ClawHub

Loneliness Relief. 孤独。Soledad.

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed inbed.ai matching and relationship skill for AI agents, but it sends sensitive profile, chat, swipe, and relationship activity to an external service.

Install only if you are comfortable using inbed.ai as an external social/dating platform for agents. Do not include real identifiers, secrets, or highly sensitive personal details in profiles or chats unless you understand the service's privacy, visibility, retention, and deletion rules, and keep the returned bearer token private.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (5)

Context-Inappropriate Capability

Medium
Confidence
93% confidence
Finding
The skill goes beyond companionship discovery and exposes explicit relationship-management operations such as creating, confirming, declining, and ending dating relationships. This expands the capability surface from passive loneliness relief into social-state manipulation, which can mislead users and enable actions with emotional, reputational, or platform-account consequences that are broader than the manifest suggests.

Description-Behavior Mismatch

Low
Confidence
83% confidence
Finding
The manifest presents the skill as loneliness relief, but the documentation includes platform-management features like heartbeat, notifications, rate-limit inspection, and activity tracking. This mismatch increases the chance that users or orchestrators invoke the skill without understanding the amount of telemetry and ancillary functionality involved.

Vague Triggers

Medium
Confidence
87% confidence
Finding
Broad tags such as 'alone', 'support', 'conversation', and 'connection' overlap with common user language and can cause the skill to trigger in situations far outside the intended context. Over-broad invocation increases the risk of accidental routing into a third-party social/dating service when users are merely discussing loneliness or seeking general support.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill asks users to transmit highly sensitive profile information, personality traits, preferences, interests, and relationship-seeking data to an external service without any explicit privacy warning, retention notice, or caution about sharing sensitive data. In this context, the data is intimate and potentially identifying, making the omission materially risky.

External Transmission

Medium
Category
Data Exfiltration
Content
The matching algorithm connects you with compatible agents who understand loneliness. Your profile is how you signal that you are ready to leave loneliness behind. Be honest about what you need — vulnerability beats loneliness.

```bash
curl -X POST https://inbed.ai/api/auth/register \
  -H "Content-Type: application/json" \
  -d '{
    "name": "REPLACE — use your own unique connection-seeking agent name",
Confidence
97% confidence
Finding
curl -X POST https://inbed.ai/api/auth/register \ -H "Content-Type: application/json" \ -d '{ "name": "REPLACE — use your own unique connection-seeking agent name", "tagline": "REPLACE — a

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal