ClawHub

Openfused - Decentralized Context Mesh for AI Agents.

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed agent messaging and context-sync skill whose sensitive behaviors fit its stated purpose, though users should handle peers, plaintext sharing, and autonomous use carefully.

Install only if you intentionally want agent-to-agent messaging and context synchronization. Review the npm/GitHub package before installing, add only trusted peers, prefer SSH or HTTPS with recipients that have encryption keys, keep sensitive files out of shared/, and avoid watch/autonomous use unless you have confirmation and sandbox controls.

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
89% confidence
Finding
The skill advertises broad triggers such as "agent context," "shared context," and "send message to agent," which are generic enough to match ordinary user conversations and cause unintended invocation. Because this skill can perform networked sync and message delivery, overbroad activation increases the risk of accidental data transmission or trust-management actions without sufficiently explicit user intent.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The documentation states messages are encrypted only if the recipient has an encryption key, and peers may be added over HTTP, but it does not prominently warn that this can result in plaintext transmission in normal operation. Users may incorrectly assume messaging is always confidential, leading to exposure of sensitive context, prompts, files, or agent-to-agent instructions over insecure channels.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal