ClawHub
Back to skill

Security audit

agent-architecture-evaluator

Security checks across malware telemetry and agentic risk

Overview

This is a straightforward architecture review skill with an optional local report-formatting script and no evidence of hidden access or unsafe behavior.

Reasonable to install for architecture-review work. Users should avoid including secrets or unnecessary private system details in architecture descriptions, and should run the optional renderer only on JSON inputs and output paths they intentionally choose.

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
86% confidence
Finding
The invocation guidance says to use the file broadly for assessing any agent architecture or multi-agent system, but it does not define clear entry conditions, scope limits, or exclusion criteria. In an agent environment, overly broad routing guidance can cause the skill to be selected in inappropriate contexts, leading to misanalysis, unnecessary disclosure of architectural details, or unsafe reliance on a generic review framework where a narrower specialized skill should be used.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.