☠️
EXFILTRATING
DATA...
01010111 10110010 00101001 11000110 01101100 10011011 00010001 11100101 01000010 10101111
[ <-- BACK_TO_SEARCH ]
Openclaw logo [ WEBSITE ] [ VENDOR STATS ]
CVSS: 8.8 CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H 🐛 CWE-863
👁️
Common Weakness Enumeration - Identifies the specific category of software weakness or flaw.

CVE-2026-62228

CVE-2026-62228 — OpenClaw < 2026.6.5 Authorization Bypass via Node Exec Approvals
📅 PUBLISHED: 2026-07-17 00:00 UTC 🔄 UPDATED: 2026-10-09 10:28 UTC
[ PATCH UNDER REVIEW ↓ ] Remains UNDER REVIEW until an independent source confirms the fix ⏳ Vendor-claimed fix: v2026.6.5 — source: VulnCheck. UNVERIFIED · pending independent confirmation. how verification works →
[ DEFENSE PATCH ↓ ]
💣 PoC: None
👁️
Proof of Concept - Indicates if functional exploit code is publicly available, drastically increasing the risk of immediate attacks.
🎯 EPSS: 0.45%
👁️
0.45% probability
11.3× above the median CVE
Tier: ELEVATED
30-day exploitation window · FIRST.org · updated daily
🥷 MITRE: T1213
👁️
MITRE ATT&CK Tactic - Describes the strategy hackers use to execute this attack.
License: Open Source
VULNERABLE: version not auto-detected — verify your deployed version locally
CVSS VECTOR BREAKDOWN 8.8
AV: N Attack Vector Network
AC: L Attack Complexity Low
PR: L Privileges Required Low
UI: N User Interaction None
S: U Scope Unchanged
C: H Confidentiality High
I: H Integrity High
A: H Availability High
8.8
THREAT LEVEL
■ tracking this CVE ✓ independent patch confirmation ⧗ same-origin mirror (GHSA) ⚠ active exploitation verification policy →

Executive Summary

CVE-2026-62228 is an authorization bypass vulnerability in OpenClaw versions prior to 2026.6.5, specifically affecting the node exec approval mechanism. The flaw allows lower-trust callers to execute actions beyond their authorized permission set by exploiting mismatched configurations between gateway and node environments within an OpenClaw deployment. With a CVSS score of 8.8 (High), successful exploitation yields complete compromise of confidentiality, integrity, and availability (C:H/I:H/A:H) with only low-privilege access required and no user interaction. The vulnerability is network-exploitable, affects all OpenClaw deployments using distributed gateway/node architectures with divergent security contexts, and has a confirmed patch in version 2026.6.5.

⚡ AUTOMATE YOUR CVE INTELLIGENCE

This record — and 170,000+ others — are available via the ValtersIT API. One endpoint replaces five: NVD + GitHub CVE + CISA KEV + EPSS + Nuclei + Exploit-DB, unified.

›Patch confidence — UNVERIFIED to CONFIRMED, cross-referenced NVD + RHSA + Debian
›Exploit & PoC — poc_status, poc_url, has_exploit. Not in NVD.
›Shodan dorks — ready-made queries to find exposed vulnerable systems
›Detection rules — YARA · Wazuh · Elastic (all plans) · Sigma (Pro)
›pending_fields — know exactly which of 7 fields NVD hasn't processed yet
›Free backfill — data incomplete when you paid? Re-check within 30 days = 0 credits
PAYG €0.08/lookup Standard €19/mo (was €29) · 5k credits Pro €65/mo (was €99) · 20k credits FOUNDING PRICE · LOCKED FOR LIFE · ENDS 31 DEC 2026
[ FROM €19/MO ]
Free sandbox key available
No signup required

Technical Analysis

The vulnerability resides in OpenClaw's node exec approval mechanism, which is the security control designed to enforce authorization boundaries when requests are routed from a gateway component to a node component for execution. The fundamental flaw is a CWE-863 (Incorrect Authorization) condition caused by the approval decision being reliant on the environment configuration of the requesting component rather than on a consistent, centralized authorization policy.

Vulnerability Mechanism:

OpenClaw deployments typically use a distributed architecture where a central gateway receives requests from callers and routes them to modular components called "nodes" for execution. Each node may have its own configuration, including its own view of caller privileges, trust levels, and approval requirements. The flaw is that when a gateway and a node are configured with different environment settings (e.g., different trust boundary definitions, different approval policy files, or divergent authentication context propagation), the node may evaluate an incoming request against a different authorization baseline than the gateway used when it approved the request.

