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-2026-100566 Overview
CVE-2026-100566 is an access control vulnerability in OpenClaw LINE versions prior to 2026.8.1. The flaw resides in the group allowlist mode, which silently inherits direct message (DM) allowFrom values when groupAllowFrom is not explicitly configured. Attackers who participate in a group can trigger the agent despite configured group allowlist restrictions whenever DM access is broader than the intended group access. The issue is classified as improper authorization under [CWE-863].
Critical Impact
Group participants can bypass intended group allowlist restrictions and invoke the OpenClaw LINE agent when DM configuration is permissive, resulting in unintended agent access across group contexts.
Affected Products
- OpenClaw LINE versions before 2026.8.1
- Deployments relying on groupAllowFrom without explicit configuration
- Instances where DM allowFrom is broader than intended group access
Discovery Timeline
- 2026-09-26 - CVE-2026-100566 published to the National Vulnerability Database
- 2026-09-29 - Last updated in NVD database
Technical Details for CVE-2026-100566
Vulnerability Analysis
OpenClaw LINE enforces access to its agent through separate allowlist configurations for DMs and group chats. When administrators configure a group allowlist but omit the groupAllowFrom setting, the application does not fail closed. Instead, it silently inherits the DM allowFrom list and applies it as the authorization decision for group invocations.
This inheritance occurs without warning, logging, or administrator confirmation. Administrators who assume group access is restricted by their explicit group allowlist mode receive a different, broader policy than intended. The vulnerability is a logic flaw in the authorization resolution path, not a memory or injection issue.
Root Cause
The root cause is improper authorization logic ([CWE-863]) in how OpenClaw LINE resolves allowlist policy. When groupAllowFrom is unset, the code path falls back to the DM allowFrom value rather than enforcing a deny-by-default stance for the group context. The two scopes should be evaluated independently.
Attack Vector
An attacker must be a participant in a group where OpenClaw LINE is active. If the operator's DM allowFrom list includes the attacker, or is more permissive than the intended group policy, the attacker can trigger the agent inside the group even when the group allowlist should block them. The attack requires no privileges beyond group membership and no user interaction from the operator.
See the GitHub Security Advisory GHSA-fgwg-c3wv-8f45 and the VulnCheck Advisory for OpenClaw for technical details.
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 AttacksDetection Methods for CVE-2026-100566
Indicators of Compromise
- Agent invocation events in group chats from identities not listed in the intended groupAllowFrom policy.
- OpenClaw LINE deployments running a version earlier than 2026.8.1 with group allowlist mode enabled and no explicit groupAllowFrom value.
- Unexpected agent responses in group contexts where the operator expected the request to be rejected.
Detection Strategies
- Audit OpenClaw LINE configuration files for group allowlist mode entries that omit groupAllowFrom.
- Correlate agent trigger logs against the explicit group allowlist to flag any invocation originating from a non-listed identity.
- Review historical agent activity in groups to identify prior unauthorized invocations that occurred under the inherited DM policy.
Monitoring Recommendations
- Alert on any OpenClaw LINE agent invocation that references a user not present in the configured groupAllowFrom list.
- Track configuration drift so that upgrades or restarts cannot silently reintroduce the inherited policy.
- Monitor version metadata and flag any OpenClaw LINE instance still running a release prior to 2026.8.1.
How to Mitigate CVE-2026-100566
Immediate Actions Required
- Upgrade OpenClaw LINE to version 2026.8.1 or later on all deployments.
- Explicitly set groupAllowFrom in every configuration that enables group allowlist mode, even when the intended list is empty.
- Review and tighten DM allowFrom values so that any inherited policy would still be acceptable.
Patch Information
The vendor fixed the inheritance behavior in OpenClaw LINE 2026.8.1. Patched versions evaluate groupAllowFrom independently and no longer fall back to DM allowFrom when the group value is omitted. Refer to GitHub Security Advisory GHSA-fgwg-c3wv-8f45 for release details.
Workarounds
- Disable group allowlist mode until the patched version can be deployed.
- Set groupAllowFrom to an explicit, minimal list of trusted identities to prevent inheritance from DM configuration.
- Remove the OpenClaw LINE agent from groups containing untrusted participants until the upgrade is complete.
# Configuration example: enforce explicit group allowlist
# Replace with the identities authorized to invoke the agent in group chats
groupAllowFrom:
- trusted-user-1
- trusted-user-2
# Ensure DM scope is also minimized
allowFrom:
- trusted-user-1
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

