Security audit

大六壬技能

Security checks for vulnerabilities and agentic risk

Overview

This skill does not show malicious system access, but it should be reviewed because its advertised commands and capabilities do not match the shipped files and it includes consequential health-style use cases without safeguards.

Review this package before installing. It does not appear to seek credentials, network access, persistence, or privileged control, but users should expect it may not run as documented and should treat any divination output as entertainment only, not medical, legal, financial, or other professional advice.

A.I.G

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

SkillSpector

Vulnerability Patterns

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
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (5)

Tp4

High

Category: MCP Tool Poisoning
Confidence: 97% confidence
Finding: 从安全与能力范围看，代码没有出现未声明的敏感能力：没有网络访问、文件读写、系统调用、外部触发器或权限滥用，整体主目的与声明基本一致，仍属于占卜/排盘算法。之所以判定为 mismatch，是因为声明描述了较完整的“大六壬排盘与断卦”能力，而实际代码只实现了大量“简化版”逻辑，和声明的能力精度存在实质偏差：1) 月将加时函数实际上未使用 shiChen 参与计算，只是按月将顺排天盘；2) 四课、三传、天将、断课分析均为简化占位实现，不符合“完整支持”的表述；3) 导出的 getJiuZongMen / getShenSha 虽存在，但前者依赖不存在的字段（siKe.some(ke => ke.ke === true)），后者将 riGan 当天干使用，而主流程中 riGan 实际取自地支数组；4) getShenSha 中“吉凶': '中'”存在明显语法错误，代码整体可能无法正常加载。综上，代码主题与描述相关，但描述高估了实现完整度与可用性，因此构成描述与实际行为不一致。

Missing User Warnings

Medium

Confidence: 96% confidence
Finding: The skill explicitly supports guidance on illness, travel, career, relationships, and other high-stakes matters, including health-related questions like '这个病严重吗', yet provides deterministic-sounding recommendations without any disclaimer or referral to qualified professionals. This can mislead users into relying on divination outputs for consequential decisions, increasing the risk of delayed medical care or other harmful real-world choices.

Natural-Language Policy Violations

Medium

Confidence: 89% confidence
Finding: The file’s natural-language header and comments are entirely in Chinese, which imposes a specific language context on maintainers or users of the skill without any opt-in or documented locale constraint. Under the policy, language-specific behavior should either offer choice or clearly justify the locale limitation.

Intent-Code Divergence

Medium

Confidence: 95% confidence
Finding: The inline comment at L185 says these are newly added exports, implying working support for those features. However, `getJiuZongMen` and `getShenSha` are exported before their later definitions, and `getShenSha` contains an invalid object literal at L257 (`吉凶'`), which would prevent the module from loading at all; this contradicts the documented claim that the skill now supports these additions.

Intent-Code Divergence

Low

Confidence: 81% confidence
Finding: The manifest claims support for full chart construction and analysis, while multiple code comments label core routines as `简化版`, including the main divination steps and later added systems. This is not just omitted detail: the inline documentation acknowledges the implementation is simplified, which diverges from the stronger capability claim that the skill supports these systems as such.

Static analysis

No suspicious patterns detected.