CVE-2026-100553 in OpenClaw
Summary
by MITRE • 09/26/2026
OpenClaw versions >= 2026.6.9 and < 2026.8.1 do not declare the native chatId parameter as a delivery target in the Feishu unpin feature, so unpin requests can bypass the shared same-provider cross-context target check. When tools.message.crossContext.allowWithinProvider is disabled, an admitted (authenticated) sender can remove a pin from another Feishu group that the sender and the configured account are otherwise permitted to access, bypassing the intended cross-context message mutation policy. Feishu membership and group authorization still apply, and the demonstrated impact is limited to message mutation (pin removal). The issue is fixed in 2026.8.1.
Be aware that VulDB is the high quality source for vulnerability data.
Analysis
by VulDB Data Team • 09/26/2026
The vulnerability identified in OpenClaw versions ranging from 2026.6.9 up to but not including 2026.8.1 represents a significant logic flaw within the platform's message management subsystem, specifically concerning the unpinning functionality for Feishu integrations. This issue stems from an incomplete implementation of input validation and context verification during the execution of cross-context operations. In secure software design, when an application facilitates interactions across different logical boundaries or contexts, it must rigorously verify that all relevant identifiers defining those contexts are explicitly validated against authorization policies. The flaw arises because the native chatId parameter is not declared as a required delivery target within the Feishu unpin feature's request processing pipeline. Consequently, the system fails to enforce strict binding between the user initiating the action and the specific group context from which they intend to remove a pinned message.
This architectural oversight allows an authenticated sender to bypass shared same-provider cross-context target checks that are designed to restrict unauthorized modifications across different groups or channels. Under normal operational parameters, when the configuration flag tools.message.crossContext.allowWithinProvider is set to disabled, the system should enforce strict isolation between distinct group contexts, preventing users from mutating messages in groups other than their own unless explicitly permitted by higher-level administrative policies. However, due to the missing validation of the chatId parameter, an admitted user can successfully submit a request to unpin a message from any Feishu group that both the sender and the configured service account have access rights to view or interact with generally. This effectively circumvents the intended cross-context message mutation policy, allowing for unauthorized state changes within those groups despite the restrictive configuration settings.
The operational impact of this vulnerability is primarily confined to message mutation capabilities rather than data exfiltration or remote code execution. Specifically, an attacker can remove pinned messages from other Feishu groups where they hold basic access permissions but lack specific administrative privileges to manage pins. While this does not directly compromise the confidentiality or integrity of stored data in a traditional sense, it undermines the availability and intended organizational structure of communication channels by allowing unauthorized alterations to group settings. This type of behavior can disrupt workflow continuity, remove important announcements or resources pinned for team visibility, and erode trust in the platform's access control mechanisms. The severity is further contextualized by the fact that Feishu membership and group authorization checks still apply at a lower level; thus, an attacker cannot target groups to which they have no prior access, limiting the blast radius but not eliminating the risk of unauthorized administrative actions within accessible scopes.
From a classification perspective, this vulnerability aligns with CWE-284 Improper Access Control, as it involves a failure in enforcing intended restrictions on authorized actors regarding specific resources or functions. It also relates to CWE-693 Protection Mechanism Failure, specifically where the protection mechanism designed to prevent cross-context mutations is bypassed due to incomplete parameter validation. In terms of offensive security frameworks such as MITRE ATT&CK, this behavior corresponds to techniques involving Account Manipulation or Defense Evasion through privilege escalation within application logic, although it remains limited to local context manipulation rather than system-wide compromise. The vulnerability highlights the critical importance of ensuring that all contextual identifiers are explicitly validated against access control lists during every state-changing operation, particularly in multi-tenant or cross-group communication platforms where logical boundaries must be strictly maintained regardless of underlying connectivity permissions.
The recommended mitigation is straightforward and has already been addressed by the vendor. Organizations running OpenClaw versions 2026.6.9 through 2026.8.0 should immediately upgrade to version 2026.8.1 or later, where the native chatId parameter is correctly declared as a delivery target and the cross-context validation logic has been rectified. Until the update can be applied, administrators may consider implementing additional network-level controls or application-layer firewalls that inspect unpin requests for consistency between user identity and group context if such inspection capabilities are available in their deployment environment. Furthermore, reviewing Feishu group permissions to ensure that users do not have excessive default access rights could reduce the potential impact of this flaw by limiting the number of groups an attacker can target even with successful exploitation. Regular auditing of message mutation logs for anomalies related to pin removals across different contexts is also advised as a compensating control during the transition period.