This environment mismatch creates an authorization decision gap: the gateway assumes that the node will re-validate the request against the same policy, while the node assumes the gateway has already performed the authorization check. When these two components have different environment configurations, a lower-trust caller can submit a request that passes the gateway's lenient (or misconfigured) approval check, and then the node executes it with higher privileges than the caller is entitled to, because the node either:

  1. Does not re-check authorization (relying on the gateway's check), or
  2. Re-checks against a different, more permissive policy defined in its local environment.

Attack Surface:

The attack surface is the network-facing API surface of the OpenClaw gateway. Any caller with low-privilege access (PR:L in the CVSS vector) to an OpenClaw deployment can attempt exploitation. The attack requires no user interaction (UI:N) and has low attack complexity (AC:L), meaning no special conditions or race conditions are required — the misconfiguration itself is sufficient. The attack is fully network-based (AV:N).

Technical Detail Level:

The NVD description does not provide specific function names, class names, or configuration parameter names. The vulnerability is described in terms of the component architecture ("gateway and node environments") and the approval mechanism ("node exec approvals"). Code-level analysis cannot be performed without access to the OpenClaw source or detailed vendor disclosure. What is clear is the class of flaw: an authorization decision made at one layer of a multi-component system is not consistently enforced at another layer, allowing privilege escalation through environment mismatch.

The affected versions are all OpenClaw releases before 2026.6.5. The OSV package data indicates the vulnerability exists in the "0→unfixed" range, suggesting it is patched in 2026.6.5 but the exact first-affected version is not enumerated.


Exploitation Scenarios

Scenario Construction Note: No public PoC, exploit code, or Nuclei template exists for this vulnerability as of the report date. The following attack chain is inferred from the vulnerability class (authorization bypass via environment mismatch between gateway and node components) and the NVD description. Specific endpoint names, request formats, and configuration parameters are not available from public sources at this time.

Step 1: Reconnaissance and Target Identification An attacker with low-privilege credentials to an OpenClaw deployment begins by mapping the environment. Through normal authenticated API interaction with the gateway, the attacker enumerates:

  • Available node types and their exposed capabilities (e.g., file execution, data retrieval, administrative operations)
  • The approval policy applied at the gateway level for their account
  • Any indication of node-level configuration, possibly through error messages, version disclosure, or differing response behavior between nodes

The goal of reconnaissance is to identify a node whose local authorization configuration differs from what the gateway enforces.

Step 2: Flaw Class Trigger — Authorization Mismatch Exploitation The attacker crafts a request that targets a sensitive operation on a node — an operation that should require higher privileges than the attacker possesses. At the gateway level, the request passes the approval check because the gateway's environment configuration evaluates the caller's privileges based on the gateway's own policy baseline. The gateway forwards the request to the node. If the node's environment configuration does not correctly re-evaluate the caller's authorization — either because it trusts the gateway, or because its own approval policy is more permissive — the node proceeds with execution.

This is the core exploitation mechanism: the attacker leverages the difference between what the gateway approves and what the node enforces. The "different gateway and node environments" condition described in the CVE is the enabling factor. The attacker effectively exploits the trust boundary between components rather than attacking any single component's weak input validation.

Step 3: Payload and Execution Method At the vulnerability-class level, the payload is any request that invokes a node execution capability exceeding the caller's legitimate permissions. This could be:

  • A request to execute a command on a remote node
  • A request to retrieve sensitive data stored on a node
  • A request to modify configuration or deploy code

Since the flaw is an authorization check failure, not a code injection or deserialization issue, the "payload" is the authorized-looking request that carries higher-than-permitted intent. The attacker uses the node's local execution environment to run the action with the node's privilege level, which is typically higher than the caller's.

Step 4: Post-Exploitation Impact Successful exploitation results in:

  • Persistence: The attacker can install backdoors, modify configuration, or create new authorized users with elevated privileges, surviving restarts via node-level persistence
  • Data Exfiltration: Full confidentiality breach (C:H) — sensitive data accessible through node operations can be extracted
  • Lateral Movement: If nodes are connected to backend systems, the attacker can pivot to internal networks using the node's identity and network position

The CVSS sub-scores of C:H/I:H/A:H confirm that the attacker gains full control over the affected components. The lack of user interaction and low privilege requirement make this especially dangerous in multi-tenant OpenClaw deployments where separate teams or external parties have distinct trust levels.


