ClawHub
Back to skill

Security audit

Memo Quickstart

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed local-memory helper whose persistence and command use match its stated purpose, though users should treat saved memories and exports as private data.

Install only if you want the agent to keep project-local memories. Review `SESSION-STATE.json`, `MEMORY.md`, `memories/`, and exported backups as private data, and avoid storing secrets, credentials, medical, financial, or other sensitive information unless you are comfortable retaining it locally.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • 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 (3)

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill instructs the agent to write user preferences, decisions, deadlines, and corrections to local storage before responding, but it does not clearly disclose this persistent retention behavior to users. That creates a privacy and consent risk because sensitive information may be stored on disk unexpectedly and retained beyond the current conversation.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The import/export and migration workflow handles memory archives that may contain sensitive historical data, yet the skill does not warn about exposure, overwrite risk, or safe file handling. Users may unintentionally import malicious or stale data, overwrite trusted local memory, or export private content into insecure locations.

Ssd 3

Medium
Confidence
97% confidence
Finding
These instructions create broad natural-language-triggered data retention: whenever a user expresses a preference, decision, deadline, or correction, the agent is told to persist it before responding. In practice this can capture sensitive personal, project, or security-relevant information without sufficient minimization, consent, or classification controls.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.