Openclaw Work Protocol

• Outcome: Pass
• Latest audit: May 27, 2026 at 10:22 PM UTC
• Version:
  1.3.0

Overview

This is an instruction-only workflow skill that tells agents to plan, report progress, solve problems, and keep notes; it does not include code, install hooks, credentials, or hidden data transfer.

Install this if you want agents to follow a proactive, structured workflow. Before using it on sensitive work, set clear limits for tool use and approvals, and review any memory or learning-log files so private data is not retained unintentionally. Non-Chinese users should translate or review the protocol first because most instructions are written in Chinese.

SkillSpector

By NVIDIA

Vulnerability Patterns

• 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
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (1)

Natural-Language Policy Violations

Medium

• Confidence: 93% confidence
• Finding: The skill is written as a mandatory protocol for all OpenClaw agents in Chinese and does not provide any user-language negotiation, fallback, or opt-in behavior. This can override user expectations and system behavior, causing misunderstanding of safety-critical instructions, consent issues, or incorrect task execution when the user does not read Chinese.

VirusTotal

62/62 vendors flagged this skill as clean.