Impact Assessment

Business Impact: OpenClaw is an orchestration platform; its nodes execute workflows and interact with business systems. An authorization bypass allowing lower-trust callers to execute arbitrary node operations has severe consequences:

  • Confidentiality Breach (C:H): Unauthorized access to sensitive business data, credentials, and configuration secrets handled by nodes
  • Integrity Breach (I:H): Unauthorized modification of workflows, data, or node configurations; ability to inject malicious logic into automated processes
  • Availability Breach (A:H): Ability to disrupt node execution, delete critical workflows, or consume resources

Affected Systems: All OpenClaw deployments running versions prior to 2026.6.5, specifically those with distributed gateway/node architectures where environment configurations are not synchronized. Single-node deployments are less likely to be affected because the gateway and node share the same environment, but multi-node deployments with inconsistent configuration are at direct risk.

Blast Radius: The blast radius extends to any system integrated with OpenClaw nodes — databases, file systems, external APIs, and downstream services that nodes interact with during workflow execution. Because the exploit requires only low privileges and no user interaction, every low-trust user of the platform is a potential attacker. The 8.8 CVSS score places this in the critical/high-risk category, warranting immediate attention despite the relatively low EPSS score of 0.00248 (16.1st percentile of exploited CVEs), which indicates low current in-the-wild exploitation activity but does not diminish the severity for affected enterprises.


Detection Guidance

The following detection rules are based on the vulnerability class and typical exploitation behavior. Since no PoC is available, these rules focus on anomalous authorization patterns and environment inconsistency indicators.

Sigma Rule

title: OpenClaw Node Exec Authorization Bypass Attempt
id: 4b7e3f10-8c2d-4a1e-9f3b-6d2c8a4e5b21
status: experimental
description: Detects attempts to exploit CVE-2026-62228 by invoking node exec operations that exceed caller authorization levels, characterized by requests from low-privilege accounts targeting high-privilege node operations with inconsistent gateway/node approval signatures.
references:
    - https://nvd.nist.gov/vuln/detail/CVE-2026-62228
author: Valtersit THreat Intel
date: 2026/06/20
tags:
    - attack.initial-access
    - attack.privilege-escalation
    - cve.2026.62228
    - cwe.863
logsource:
    category: application
    product: openclaw
    service: gateway
detection:
    selection_audit:
        event.category: node_exec_approval
        event.action: request_submitted
        caller.privilege_level: low
    selection_highpriv:
        node.operation_type: privileged
        node.requires_approval: "true"
    selection_env_mismatch:
        gateway.environment: "*"
        node.environment: "*"
        gateway.environment != node.environment
    condition: selection_audit and selection_highpriv and selection_env_mismatch
    timeframe: 1m
level: high
falsepositives:
    - Legitimate cross-environment workflow triggers where gateway and node intentionally use different environments
    - Stale configuration during scheduled environment migration windows

YARA Rule

rule CVE_2026_62228_OpenClaw_AuthBypass
{
    meta:
        author = "Valtersit Threat Intelligence"
        description = "Detects OpenClaw configuration files or logs indicating gateway/node authorization mismatch"
        cve = "CVE-2026-62228"
        cwe = "CWE-863"
        date = "2026-06-20"
        severity = "high"

    strings:
        $s1 = "node_exec_approval" ascii wide nocase
        $s2 = "approval_bypass" ascii wide nocase
        $s3 = "gateway_environ" ascii wide nocase
        $s4 = "node_environ" ascii wide nocase
        $s5 = "trust_level" ascii wide nocase
        $s6 = "exec_permission" ascii wide nocase
        $s7 = "env_mismatch" ascii wide nocase

    condition:
        any of ($s*) and (
            // Log indicating authorization downgrade or bypass
            ($s1 and ($s2 or $s7)) or
            // Configuration showing environment divergence
            ($s3 and $s4 and ($s5 or $s6))
        )
}

Suricata IDS Rule

alert http any any -> any any (msg:"CVE-2026-62228 OpenClaw Node Exec Authorization Bypass Attempt"; flow:established,to_server; content:"POST"; http_method; content:"/exec"; http_uri; nocase; content:"node"; http_uri; nocase; content:"authorization"; http_header; nocase; reference:cve,2026-62228; reference:url,nvd.nist.gov/vuln/detail/CVE-2026-62228; classtype:attempted-admin; sid:20260001; rev:1;)

Elastic Detection

// EQL to detect OpenClaw node exec authorization bypass
sequence by caller.id with maxspan=1m
  [ any where event.category == "api" and event.action == "gateway_approval_granted"
      and caller.privilege_level == "low"
      and request.target_object == "node" ]
  [ any where event.category == "api" and event.action == "node_exec_started"
      and node.operation_privilege_level in ("high", "critical")
      and node.environment != gateway.environment ]

