Security audit

TaxWise — 智能财税

Security checks for vulnerabilities and agentic risk

Overview

TaxWise appears to be a local tax/bookkeeping helper, but its tax-filing feature overclaims what it calculates for a high-stakes financial workflow.

Review carefully before installing for real tax work. It appears local and non-exfiltrating, but do not rely on its filing, compliance, or planning output as complete tax advice, especially for personal income tax. Avoid adding real OCR, tax-authority, or business-credit API credentials unless you first review and harden the code and secret handling.

SkillSpector

Vulnerability Patterns

MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
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)

Lp3

Medium

Category: MCP Least Privilege
Confidence: 87% confidence
Finding: The skill documents scripts that read from user-supplied file paths and write local data, but it does not declare any permissions for those capabilities. This creates a trust and policy gap: the platform and users are not clearly informed that the skill can access local files, increasing the risk of unintended file exposure or modification if the scripts are invoked with sensitive paths.

Description-Behavior Mismatch

Medium

Confidence: 95% confidence
Finding: The module documentation explicitly claims to calculate and generate drafts for personal income tax, but the implementation only includes VAT, surcharges, and corporate income tax. In a tax-filing skill, this mismatch can cause users to rely on an incomplete filing output and omit a required tax category, creating real compliance and financial risk.

Static analysis

No suspicious patterns detected.