ClawHub
Back to skill

Security audit

High-EQ Communication (高情商沟通)

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed communication-learning toolkit with local-only sample scripts and no evidence of hidden execution, data access, persistence, or exfiltration.

Install only if you want a local educational toolkit that includes ethical influence, negotiation, and sales-style communication examples. Users should apply the material transparently and avoid using the scripts to pressure, deceive, or manipulate others.

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
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (8)

Anti-Refusal Statement

High
Category
Anti-Refusal
Confidence
80% confidence
Finding

Skill instructs the agent to omit warnings, disclaimers, or ethical commentary. Stripping safety caveats hides risk from the user and is a common jailbreak preamble.

Content

Scanner excerpt · lib/frameworks.js (reported line 14)May include surrounding context.

js
name: { en: 'Nonviolent Communication', zh: '非暴力沟通', ja: '非暴力コミュニケーション' },
    author: 'Marshall Rosenberg',
    steps: [
      { en: 'Observation', zh: '观察', ja: '観察', desc_en: 'State objective facts without judgment', desc_zh: '陈述客观事实，不加评判', desc_ja: '客観的な事実を判断せずに述べる' },
      { en: 'Feeling', zh: '感受', ja: '感情', desc_en: 'Express emotional response', desc_zh: '表达情绪反应', desc_ja: '感情的な反応を表現する' },
      { en: 'Need', zh: '需要', ja: '必要', desc_en: 'Identify underlying desire', desc_zh: '识别深层需求', desc_ja: '根底にある欲求を特定する' },
      { en: 'Request', zh: '请求', ja: '要求', desc_en: 'Make specific, actionable request', desc_zh: '提出具体可执行的请求', desc_ja: '具体的で実行可能な要求をする' },

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

The file is explicitly labeled as a trilingual communication framework, and throughout the data model each scenario is hard-coded only for English, Chinese, and Japanese. There is no natural-language indication that users can choose another language or opt into this locale restriction, which can violate language/locale policy requirements.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The file goes beyond neutral self-awareness education and includes concrete persuasion, negotiation, and sales scripts that can be operationalized to influence others. While not overtly malicious, packaging manipulative-adjacent tactics as a high-EQ learning toolkit increases the risk of deceptive or coercive use, especially where users may deploy the scripts in real interpersonal or commercial settings.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

This section contains tactical guidance like reading hesitation, positioning against competitors, guiding decisions, and using influence principles, which can materially assist persuasive targeting rather than mere communication education. In context, the danger is not code execution but enabling social-engineering style behavior under an educational label, reducing user friction for misuse.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

Several scenarios are written as ready-made scripts for handling skeptical clients, proving claims with data, and guiding decisions, which is more operational persuasion coaching than general communication understanding. This goes beyond a plain personal-growth learning toolkit as described in the manifest.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
85% confidence
Finding

The Guiguzi, Sun Tzu, and Cialdini sections include tactics such as controlling conversation flow, strategically presenting alternatives, guiding decisions, and using influence principles. While some of this can be educational, these capabilities are more aligned with persuasive influence training than the manifest’s narrower framing around personal growth and understanding communication dynamics.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The manifest describes an educational resource for personal growth, self-awareness, and understanding communication dynamics. These Sun Tzu entries frame the content around researching clients and competitors, pitching, pressure dynamics, and winning against rivals, which shifts the behavior toward business persuasion and competitive strategy rather than the stated self-development focus.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
80% confidence
Finding

The CLI initializes lang = 'en' as the default output language. Under the stated policy, forcing a specific language without user opt-in can be a natural-language policy concern, and this file does not indicate that English was explicitly chosen by the user before output is generated.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.