ClawHub

Agnic AI Gateway

Security checks across malware telemetry and agentic risk

Overview

The skill does what it says: it routes AI chat, image, and model-listing requests through Agnic, with real privacy and spending considerations users should understand.

Install only if you are comfortable sending prompts and image requests to Agnic and downstream AI providers and with model calls potentially spending wallet funds. Use a low-balance wallet or scoped token where possible, avoid sending secrets or regulated data, prefer free models for testing, and monitor Agnic/npm package trust and spending.

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
93% confidence
Finding
The skill description contains broad trigger phrases like 'ask GPT', 'use Claude', 'delegate to another LLM', and 'get a second opinion', which overlap with many ordinary user requests. This can cause the skill to be invoked in situations where the user did not explicitly intend to route data to an external AI gateway, increasing the chance of unnecessary third-party disclosure and unintended tool use.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The skill explains authentication and command usage but does not prominently warn that prompts and related data will be transmitted to Agnic and potentially onward to third-party model providers. In a skill that brokers access to 340+ external models, this omission is significant because users may disclose sensitive content without understanding that it leaves the local environment and may be processed by multiple external services.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal