ClawHub
Back to skill

Security audit

职场高效能习惯养成教练

Security checks for vulnerabilities and agentic risk

Overview

This is a text-only workplace coaching skill whose risks are mainly privacy and over-broad activation, not hidden code or malicious behavior.

Install only from the reviewed source or pinned release, and be mindful that coaching prompts may involve private workplace situations, relationship details, and personal values. Do not record or store other people’s statements or reactions without consent, and avoid putting confidential employer, HR, or regulated information into the chat unless your agent environment is approved for it.

SkillSpector

Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (22)

Context Leakage

High
Category
Data Exfiltration
Content
#### 📏 Every Check-in Task Is Verifiable

Most habit plans fail because "be more proactive today" isn't checkable. Here each daily task carries an explicit completion standard — "log at least 3 language substitutions", "produce a list tagged with all four quadrants", "record the conversation and the other person's reaction". Done or not done is unambiguous, and each task is capped at 15-30 minutes.

#### 🎯 Diagnosis Covers All Seven, No Cherry-Picking
Confidence
85% confidence
Finding
The instruction to 'record the conversation and the other person's reaction' can encourage capture of third-party communications and behavioral details that may be sensitive, confidential, or regulated depending on workplace and jurisdiction. In a coaching skill, this is more dangerous because users may follow the advice in real employment contexts involving coworkers, managers, or HR-sensitive discussions.

Hidden Instructions

High
Category
Prompt Injection
Content
</head>
<body>

<!-- Nav -->
<nav><div class="nav-inner">
<div class="nav-logo">🧭 mu-highly-effective · <span data-en="Effective Habits Coach" data-zh="职场高效能习惯养成教练">职场高效能习惯养成教练</span></div>
<ul class="nav-links">
Confidence
70% confidence
Finding
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Hidden Instructions

High
Category
Prompt Injection
Content
</div>
</div></section>

<!-- ✨ 核心亮点 -->
<section id="highlights" style="background:var(--bg-warm)"><div class="container">
<h2 class="section-title"><span data-en="✨ Core <em>Highlights</em>" data-zh="✨ <em>核心</em>亮点">✨ <em>核心</em>亮点</span></h2>
<p class="section-desc"><span data-en="Not another book summary &mdash; a habit operating system you can actually run." data-zh="不是又一个读书笔记，而是一套可以直接跑起来的习惯操作系统。">不是又一个读书笔记，而是一套可以直接跑起来的习惯操作系统。</span></p>
Confidence
70% confidence
Finding
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Hidden Instructions

High
Category
Prompt Injection
Content
</div>
</div></section>

<!-- 🚀 工作流 -->
<section id="workflows" style="background:var(--bg-warm)"><div class="container">
<h2 class="section-title"><span data-en="🚀 <em>Workflows</em>" data-zh="🚀 <em>工作流</em>">🚀 <em>工作流</em></span></h2>
<p class="section-desc"><span data-en="6 independent workflows. Trigger and run on demand." data-zh="6 个独立工作流，触发即跑，互不依赖。">6 个独立工作流，触发即跑，互不依赖。</span></p>
Confidence
70% confidence
Finding
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Vague Triggers

Medium
Confidence
94% confidence
Finding
The trigger for Workplace Q&A is described as 'Any pasted workplace situation, conflict, or confusion,' which is not specific enough to avoid collisions with normal user chat. It does not define scope limits or provide examples of non-triggering workplace conversations.

Vague Triggers

Medium
Confidence
88% confidence
Finding
Phrases such as 'check-in,' 'what do I really want,' 'life direction,' 'no time,' and 'trust issue' are common natural-language expressions that can appear in many contexts unrelated to this skill. Without stronger contextual limits, these triggers are susceptible to unintended activation.

Vague Triggers

Medium
Confidence
95% confidence
Finding
The Quick Mode trigger says the skill activates when the user pastes a workplace situation, problem, or question directly. That condition is very broad and overlaps with ordinary conversation about work, without clear exclusion conditions or negative examples to distinguish when the skill should versus should not activate.

