ClawHub
Back to skill

Security audit

夸克扫描王 - 文件扫描增强 Alibaba-Quark-Scan

Security checks across malware telemetry and agentic risk

Overview

This skill appears to do what it says: send a user-provided image to Quark's scan service for enhancement and save the returned image locally.

Install only if you are comfortable sending the images you process, including IDs, contracts, receipts, exams, or notes, to Quark's remote service using your SCAN_WEBSERVICE_KEY. Review the service's privacy terms, avoid sensitive images unless necessary, and periodically clean the temporary output folder if local persistence matters.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Lp3

Medium
Category
MCP Least Privilege
Confidence
93% confidence
Finding
The skill clearly handles sensitive capabilities: it reads local files or base64 image content, accesses an API key from the environment, and sends images to a third-party network endpoint. Yet no explicit permission declaration is present, which weakens platform enforcement and user visibility into data access and exfiltration behavior.

Vague Triggers

Medium
Confidence
87% confidence
Finding
The description instructs the agent to trigger the skill even when the user does not explicitly ask for enhancement, as long as the request loosely relates to clarity or cleanup. This broad auto-trigger can cause unintended transmission of user images to a third-party service, especially when users may expect local handling or are discussing images only generally.

Vague Triggers

Medium
Confidence
90% confidence
Finding
The fallback 'scan document' intent is defined so broadly that generic optimization requests can be captured even when the user's goal is ambiguous. In context, this increases the chance of over-collection and unnecessary third-party processing of documents that may contain sensitive information such as contracts, IDs, receipts, or notes.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The runner sends user-supplied image paths, URLs, or base64 image content to an external OCR service without any user-facing disclosure, consent gate, or visible privacy warning in this execution path. In the context of a skill advertised as image enhancement rather than recognition, this increases the risk that users provide sensitive documents or IDs without understanding they are being transmitted to a remote service for OCR-related processing.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.