ClawHub
Back to skill

Security audit

神经记忆增强系统

Security checks across malware telemetry and agentic risk

Overview

This is a local memory skill, but it directs agents to automatically persist conversation-derived facts, preferences, decisions, errors, and full dialogue text without a clear opt-in, deletion, or sensitivity-control flow.

Install only if you want your agent to keep a long-lived local memory of conversations and project decisions. Avoid using it for secrets, regulated data, confidential business discussions, or personal information unless you first establish clear rules for what may be stored, how to review it, and how to delete or isolate memories by project.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)

Missing User Warnings

High
Confidence
97% confidence
Finding
The skill describes automatic storage and processing of conversation content without a clear, prominent warning or consent flow about retention, sensitivity, and persistence. In an agent context, this can lead to silent collection of personal data, secrets, internal decisions, or regulated information into long-lived local storage without the user realizing it.

Ssd 3

Medium
Confidence
98% confidence
Finding
The skill explicitly instructs the agent to automatically extract and persist user preferences, decisions, facts, errors, and even full dialogue text as memory. This creates a real privacy and data-governance risk because it encourages indiscriminate retention of potentially sensitive content, expands the blast radius of prompt injection or accidental disclosure, and can make later recall surface confidential information in unrelated contexts.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.