CVE-2026-53866
Description
OpenClaw before 2026.5.12 contains an allowlist bypass in shell inline-command parsing, allowing authenticated operators to execute unapproved commands without intended approval prompts.
AI Insight
LLM-synthesized narrative grounded in this CVE's description and references.
OpenClaw before 2026.5.12 contains an allowlist bypass in shell inline-command parsing, allowing authenticated operators to execute unapproved commands without intended approval prompts.
Vulnerability
OpenClaw before version 2026.5.12 contains an allowlist bypass vulnerability in its shell inline-command parsing. When a command request uses shell inline-command forms, it can route through a parser case that misses the expected allowlist decision, enabling execution of shell content without the intended approval prompts. The vulnerability is present in all versions prior to the patched release [1][2].
Exploitation
An attacker must be an authenticated Gateway operator with the ability to issue command requests. By crafting a command using shell inline-command forms that routes through the missing parser case, the attacker can bypass the allowlist check and execute unapproved shell commands without triggering the intended approval prompt [1][2].
Impact
Successful exploitation allows an authenticated operator to execute shell commands that were not approved by the allowlist, potentially leading to unauthorized data access or system modification. The CVSS v3 score is 8.1 (High) with high confidentiality and integrity impacts [2]. Practical impact depends on configuration and whether lower-trust input can reach the vulnerable path [1].
Mitigation
The vulnerability is fixed in OpenClaw version 2026.5.12 [1]. Until patched, administrators should require approval for shell inline-command forms, keep channel and tool allowlists narrow, avoid sharing a Gateway between mutually untrusted users, and disable the affected feature when not needed [1]. No known exploitation in the wild has been reported in the available references.
AI Insight generated on Jun 16, 2026. Synthesized from this CVE's description and the cited reference URLs; citations are validated against the source bundle.
Affected products
2Patches
0No patches discovered yet.
Vulnerability mechanics
No source-code context for this CVE — mechanics is only generated when we can read the actual fix diff. Without that, the four sections (root cause, attack vector, affected code, fix) would be speculation rather than analysis.
References
2News mentions
0No linked articles in our index yet.