ClawHub
Back to skill

Security audit

hippocampus-memory-engine

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed memory engine, but it asks agents to persist, rebuild, clear, and sync user memory data with too little user control or safety guidance.

Review before installing. This skill is not clearly malicious, but it can cause agents to store conversation content and local memory files into a long-term knowledge base, rebuild/export that data, and run a destructive full-sync workflow. Only use it if you understand the hardcoded local paths, have backups, and are comfortable with what may be written to Obsidian or GitHub-related outputs.

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
92% confidence
Finding
The invocation guidance is overly broad and can match ordinary conversation about memory, note-taking, or end-of-session summaries, causing the skill to trigger when the user did not explicitly request this engine. Because the skill performs persistence and downstream sync actions, accidental activation can lead to unintentional storage and propagation of user content into external memory systems.

Missing User Warnings

High
Confidence
97% confidence
Finding
The skill documents a destructive full-sync path that clears data layers and also supports export/sync to Obsidian and GitHub, but it provides no safety guidance such as confirmation, dry-run, backup, scope preview, or destination review. In practice, this can cause accidental data loss locally and unintended disclosure by pushing conversation-derived content into external repositories or knowledge stores.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.