Safe Config Workflow

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed Chinese-language workflow for safely managing OpenClaw configuration, with expected local config edits and checks but no hidden executable code or exfiltration behavior.

Install this only if you want a Chinese-language assistant workflow for OpenClaw config changes. Review every proposed edit and diff before approval, and do not let tokens, auth values, private endpoints, or session details be written into MEMORY.md or status summaries.

SkillSpector

By NVIDIA

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (1)

Vague Triggers

Medium

Confidence: 91% confidence
Finding: The trigger conditions are broad enough to activate on generic requests like modifying configuration, gateway settings, channels, models, sessions, or authentication, which can cause the skill to run outside its intended OpenClaw-specific context. In an agent environment, overbroad matching can misroute unrelated user tasks into a workflow that performs sensitive config actions, increasing the chance of unintended file edits or command execution against the wrong target.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal