Social Hub Server

• Outcome: Review
• Latest audit: May 28, 2026, 8:15 AM
• Version:
  1.0.0

Overview

This is a coherent social-matching skill, but it centralizes sensitive multi-user profile and relationship data with recurring automation and insufficient access, consent, and retention boundaries.

Install only if you intend to run a trusted central matching service. Before use, define who can join the agent group, how agent messages are authenticated, what user data may be stored or sent to embedding/LLM providers, who can read logs, how long data is retained, and how users can consent, review, and delete their data.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (1)

Vague Triggers

Medium

• Confidence: 87% confidence
• Finding: The trigger condition says the skill should run whenever a new group message appears or on scheduled scans, while the skill also emits many non-protocol observability log messages into that same group. This creates a realistic risk of self-triggering or trigger amplification, causing unnecessary processing, repeated matching actions, or feedback loops in a centralized engine with broad access to user profile data.

VirusTotal

66/66 vendors flagged this skill as clean.