ClawHub

mnemospark memory-backup

Security checks across malware telemetry and agentic risk

Overview

This is a transparent memory-backup skill that handles sensitive agent memory, but its access is disclosed, purpose-aligned, and gated by user confirmation.

Before installing, review the proposed included paths carefully because agent memory can contain private or sensitive context. Prefer local backup unless you trust mnemospark, and only approve the optional unsafe-force mnemospark install or cloud upload if you understand that separate plugin and destination.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Rogue AgentSelf-Modification, Session Persistence
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

Lp3

Medium
Category
MCP Least Privilege
Confidence
83% confidence
Finding
The skill instructs the agent to read environment-scoped configuration and local filesystem contents (`~/.openclaw/openclaw.json`, workspace files, plugin-owned config) but declares no permissions. That mismatch creates hidden capability exposure: an orchestrator may permit sensitive file inspection without an explicit trust boundary, increasing the chance of unintended access to secrets or private data during backup discovery.

Session Persistence

Medium
Category
Rogue Agent
Content
# Memory Backup

Create or update backups for OpenClaw memory artifacts only.

Follow this workflow:
Confidence
86% confidence
Finding
Create or update backups for OpenClaw memory artifacts only. Follow this workflow: 1. Inspect memory-related config before touching backup payloads. 2. Validate enabled memory systems against files

VirusTotal

67/67 vendors flagged this skill as clean.

View on VirusTotal