ClawHub

Claw Memory

Security checks across malware telemetry and agentic risk

Overview

This is a coherent shared-memory skill, but it can send and persist local memory files to a third-party service with broad activation and weak consent boundaries.

Install only if you intentionally want selected memories stored in this external service. Review and redact MEMORY.md or notes before upload, avoid credentials and highly sensitive personal data, protect the bearer token and encryption key, and prefer a pinned or reviewed install source over a raw main-branch download.

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
90% confidence
Finding
The skill description says to use the service whenever a user asks to store, retrieve, search, or share memories across agent instances, which is broad and lacks clear consent and data-scope boundaries. In a memory-sharing skill, vague activation criteria increase the chance an agent will transmit sensitive conversational or local data to a remote service without an explicit, informed user decision.

Missing User Warnings

Medium
Confidence
98% confidence
Finding
The MEMORY.md import example reads a local memory file and uploads its full contents to an external API, but the skill does not present a prominent privacy warning immediately before doing so. Because memory files often contain secrets, personal data, or prior conversation context, this omission can lead to unintentional bulk exfiltration of highly sensitive local data.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal