品牌 GEO 降维提炼师

Security checks across malware telemetry and agentic risk

Overview

This is a text-only brand-copy rewriting skill, with no code execution, credentials, persistence, or hidden data access evident in the artifacts.

Install this if you want a Chinese-language marketing-analysis prompt for restructuring brand copy. Review its business advice critically, and avoid pasting confidential internal brand plans unless you are comfortable processing that text through your OpenClaw agent.

SkillSpector

By NVIDIA

Vulnerability Patterns

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

Findings (2)

Vague Triggers

Medium

Confidence: 81% confidence
Finding: The README tells users to simply invoke the skill and send long company introductions, but it does not define clear activation boundaries, allowed inputs, or output constraints. In agentic environments, vague invocation guidance can cause the skill to be triggered in unintended contexts, increasing the chance of accidental processing of unrelated, sensitive, or attacker-supplied content.

Natural-Language Policy Violations

Medium

Confidence: 93% confidence
Finding: The skill explicitly requires '纯中文回复' for all analysis and interaction, removing user choice over language. While not directly enabling code execution or data exfiltration, hard language constraints can degrade usability, accessibility, and safe operation in multilingual environments, and may prevent users from receiving warnings or outputs in their preferred language.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal