ClawHub
Back to skill

Security audit

SQL查询引擎

Security checks across malware telemetry and agentic risk

Overview

This SQL guidance skill is not malicious, but it asks for broad execution, write, API, and credential-related authority that is not clearly scoped for a documentation-style SQL helper.

Install only if you are comfortable with an agent receiving broad read/write/exec-style instructions for SQL work. Treat generated SQL as advisory, review every statement before running it, avoid production databases by default, and do not provide API keys unless a specific, trusted service and purpose are clearly identified.

SkillSpector

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)

Intent-Code Divergence

High
Confidence
92% confidence
Finding
The skill claims it cannot connect to a real database, yet elsewhere provides operational and mutating SQL guidance such as UPDATE, ALTER SYSTEM, VACUUM, and transaction examples. This mismatch can mislead an agent or user into treating the skill as low-risk documentation while still enabling dangerous real-environment actions, increasing the chance of unintended data modification or service disruption.

Intent-Code Divergence

Medium
Confidence
95% confidence
Finding
The manifest declares exec/write/read tools and the document advertises command execution, file handling, and API-style capabilities, but the dependency section later labels the skill as pure Markdown only. This inconsistency obscures the real capability surface, which may cause downstream policy engines or users to underestimate the skill's ability to change files or run commands.

Context-Inappropriate Capability

Medium
Confidence
83% confidence
Finding
The skill requests API key configuration and references external API integration without a clear, necessary connection to its stated role as an SQL guidance engine. Unnecessary secret handling expands attack surface by encouraging users or agents to expose credentials for functionality that is not clearly bounded or justified.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The skill describes command execution, file writes, and external/API interaction without a prominent warning that these actions can alter systems, data, or credentials. In the context of a tool-enabled agent, missing safety interlocks increases the likelihood of destructive or privacy-impacting actions being taken without informed consent.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.