Security audit

科研技能鉴赏

Security checks across malware telemetry and agentic risk

Overview

This is a lightweight research-skill directory that routes users to other skills, with broad triggers that may be noisy but no evidence of hidden execution, data access, or persistence.

Install this as a research-skill menu if you want help choosing among related academic tools. Review any specialist skill it recommends before installing, especially those that use SmartLib quotas, upload documents, call meeting services, or manage a knowledge base.

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 (2)

Vague Triggers

Medium

Confidence: 90% confidence
Finding: The trigger list is extremely broad and includes common research-related phrases such as '找论文', '查文献', '组会', and '答辩'. This can cause the routing skill to activate in many ordinary conversations and intercept requests that should go directly to a more specific skill, increasing prompt-surface exposure and the chance of misrouting or inappropriate installation guidance.

Vague Triggers

Medium

Confidence: 84% confidence
Finding: The skill says to first classify any research-related need and then redirect to another skill, but its activation boundary is described at a very high level. In practice, that ambiguity can make the entry skill seize control of requests better handled by specialist skills, creating unnecessary indirection, broader access to user context, and potential workflow confusion or unsafe delegation.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.