Skill Enumeration

Medium
Category
Agent Snooping
Content
git clone https://github.com/muippt/mu-highly-effective.git ~/.claude/skills/mu-highly-effective
```

> Using a different agent? Just drop the folder wherever your tool loads skills from. Project-level works too: `.claude/skills/mu-highly-effective`.

**2. Verify** — restart your agent and confirm the skill is picked up
Confidence
85% confidence
Finding
Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.

Vague Triggers

Medium
Confidence
95% confidence
Finding
The workflow table says the workplace-confusion flow can be triggered by essentially any pasted workplace scenario, conflict, or confusion. In an agent environment, such broad natural-language matching can cause unintended invocation during ordinary conversation, leading the skill to steer responses when the user did not explicitly ask for this framework. The issue is not code execution, but prompt-scope overreach and misrouting of agent behavior.

Vague Triggers

Medium
Confidence
93% confidence
Finding
The interactive mode triggers are very general, including direct workplace scenarios, generic requests for plans, and common phrases like time shortage or relationship tension. Such low-specificity triggers raise the probability of accidental activation during everyday language, which can bias the agent toward this skill's framework even when the user intended a normal conversation. The skill is content-only, so the danger is behavioral overreach rather than system compromise.

Vague Triggers

Medium
Confidence
92% confidence
Finding
The quick-start text says that pasting any workplace scenario will automatically identify habits and respond in the skill's format. This encourages ambiguous auto-activation from broad user input and increases the chance the agent applies the skill without clear consent or context boundaries. In multi-skill environments, that can override more appropriate behaviors or cause surprising responses.

Natural-Language Policy Violations

Medium
Confidence
91% confidence
Finding
L003 的展示名称及全文指令均以中文固定描述该技能的交互方式，但未说明是否允许根据用户偏好切换语言。对于通用公开技能，这构成潜在的语言/locale 强制约束，缺少用户选择或显式 opt-in。

Vague Triggers

Medium
Confidence
93% confidence
Finding
The quick-mode trigger is defined as any user directly posting a workplace scenario, confusion, or question, which is broad enough to overlap with ordinary conversation. In a shared routing environment this can cause unintended activation of the skill, leading to context hijacking, misrouting, or the assistant applying this coaching workflow when the user did not ask for it.

Vague Triggers

Medium
Confidence
89% confidence
Finding
The guided-mode trigger activates on broad requests like diagnosis, mission statement writing, check-in planning, emotional account review, or time-matrix organization without clear scope limits or exclusion rules. This increases the chance that unrelated user requests are captured by the skill, producing unintended behavior or overriding higher-priority/general assistant handling.

Vague Triggers

Medium
Confidence
95% confidence
Finding
The check-in mode relies on generic terms like '打卡' and common day counts such as 7, 21, and 49 days, which are likely to appear in everyday planning discussions. This can cause accidental mode switching and unauthorized skill routing, especially in multilingual or broad productivity contexts where such phrases are common.

Natural-Language Policy Violations

Medium
Confidence
84% confidence
Finding
L267-L271 规定了固定的话语风格与措辞习惯，但结合全文中文规则使用，未说明这些规范是否仅适用于中文场景或可按用户语言偏好调整。对公开技能而言，这会形成默认强制语言/locale 的风险。

Intent-Code Divergence

Medium
Confidence
97% confidence
Finding
The page claims to be self-contained and zero-dependency, but it loads remote resources and sends users to external sites for installation and promotion. This creates a trust-boundary problem: remote badge images and off-site install links can change over time, enable tracking, or deliver different content than what was reviewed.

Natural-Language Policy Violations

Low
Confidence
82% confidence
Finding
The document is initialized with `lang="zh"`, making Chinese the default language presentation. Although a language toggle exists later, the file still forces a specific locale on first load rather than offering an explicit initial language choice or opt-in.

Description-Behavior Mismatch

Low
Confidence
94% confidence
Finding
The manifest describes a Seven Habits coaching skill that diagnoses habit gaps, generates plans, guides mission statements, audits emotional bank accounts, and organizes time matrices. In this file, the operative behavior is rendering marketing content and linking users to external sites such as the skill hub, GitHub, social media, and product pages; none of the claimed coaching functionality is actually implemented here.

Natural-Language Policy Violations

Low
Confidence
95% confidence
Finding
This markdown file presents all instructions, templates, and user-facing content exclusively in Chinese. Under the policy criteria, forcing a specific language without user opt-in can be a natural-language policy violation unless the locale restriction is explicitly justified.

Natural-Language Policy Violations

Low
Confidence
97% confidence
Finding
This markdown file contains user-facing instructions, prompts, and output templates only in Chinese, which effectively forces a specific language/locale. Under the policy, language constraints should either provide user opt-in or be explicitly documented and justified.

Natural-Language Policy Violations

Low
Confidence
94% confidence
Finding
This markdown file contains natural-language instructional content exclusively in Chinese, and nowhere indicates that the user can select another language or that the skill is intentionally limited to a Chinese-speaking context. Under the language/locale policy criterion, forcing a specific language without user opt-in is a policy concern.

Static analysis

No suspicious patterns detected.