Back to skill

Security audit

个人财务管理助手

Security checks for vulnerabilities and agentic risk

Overview

The skill mostly performs local financial CSV analysis, but it also documents recurring Feishu report delivery in a way that conflicts with its offline/no-network privacy claims.

Review this skill before installing. Its local CSV analysis is straightforward, but do not enable or copy the Feishu cron report configuration unless you have explicit control over the schedule, recipient, redaction, and removal process. Use cautious output paths, avoid overwriting existing files, and open exported CSVs with spreadsheet formula execution disabled when possible.

Vulnerability Patterns
  • System PersistenceInstalls backdoors, hooks, services, or scheduled tasks that survive the run
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (3)

T06 · System Persistence

Error
Location
SKILL.md:274
Finding

Persistent Scheduled Transmission of Sensitive Financial Reports

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
personal-finance.sh:285
Finding

Spreadsheet Formula Injection in Categorized CSV Exports

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
personal-finance.sh:279
Finding

Unsafe Output Validation Allows File Overwrite and Symlink Attacks

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • System Prompt LeakageDirect Leakage, Indirect Extraction, Tool-Based Exfiltration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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
Findings (9)

Tp4

High
Category
MCP Tool Poisoning
Confidence
86% confidence
Finding

大体上,这段代码确实属于个人财务/银行账单分析工具:它读取交易CSV,验证字段与金额、按时间周期汇总、按规则分类,并生成简要统计报表,因此与“银行账单分析、消费分类、财务报表”基本一致。但声明中的“预算管理”和“储蓄建议”是明显未实现的核心能力,属于功能描述高于实际行为。此外,虽然默认是只读,但 categorize 命令在提供 --output 时会写出文件,这是一项应当在描述或权限层面更明确说明的能力。综合来看,描述与实际存在实质性不一致,但不是完全不相关,而是部分功能缺失与少量未声明写文件行为。

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill claims offline, no-network processing, yet also defines scheduled Feishu report delivery. In a personal-finance context, this contradiction is dangerous because monthly reports and budget alerts may expose sensitive banking and spending data to an external messaging service despite user expectations of fully local processing.

Content

No source excerpt is available for this finding.

Direct Prompt Extraction

High
Category
System Prompt Leakage
Confidence
85% confidence
Finding

Skill contains instructions that could directly expose system prompts, internal rules, or hidden instructions to users or external parties.

Content

Scanner excerpt · personal-finance.sh (reported line 257)May include surrounding context.

sh
for rule in rules:
        for keyword in rule["keywords"]:
            if keyword in text:
                return rule["category"]
    return "Uncategorized"

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The manifest description forces a single language/locale presentation without offering an alternative or documenting that the skill is intentionally limited to Chinese-speaking users. Under the stated policy, language-specific behavior should be opt-in or clearly justified.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
83% confidence
Finding

The markdown content consistently assumes a Chinese locale for all usage instructions and outputs, and there is no opt-in, fallback, or justification for this restriction. This can violate language/locale policy when users are not given a choice.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

Contradictory security and privacy claims can mislead users into providing highly sensitive financial data under false assumptions. Even if the network push is intended, inconsistent documentation undermines informed consent and increases the chance of accidental data exposure through Feishu integrations or scheduled automation.

Content

No source excerpt is available for this finding.

Vague Triggers

Low
Category
Not specified by scanner
Confidence
91% confidence
Finding

This JSON manifest-like file defines category-matching keywords, so SQP-1 applies. The keyword "market" is highly generic and can match many everyday descriptions unrelated to groceries, making the trigger scope overly broad and increasing the chance of unintended invocation or misclassification.

Content

No source excerpt is available for this finding.

Vague Triggers

Low
Category
Not specified by scanner
Confidence
94% confidence
Finding

The standalone keyword "deposit" is not specific to payroll or income and could match many unrelated financial events such as transfers, savings deposits, or merchant deposits. Because the trigger list provides no narrowing constraints or negative examples, the activation condition for this category is too broad.

Content

No source excerpt is available for this finding.