ClawHub
Back to skill

Security audit

Biomorphic Memory

Security checks across malware telemetry and agentic risk

Overview

The skill’s memory feature is coherent, but its installer can automatically delete a user-configurable local directory if an update fails.

Review the installer before running it. Do not set BIOMORPHIC_INSTALL_DIR to any directory that contains unrelated files, and consider manually cloning a pinned revision or removing the rm -rf fallback. Confirm where memories are stored, how they can be deleted, and what text is sent to the embedding provider before using it with sensitive context.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • 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 (2)

Tp4

High
Category
MCP Tool Poisoning
Confidence
90% confidence
Finding
The skill presents itself as a memory capability, but the installation flow invokes a shell script that can clone/update remote code, install packages, and write into the local environment. This is dangerous because users may treat it as a benign memory feature while actually granting execution of unreviewed remote installer logic, increasing supply-chain and arbitrary code execution risk.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The installer removes the entire installation directory and reclones it automatically when `git pull` fails, without prompting the user or validating what path is being deleted. Because `INSTALL_DIR` is influenced by `BIOMORPHIC_INSTALL_DIR`, a user can accidentally point it at an important directory and lose local data or unrelated files during a routine install/update.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.