OpenClaw NLP 超级技能

Security checks across malware telemetry and agentic risk

Overview

This is a coherent NLP text-processing skill with normal model-download and privacy caveats, but no evidence of hidden access, credential theft, destructive behavior, or exfiltration.

Install only if you want a Chinese-focused local NLP helper and are comfortable with Python NLP dependencies plus optional transformer models that may be downloaded and cached. Do not process secrets, credentials, regulated personal data, or confidential business text unless you understand your local logging, cache, and model-download controls; review outputs manually for important translation, summarization, generation, or correction tasks.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (9)

Vague Triggers

Medium

Confidence: 92% confidence
Finding: 该文件是 markdown，适用 SQP-1。L507 将“文本分析”列为触发词，但未限定具体任务边界、上下文或排除条件；这类短语在日常请求中非常常见，可能导致误触发或与其他分析类技能冲突。

Vague Triggers

Medium

Confidence: 88% confidence
Finding: “文本生成”是高度通用的能力描述，不足以区分该技能与其他写作、摘要、翻译或内容创作技能。文档虽列出触发词，但未提供负例、上下文限定或明确激活条件，存在歧义触发风险。

Missing User Warnings

Medium

Confidence: 84% confidence
Finding: 该文件是 markdown，适用 SQP-2。文档详细描述了会接收并处理用户文本内容的功能，但“注意事项”仅提及模型下载、内存和性能，没有提醒用户避免提交敏感个人、机密或受监管数据，也未说明生成/摘要/翻译结果可能不准确。

Vague Triggers

Medium

Confidence: 93% confidence
Finding: 该文件是 markdown，适用 SQP-1。L507 将“文本分析”列为触发词，但未限定具体任务边界、上下文或排除条件；这是一个高频泛化表达，可能覆盖摘要、分类、情感分析等多种能力，导致误触发或与其他文本处理技能冲突。

Vague Triggers

Medium

Confidence: 88% confidence
Finding: “文本生成”本身覆盖续写、扩写、改写、创作等多种常见用户意图，且文档未说明何时由该技能接管、何时不应触发。对于超级技能包而言，这种宽泛触发词会增加与通用写作或聊天能力的重叠风险。

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: 该文件是 markdown，适用 SQP-2。L548 仅提到“首次使用transformers功能会自动下载模型”，但未明确提示这意味着会发生网络连接、下载第三方模型文件并占用本地存储；这属于可能影响用户网络、隐私预期和系统资源的行为，应有更显著的用户告知。

Natural-Language Policy Violations

Low

Confidence: 78% confidence
Finding: SQP-3 适用于所有文件类型。文档从标题、功能说明到示例均默认中文语境，并在 L551-L562 强调“中文优先”“重点支持中文”，但未明确说明是否允许用户选择其他交互语言或在非中文场景下如何处理，存在一定的语言/区域偏好约束。

Natural-Language Policy Violations

Medium

Confidence: 92% confidence
Finding: This code defines the skill name, description, comments, trigger phrases, error messages, and many result labels in Chinese only. That creates a language policy concern because users are not offered a locale choice or opt-in, and the skill behavior appears to assume Chinese as the default interaction language.

Missing User Warnings

Low

Confidence: 87% confidence
Finding: This is a markdown file, so SQP-2 applies to omissions in user-facing safety disclosures. The document instructs users to send arbitrary text to translation, NER, summarization, correction, and generation features, but it does not warn against providing confidential, personal, or sensitive content despite these behaviours affecting user data and privacy.

VirusTotal

61/61 vendors flagged this skill as clean.

View on VirusTotal