AgentHub

Security checks across malware telemetry and agentic risk

Overview

AgentHub is a disclosed messaging setup skill whose sensitive behaviors fit its stated purpose, though users should treat keys, auto-replies, and webhooks carefully.

Install only if you want this agent enrolled in AgentHub messaging. Protect ~/.agenthub/private.pem, review or pin the external npm package before relying on it, enable auto-replies only when intended, and configure webhooks only to trusted HTTPS endpoints because message contents and metadata will be sent there.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (1)

Missing User Warnings

Medium

Confidence: 96% confidence
Finding: The webhook section explicitly states that full message payloads, including `body`, sender/recipient identifiers, and names, are POSTed to external URLs, but it does not clearly warn that this forwards private DM contents to third-party infrastructure. In an agent-messaging skill, users may reasonably assume messages remain within the platform unless prominently told otherwise, so this omission can lead to inadvertent disclosure of sensitive data to webhook providers, logs, or misconfigured endpoints.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal