ClawHub
Back to skill

Security audit

long-doc-translation

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed long-document translation workflow with local quality-check and HTML-building scripts, not hidden data access or unsafe automation.

Install this if you want a Chinese full-book or long scholarly translation workflow. Be aware that broad translation triggers may activate it for smaller translation requests, and the deduplication script only modifies chunk files when run with --apply, so review dry-run output before applying changes.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (3)

Vague Triggers

Medium
Confidence
85% confidence
Finding
The trigger list includes broad phrases such as general book translation requests, which can cause the skill to activate outside narrowly intended scenarios. Overbroad activation increases the chance of unintended routing, mishandling sensitive documents, or surprising users by invoking a heavyweight workflow when they asked for ordinary translation help.

Natural-Language Policy Violations

High
Confidence
92% confidence
Finding
The skill mandates Chinese output without confirming the user's preferred output language, which can lead to unintended data transformation and user surprise, especially for multilingual or sensitive scholarly materials. In automated routing contexts, forcing a language choice without opt-in can cause incorrect handling of requests and reduce user control over how their content is processed.

Vague Triggers

Medium
Confidence
89% confidence
Finding
The trigger list contains broad phrases such as '全文翻译', '学术翻译', and '长篇翻译' that can match many ordinary user requests and cause the skill to auto-load without clear user intent to invoke this specific workflow. In an agent marketplace context, unintended auto-loading can expose user content to a more powerful or specialized processing pipeline than expected, increasing the chance of over-collection, misrouting, or surprise actions.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.