ClawHub
Back to skill

Security audit

TencentDB PostgreSQL Skill

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed Tencent Cloud PostgreSQL assistant that can read and change cloud database settings when the user provides credentials and an explicit target.

Install only if you are comfortable giving the host process Tencent Cloud PostgreSQL permissions. Use a least-privilege CAM subaccount, avoid long-lived high-privilege keys, do not paste secrets into chat, and be precise about region and instance ID before asking it to open, close, or modify services.

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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
Findings (2)

Intent-Code Divergence

Medium
Confidence
93% confidence
Finding
The document defines non-public opens (`EnableWanNet=false`) as a valid success state earlier, but the later guardrail says to never claim open success without a ready state and a usable public access address. That contradiction can cause the agent to misreport successful private-only enablement as failure, leading to unsafe operator decisions, repeated write attempts, or unnecessary exposure changes to obtain a public address.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The skill explicitly authorizes direct execution of OpenMem0Service and CloseMem0Service on the basis of a single imperative sentence, without a separate confirmation step for a state-changing cloud operation. Even though it checks for ambiguity and performs preflight validation, a mistaken, spoofed, or misinterpreted prompt could still trigger service enablement or shutdown on production infrastructure.

Static analysis

No suspicious patterns detected.