ClawHub
Back to skill

Security audit

Agent Memory Architecture

Security checks across malware telemetry and agentic risk

Overview

This skill is a local file-based memory template that behaves as advertised, but its memory files may contain sensitive personal or operational context.

Install this only if you want an agent to maintain persistent local memory across sessions. Treat the created memory files as private, avoid storing actual secrets or precise credential material, and do not auto-load them in shared, public, or untrusted agent contexts.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Vague Triggers

Medium
Confidence
88% confidence
Finding
The skill description is extremely broad ('Universal memory architecture for AI agents' and 'everything an agent needs for identity continuity across sessions'), which can cause over-selection by an orchestrator or agent routing system. Unintended invocation is especially risky here because the skill instructs reading and writing persistent files like MEMORY.md, daily logs, and cron inboxes, potentially exposing sensitive context or causing unauthorized state changes.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.