Back to skill

Security audit

税务策略师

Security checks for vulnerabilities and agentic risk

Overview

This is a plain Chinese tax-planning advisory skill with no executable code, persistence, credential handling, or hidden data movement, though its activation triggers are broader than ideal.

Install this only if you want a China-focused tax advisory persona. Because it may trigger on broad finance, analysis, report, or budget prompts, confirm that its advice is relevant to your jurisdiction and situation, and verify any tax position against current law or a qualified tax professional before acting.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (6)

Vague Triggers

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The trigger set is overly broad for a high-authority tax strategy skill, making accidental activation likely during ordinary finance discussions. When mis-triggered, the agent may begin giving specialized tax, compliance, or structuring guidance in the wrong context, which can mislead users or cause unintended data exposure in sensitive financial workflows.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

技能名称与描述完全固定为中文,并明确限定在“中国及地方税务法规”框架内,但文档没有说明这是面向特定地区用户的专用技能,也未向用户提供语言或地区适用范围的选择。根据语言/locale 政策,若未明确提供选择或合理限定,强制单一语言/法域语境可能构成自然语言层面的策略违规。

Content

No source excerpt is available for this finding.

Overly Broad Trigger

Low
Category
Trigger Abuse
Confidence
89% confidence
Finding

The trigger '财务' is extremely broad and can match many unrelated requests, causing this specialized skill to activate outside intended tax-planning scenarios. In a finance environment, mistaken activation can inject authoritative tax guidance into general conversations and increase the chance of inappropriate recommendations or unnecessary access to sensitive business context.

Content

No source excerpt is available for this finding.

Overly Broad Trigger

Low
Category
Trigger Abuse
Confidence
90% confidence
Finding

The trigger '分析' is generic enough to match a huge range of ordinary prompts, creating a clear risk of unintended activation. Because the skill presents itself as an expert advisor on compliance-sensitive tax matters, accidental invocation may steer users toward irrelevant or jurisdiction-specific advice without clear consent.

Content

No source excerpt is available for this finding.

Overly Broad Trigger

Low
Category
Trigger Abuse
Confidence
90% confidence
Finding

The trigger '报表' is too short and generic, overlapping with many accounting, BI, and operational reporting requests. This can cause the tax strategy skill to activate when users only want generic reporting help, increasing confusion and the risk of unsolicited tax guidance in sensitive financial contexts.

Content

No source excerpt is available for this finding.

Overly Broad Trigger

Low
Category
Trigger Abuse
Confidence
90% confidence
Finding

The trigger '预算' is broad and commonly used in ordinary planning discussions, so it can unintentionally invoke a powerful tax-planning persona. In a business setting, this may bias budgeting conversations toward tax structuring recommendations when the user did not request them, potentially affecting decision quality or exposing unnecessary sensitive information.

Content

No source excerpt is available for this finding.