ClawHub
Back to skill

Security audit

电子书阅读助理

Security checks across malware telemetry and agentic risk

Overview

This is a straightforward ebook/PDF reader and note-generation skill, with only minor routing and language-usability caveats.

Reasonable to install for ebook analysis. Use it only on ebooks or PDFs you are comfortable having the agent parse, and confirm it is being invoked for an ebook task rather than unrelated note-taking if the broad triggers activate unexpectedly.

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 (2)

Vague Triggers

Medium
Confidence
91% confidence
Finding
The trigger phrases are very broad generic tasks such as reading ebooks, note-taking, content extraction, and chapter analysis. In a skill-routing environment, this can cause the skill to activate for ordinary user requests beyond its intended scope, increasing the chance of unintended file processing or over-collection of user content.

Natural-Language Policy Violations

Medium
Confidence
79% confidence
Finding
The description and capabilities are written only in Chinese, which can bias invocation and interaction toward a single language without indicating user choice or locale constraints. This is primarily a safety and usability issue: users or orchestrators may receive unexpected language behavior, causing confusion, misinterpretation of summaries, or mishandling of user requests.

VirusTotal

51/51 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.