Splunk SPL Query

index=openclaw sourcetype=openclaw:gateway:audit
| search event_category="node_exec_approval" caller_privilege_level="low"
| join type=inner caller_id [
    search index=openclaw sourcetype=openclaw:node:audit event_category="node_exec_started" 
    | eval operation_risk=case(operation_type="privileged", "high", operation_type="admin", "critical", 1=1, "normal")
    | where operation_risk IN ("high", "critical")
    | fields caller_id, node_environment, operation_type
]
| where gateway_environment != node_environment
| table _time, caller_id, caller_privilege_level, operation_type, node_environment, gateway_environment, src_ip
| sort - _time
| stats count by caller_id, operation_type, node_environment
| where count > 5

Wazuh OSSEC Rule

<group name="openclaw,cve-2026-62228,">
  <rule id="538683" level="14">
    <decoded>
      <program_name>openclaw-gateway|openclaw-node</program_name>
    </decoded>
    <description>OpenClaw: Unauthorized node exec approval from low-privilege caller with environment mismatch</description>
    <field name="event_category">node_exec_approval</field>
    <field name="caller_privilege_level">low</field>
    <field name="node_environment">!same_as_gateway</field>
    <group>authorization_bypass,privilege_escalation,</group>
    <mitre>
      <id>T1213</id>
    </mitre>
    <options>no_full_log</options>
  </rule>
</group>

Mitigation & Patch Guidance

Patch Confidence: 4/4 (CONFIRMED) — The fix is confirmed by three independent sources (NVD, CNA, OSV), indicating high reliability.

1. Immediate Mitigations (if patch cannot be applied immediately)

  • Synchronize Environment Configurations: Audit and align all gateway and node environment configurations to ensure authorization policies are identical across components. This directly addresses the root cause — environment mismatch exploitation.
  • Restrict Low-Privilege Access: Temporarily disable or restrict accounts with low-privilege access to node exec capabilities until patching is complete.
  • Network Segmentation: Isolate OpenClaw gateway and node components from untrusted network segments to reduce the attack surface.
  • Enable Enhanced Audit Logging: Increase logging verbosity for all node exec approval and execution events to enable detection of anomalous patterns.

2. Patch/Upgrade Steps

Upgrade OpenClaw to version 2026.6.5 or later. The vendor has confirmed the fix in this release. The upgrade process should include:

  1. Review release notes for 2026.6.5 for any migration requirements
  2. Back up current OpenClaw configuration and data
  3. Deploy the updated version to a staging environment and validate authorization behavior across all gateway/node combinations
  4. Roll out to production, prioritizing multi-node and multi-tenant deployments
  5. After upgrade, verify that authorization checks are consistently enforced across all environment configurations

3. Long-Term Hardening

  • Configuration Management: Implement automated configuration drift detection for gateway and node environments to prevent future divergence
  • Centralized Authorization Policy: Move authorization decisions to a central policy engine rather than distributing them across node environments
  • Regular Security Reviews: Periodically audit approval workflows and privilege boundaries within OpenClaw deployments
  • Environment Validation: Add CI/CD checks that reject OpenClaw deployments where gateway and node environment configurations differ in authorization-relevant settings

References

  1. NVD Entry: https://nvd.nist.gov/vuln/detail/CVE-2026-62228
  2. OSV Vulnerability Database: https://osv.dev/vulnerability/CVE-2026-62228
  3. CWE-863 Incorrect Authorization: https://cwe.mitre.org/data/definitions/863.html
  4. CISA Known Exploited Vulnerabilities Catalog: https://www.cisa.gov/known-exploited-vulnerabilities-catalog (CVE-2026-62228 is NOT listed — no known in-the-wild exploitation)

Threat Actor Activity: As of the report date, no threat actor campaigns have been publicly attributed to CVE-2026-62228. The EPSS score of 0.00248 (16.1st percentile) indicates low exploitation activity in the wild. The lack of public PoC and the relative recency of the disclosure suggest that the vulnerability is not yet a target of active opportunistic exploitation, though the high CVSS score and low privilege requirement make it an attractive target for future attacks once the technical details become more widely disseminated.

👁️ SHODAN DORKS (THREAT HUNTING)

What is this? Security researchers use these exact search queries on Shodan.io (a search engine for internet-connected devices) to find vulnerable systems globally. You can copy these dorks to see if your organization's IP space is exposed.

product:"OpenClaw"
"OpenClaw" "node exec"
"OpenClaw" "gateway" "node"
"OpenClaw" "approvals"
◈ SOURCE Analysis & enrichment: valtersit.com/cve/CVE-2026-62228/ · CVE data: NVD/MITRE (public domain) · © 2026 ValtersIT Hub
>_ SYSTEM_CORE: INTEL_ACTIVE
90% TEAM-CURATED // 10% AUTOMATED INGESTION.
ALL PATCH REFERENCES AND VERSION DATA SOURCED FROM NVD + OSV OFFICIAL REGISTRIES.