openclaw-ops

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The documentation exposes multiple high‑risk capabilities (arbitrary remote command execution on paired nodes, remote browser JS evaluation, camera/screen/location capture, webhook push endpoints, ACP/IDE bridging and exec-approval allowlists) that can be abused for remote code execution, persistent backdoor access, or data exfiltration if misused or if credentials/tokens are leaked.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.90). This skill explicitly treats the agent as a "high-privilege operator" and instructs installing/managing gateway and daemon services (including "裝成 service"), running destructive ops like reset/uninstall and deleting ~/.openclaw — all actions that change system state and can require system-level (sudo) changes.