Skip to main content
Vulnerability Database/CVE-2026-100598

CVE-2026-100598: OpenClaw npm Auth Bypass Vulnerability

CVE-2026-100598 is an authentication bypass flaw in OpenClaw npm package that incorrectly binds Signal approval reactions, allowing reactions to ordinary messages to trigger host actions. This article covers technical details, affected versions, impact assessment, and mitigation strategies.

Published:

Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.

Sources: National Vulnerability Database, CISA KEV, Exploit-DB

CVE-2026-100598 Overview

CVE-2026-100598 affects the OpenClaw npm package in versions before 2026.7.1. The vulnerability stems from incorrect binding of Signal approval reactions to the wrong message context. When a conversation contains both ordinary outbound text and a pending structured approval request, an approver's reaction to unrelated text can be misinterpreted as approving or denying the pending host action. The flaw maps to CWE-346: Origin Validation Error and is fixed in version 2026.7.1.

Critical Impact

Approval reactions can bind to unrelated outbound messages, causing unintended authorization or denial of pending host actions in OpenClaw-managed workflows.

Affected Products

  • OpenClaw (npm package openclaw) versions prior to 2026.7.1
  • Deployments integrating OpenClaw with Signal-based approval workflows
  • Host environments where OpenClaw executes privileged actions pending approver confirmation

Discovery Timeline

  • 2026-09-26 - CVE-2026-100598 published to NVD
  • 2026-09-28 - Last updated in NVD database

Technical Details for CVE-2026-100598

Vulnerability Analysis

OpenClaw implements an approval model where structured requests are sent through Signal and approvers respond with reactions to confirm or deny host actions. The approval reaction handler fails to correctly bind a reaction to the specific approval message it targets. When unrelated outbound messages share a conversation with a pending approval, the handler can resolve a reaction against the wrong message. An approver reacting to ordinary outbound text may inadvertently approve or deny a pending host action.

The practical impact depends on the pending request, the timing of conversation messages, and the actions available to the OpenClaw process. The flaw does not elevate authority or allow unauthorized approvers to act; it only misroutes reactions issued by legitimate approvers.

Root Cause

The root cause is an origin validation error [CWE-346] in the reaction-to-approval binding logic. The handler does not sufficiently distinguish reactions targeting an approval message from reactions targeting adjacent outbound text in the same Signal conversation. Message identity and context are not strictly enforced when resolving which pending approval a reaction applies to.

Attack Vector

Exploitation requires a conversation containing both a pending structured approval and other outbound text, plus an approver who reacts to the unrelated text. The condition is largely organic rather than remotely triggerable without approver interaction. See the GitHub Security Advisory and the VulnCheck Advisory on OpenClaw for additional technical context.

Detect Identity Attacks the Moment Authentication Fails

Singularity™ Identity surfaces credential misuse, lateral movement, and privilege escalation in real time, so you can contain the breach before it spreads.

Stop Identity Attacks

Detection Methods for CVE-2026-100598

Indicators of Compromise

  • Approval decisions recorded by OpenClaw that do not correspond to approvers' intended reactions.
  • Host actions executed or denied without a clearly matching structured approval request in the Signal conversation log.
  • Reactions attached to ordinary outbound messages immediately preceding or following a pending approval.

Detection Strategies

  • Audit OpenClaw approval logs and correlate each resolved approval against the specific message ID the approver reacted to.
  • Compare the content of the reacted-to message with the pending approval payload to flag mismatches.
  • Review installed OpenClaw package versions across environments and flag any below 2026.7.1.

Monitoring Recommendations

  • Enable verbose logging of approval binding events, including source message IDs and target approval request IDs.
  • Alert on any approval resolved against a message whose body does not match the approval schema.
  • Monitor npm dependency inventories for continued use of vulnerable OpenClaw versions.

How to Mitigate CVE-2026-100598

Immediate Actions Required

  • Upgrade the openclaw npm package to version 2026.7.1 or later across all environments.
  • Review recent OpenClaw-executed host actions for approvals that may have been bound incorrectly.
  • Rotate or revoke any sensitive state changes applied through approvals that cannot be verified against approver intent.

Patch Information

The issue is fixed in OpenClaw version 2026.7.1. The patched release corrects the approval binding logic so that reactions resolve against the intended structured approval message. Refer to the GitHub Security Advisory GHSA-r88x-r7jj-f2cf for release details.

Workarounds

  • Restrict OpenClaw conversations so that pending approvals do not coexist with unrelated outbound text.
  • Require approvers to react only when no other outbound messages are present in the conversation window.
  • Limit the actions available to the OpenClaw process until the patched version is deployed.
bash
# Upgrade OpenClaw to the patched release
npm install openclaw@2026.7.1
npm ls openclaw

Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.

Sources: National Vulnerability Database, CISA KEV, Exploit-DB

Experience the Most Advanced Cybersecurity Platform

See how the world’s most intelligent, autonomous cybersecurity platform can protect your organization today and into the future.