		}
+	betaReq, ok := gw.limiter.getCounter(targetKindMCP, "beta", metricReq, windowDay, now)
+	if !ok || betaReq != 1 {
+		t.Errorf("mcp/beta req:day counter = %d (ok=%v), want 1", betaReq, ok)
 	}
 }