registry / @agent-vm/openclaw-agent-vm-plugin / 0.0.114

@agent-vm/openclaw-agent-vm-plugin@0.0.114

OpenClaw sandbox-backend plugin that delegates execution to a Gondolin-managed VM.

AI Security Review

scanned 19h ago · by lpm-firewall-ai

LPM treats this as warn-only first-party agent extension lifecycle risk. A startup-activated OpenClaw extension registers a Gondolin sandbox backend and authenticated local control routes. It can execute configured SSH sandbox commands at runtime, but no unconsented install-time mutation or exfiltration was found.

Static reason

No blocking static signals were detected.

Trigger

OpenClaw loads the configured plugin at startup; sandbox/tool operations invoke it.

Impact

Provides substantial agent-extension execution capability within its configured VM sandbox.

Mechanism

registered sandbox backend with authenticated gateway-control and SSH execution

Rationale

Source inspection supports a legitimate but high-privilege agent-extension role, not malicious intent. Warn because startup activation and runtime sandbox-command capability create a guarded platform-extension lifecycle risk.

Evidence

package.jsondist/openclaw.plugin.jsondist/index.jsdist/openclaw-plugin-registration-BKalMcx0.jsdist/sdk-validate.mjs/opt/openclaw-sdk/sandbox.js

Decision evidence

public snapshot

AI called this Suspicious at 86.0% confidence as Dangerous Capability with low false-positive risk.

@agent-vm/openclaw-agent-vm-plugin@0.0.114

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code