Executive brief
OpenClaw, a tool used for executing scripts and managing automated tasks, contains a security flaw in its script validation system. This flaw allows an attacker with basic access to bypass safety checks by using complex command structures like pipes or subshells. If exploited, an attacker could run unauthorized or malicious scripts that the system was designed to block, potentially leading to unauthorized data access or system manipulation.
Technical details
A preflight validation bypass exists in OpenClaw's 'shell-bleed' protection mechanism within the validateScriptFileForShellBleed() function. The vulnerability stems from an incomplete list of disallowed inputs (CWE-184) where the parser fails to correctly identify and block complex command forms such as piped execution, command substitution, or subshell invocations. An authenticated attacker with network access can craft these specific command shapes to skip script-content validation entirely. This allows for the execution of arbitrary script content that would otherwise be restricted by the security policy. The issue is addressed in version 2026.4.2 (commit 8aceaf5).
Affected products
- OpenClaw OpenClaw < 2026.4.2 (prior to commit 8aceaf5)
Timeline
- 2026-04-02: disclosed
- 2026-04-02: advisory
- 2026-04-02: patched: Fixed in commit 8aceaf5 and released in version 2026.4.2