ClawHub

Memory Defragmenter

Security checks across malware telemetry and agentic risk

Overview

This is a mostly coherent memory-cleanup skill, but it can read and copy local memory files from broad home-directory paths and its execution behavior is less scoped and less accurate than advertised.

Install only if you are comfortable with a skill reading local agent memory markdown and creating backup copies. Use an explicit --path to a known memory directory, run --plan or --dry-run first, inspect any generated plan, and manually remove .defrag_backup copies if they contain sensitive information.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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)

Context-Inappropriate Capability

Medium
Confidence
94% confidence
Finding
The script defaults to scanning the user's home directory (`~`) and accepts an arbitrary `--path`, then reads all matching `.md` files from several relative locations under that base. In an agent skill context, this can expose unrelated personal or sensitive markdown content if the tool is invoked with a broad path or without explicit scoping, creating unintended local file disclosure.

Vague Triggers

Medium
Confidence
82% confidence
Finding
The trigger conditions are broad enough that ordinary phrases like 'clean', 'optimize', or 'maintenance' could invoke a skill that analyzes and modifies memory files. Over-broad activation increases the chance of the skill running in contexts where the user did not intend filesystem changes, which is more dangerous here because the skill can write, archive, and reorganize persistent memory.

VirusTotal

58/58 vendors flagged this skill as clean.

View on VirusTotal