CVE-2026-53831: OpenClaw's POSIX node system.run safe-bin allowlist could be widened by shell expansion

OpenClaw can auto-approve system.run commands through a safe-bin or allowlist-style exec policy. The check validated the argv as written. The shell then expanded a value, so something that looked like one safe-bin argument turned into extra shell words and became a file operand. This is a check-then-use gap (CWE-367) in argv validation.

The attacker needs a foothold: a paired POSIX node, and an operator or agent flow that can reach system.run under a safe-bin policy. The advisory says this is not an unauthenticated node takeover. The CVSS 3.1 vector reflects it with low privileges and high attack complexity.

The gain is file read. Depending on what the node process can access, the output may expose OpenClaw configuration data or other node-local information, such as tokens. For teams that let agents run commands on nodes, the risk is that a prompt-influenced agent turns an auto-approved command into a read of a file nobody approved. The advisory states this is a policy gap, not a claim that every safe-bin command is unsafe.

The npm package openclaw before 2026.5.18, on deployments where all of these hold:

• a POSIX node is paired to the gateway
• system.run is reachable by an authenticated operator or agent flow
• exec policy uses safe-bin or allowlist-based auto-approval
• the approved command contains shell-expanded values that can change argv shape

npm ls openclaw

Check the gateway and every node host. Versions below 2026.5.18 are affected.

Upgrade to [email protected] or later, the first stable patched version:

npm install [email protected]

Before you can upgrade, the advisory says to avoid broad safe-bin auto-approval for commands that can read arbitrary paths, and to prefer explicit approval for node commands that touch local files.

Review system.run history on POSIX nodes for approved safe-bin commands whose arguments contain shell metacharacters or variable and command substitutions, such as $, backticks or globs. Look for safe-bin commands such as file readers that touched OpenClaw config paths or other files outside the expected working directory. The sources give no specific log string. Beam records agent shell commands and file access on the machine, and its rules can flag reads of credential-like files.

Summary

On POSIX nodes, OpenClaw's system.run safe-bin checks could approve a command before shell expansion changed how the command was interpreted. A value that appeared to be a safe-bin argument could expand into additional shell words and become a file operand.

This issue is limited to paired POSIX node execution through system.run with safe-bin or allowlist-style auto-approval. It is not an unauthenticated node takeover.

Affected configurations

This affects deployments where:

• a POSIX node is paired to the gateway
• system.run is reachable by an authenticated operator or agent flow
• exec policy uses safe-bin or allowlist-based auto-approval
• the approved command contains shell-expanded values that can change argv shape

Impact

A lower-privilege operator flow could cause an approved safe-bin command to read a node-local file that was not intended by the policy. Depending on the local files available to the node process, this could expose OpenClaw configuration data or other node-local information.

The issue is a policy-enforcement gap in argv validation, not a general statement that every safe-bin command is unsafe.

Patched Versions

The first stable patched version is 2026.5.18.

Mitigations

Upgrade to [email protected] or later. Before upgrading, avoid broad safe-bin auto-approval for commands that can read arbitrary paths, and prefer explicit approval for node commands that touch local files.