ClawHub

database-toolkit

Security checks across malware telemetry and agentic risk

Overview

This is a transparent database utility that can change real databases, so it is appropriate only for users who intentionally want that capability.

Install only if you want an agent to operate on databases. Use read-only or least-privileged database users where possible, review SQL before writes/deletes, back up important databases first, avoid production root credentials, do not invoke Redis flush operations unless you intend to erase that Redis DB, and pin dependencies in controlled environments.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • 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
Findings (5)

Intent-Code Divergence

Medium
Confidence
99% confidence
Finding
The distributed lock is unsafe because it stores a constant value and releases the lock by blindly deleting the Redis key, without verifying that the caller still owns it. If the lock expires and is reacquired by another worker, a stale owner can delete the new owner's lock, breaking mutual exclusion and enabling race conditions, data corruption, or concurrent execution of critical sections.

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The skill explicitly advertises direct database update and backup operations, and later shows raw SQL execution and insert examples without enforcing read-only mode, authorization checks, scope limits, or strong irreversible-action warnings. In an agent setting, this materially increases the chance of destructive or unintended modification of local databases, data corruption, bulk exfiltration via backup, or misuse of privileged database access.

Missing User Warnings

High
Confidence
90% confidence
Finding
Providing an unrestricted flushdb() helper makes it easy for any caller with access to this wrapper to erase the entire selected Redis database. In agent or plugin contexts, such dangerous convenience methods increase the blast radius of misuse, accidental invocation, or abuse by untrusted consumers.

Unpinned Dependencies

Low
Category
Supply Chain
Content
pandas
# MySQL (optional): pymysql sqlalchemy
Confidence
95% confidence
Finding
pandas

Known Vulnerable Dependency: pandas — 1 advisory(ies): CVE-2020-13091 (** DISPUTED ** pandas through 1.0.3 can unserialize and execute commands from an)

High
Category
Supply Chain
Confidence
73% confidence
Finding
pandas

VirusTotal

67/67 vendors flagged this skill as clean.

View on VirusTotal