OpenClaw Model Router Skill

• Outcome: Pass
• Latest audit: May 28, 2026, 11:35 AM
• Version:
  0.2.1

Overview

The skill appears to perform disclosed model-routing work, with one configuration-related credential exposure risk users should understand.

Install only if you trust the configured OpenClaw binary and the agent executor it calls. Prefer a known absolute binary path, avoid running it with unnecessary secrets in the environment, and review routing configuration before enabling automatic model switching.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (1)

Context-Inappropriate Capability

Medium

• Confidence: 90% confidence
• Finding: The code executes an external binary and forwards the entire parent process environment to it via `env: process.env`. This can expose unrelated secrets, tokens, and credentials to the child process and any downstream tooling it invokes, which is risky when the binary path and behavior are configurable and not tightly constrained.

VirusTotal

64/64 vendors flagged this skill as clean.