ClawHub

LYGO FractalWeaver

Security checks across malware telemetry and agentic risk

Overview

This is a local fractal-image-to-audio skill with disclosed file output and a disclosed but automatic local 3-Brain memory write when that optional component exists.

Install if you want a local creative tool that writes audio/profile outputs and may add summaries to your LYGO/3-Brain memory. Use a clean working folder, review output paths to avoid overwrites, do not place untrusted lyra_brain.py files beside it, and ignore the publisher token/publish instructions unless you maintain the package.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Context-Inappropriate Capability

Medium
Confidence
95% confidence
Finding
The skill performs an out-of-scope side effect by importing a separate memory component and writing a summary into a shared/persistent '3-Brain' store. For a tool whose stated purpose is local image analysis and audio generation, silently propagating user-derived metadata into another system creates an unnecessary data-flow boundary and can leak filenames, workflow context, and generated analysis beyond user expectations.

Description-Behavior Mismatch

Medium
Confidence
95% confidence
Finding
The code grows persistent memory that is unrelated to the core audio-weaving function, and it does so automatically after processing. This undeclared persistence can accumulate user project names, source filenames, and derived summaries in a shared memory layer, violating least surprise and expanding the attack/privacy surface without user consent.

Missing User Warnings

Low
Confidence
77% confidence
Finding
The skill states that it produces files such as WAV, JSON, stems, and MIDI, but does not give a clear upfront warning that running it will write artifacts to disk. While expected for a media-generation tool, missing user-visible notice can still cause confusion, unintended storage use, or accidental overwrites in automated environments. The context makes this less severe than covert exfiltration, but still a transparency and safety issue.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The code adds data to an external memory component without any runtime warning or consent at the operation site, despite this action not being necessary for generating audio output. Silent writes to another subsystem are risky because users may not realize their filenames and derived summaries are being persisted or shared outside the immediate output artifacts.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal